Machine Learning Maintenance & the Internet of Things

Machine Learning Maintenance & the Internet of Things

Major lift organisations see the lift industry as an ongoing maintenance one. This point of view is important in order to ensure you can run your lift as efficiently as possible. Customers will always be frustrated by a lift that is out of action, so being able to identify faults as quickly as is possible is essential to many installations; especially in residential or busy retail areas.

Operating a passenger lift is often a fairly unglamorous service. Those using the lift get in, and without any appreciation of the mechanics at work, expect the lift to be fully functioning. Of course, lifts nearly always are. With reliable platform lifts offering a solution that’s easy to mould to any environment; be it goods, retail or a simple passenger lifts, an installation isn’t as terrifying as it was previously.

It’s obviously very difficult to take a preventative approach to lift maintenance. Often, a fault is detected after it’s occurred because traditionally there is no other way of discovering an operating error. There are maintenance schedules and checks that of course limit this, but sometimes it is an unavoidable element of the industry we operate in.

It’s becoming increasingly common for major lift manufacturers to develop and deploy advanced technological systems that ensure lifts never break down. Issues are identified and fixed before they go on to cause larger problems. This obviously differs from the previous approach of offering maintenance whereby anything that goes wrong will be put right retrospectively.

The computer data collected by the computer systems operating lifts already compiles data from the passenger lifts (especially those installed in extremely tall buildings) that record how doors are opening, the floor the button was pressed on and other essential operational data. It is also records error messages which are used by lift engineers to work on the lifts and fix faults.

These new systems – being trialled already for use in the super-tall skyscrapers- collate these codes and send them out in advance. This ‘learning’ system can use this information to build an individual profile of the way the passenger lift operates and can give a great insight into faults and operational problems before they actually happen.

An example of this would be the system monitoring the motor current over the same lift load over time. If the motor current increases during journeys with the same load, it can easily be deducted that a bearing needs to be changed. The lift is having to work harder to complete the same task and therefore the necessary maintenance can be carried out before a serious fault occurs. The systems would run automatic updates and schedule repairs, maintenance or routine services based off of this system, automatically reporting to the lift engineers.

Obviously, there is a vast range of lift type out there, with different materials, operating environments and heights. This means the data gathered is extremely complex, and the biggest challenges for the companies trying to develop this is making sense of this data and being able to interpret it efficiently. Experts from Microsoft have been called in to aid on this as the advanced system continue to develop.

This system wouldn’t be necessary for some of the more simple installations. Here, maintenance packages based off of the more reactive schedule work absolutely fine, and are by far the most cost-effective solution for those having lifts installed in their buildings. However, if you’re in the business of designing the world’s tallest skyscraper, this could well be the future of lift industry maintenance for you!