Articles
The internet of things – the connection of devices to the internet, enabling them to communicate with us, each other and applications – has long been promised to deliver benefits to industry. While some may feel it has yet to live up to its promise, manufacturing is beginning to realise a major benefit in the better monitoring and diagnosis of machines, which can prevent costly failures and halts in production.
This area has been a large focus for major companies such as GE, but as it stands a lot of the software on the market can be pricey or hard to use. New kids on the block Senseye are hoping to corner the market with an affordable and user-friendly software solution called Prognosys.
Senseye co-founders Alexander Hill, Simon Kampa and Rob Russell previously worked with condition monitoring systems in the aerospace and defence sectors, where this technology is fairly mature. In contrast, the manufacturing sector is lagging behind, using manual processes to monitor machines and dependent on specialist data scientists or engineers for analysis.
Hill says: “What we’ve done is try to automate the data science process. It means we can give people vital data much quicker and cheaper.”
Prognosys works by taking in measurements such as vibration, humidity, acoustic emissions and power usage from sensors already in the machine, or retrofitted by Senseye. The sensors stream data to the cloud via a wireless or wired connection. Prognosys analyses each measurement to identify trends and predict machine failure. It also checks whether its past predictions were correct.
Data is available to the factory floor from day one but becomes more useful in a week or two, after the software has been able to build a picture of normal operations.
Senseye aims to make the system usable by anyone. Prognosys sends simple messages to users, explaining in “natural language” where an error has been detected, when a failure is likely and which particular areas should be looked at. This is done through its easy-to-use web app and soon via text message.
Users can see data by logging onto a Facebook-style feed. A timeline will show a picture of a motor, for example, a graph with the problem area highlighted in red and a simple text explanation. Users provide feedback on whether this information is useful, in which case the software will look into it further, or whether it should be ignored in the future.
Senseye is currently piloting Prognosys with a number of manufacturers, including Parvalux, a maker of electric motors. It is too early to provide hard data but it expects to see downtime decreasing by 20-30%.
Hill says that the price of Prognosys will be orders of magnitude cheaper than what is currently out there, with some systems costing up to £5 million a year.
Senseye is offering manufacturers a free trial of Prognosys and will provide a no-cost installation and use of the cloud solution for a period of three months.
For more information visit the website:
info.senseye.io/prognosys-pilot