It is easy to say what predictive maintenance is – fixing something before it breaks – but it is much harder to say how it should be done.
By now, everyone is familiar with Internet of Things enabled predictive product maintenance for field services. Manufacturers can now continuously monitor condition using the IoT and predict when a machine, component or product is going to break down or fail, and then fix the problem before it happens.
Predictive maintenance analytics is going to transform the manufacturing industry – optimizing efficiencies that deliver cost savings and creating new revenue service models while increasing customer satisfaction and loyalty.
How you do this is easy to understand but not quite as simple to deploy. First you have to get architecture right.
- Unlimited device connectivity
For years, manufacturing and supply chain industry have been using sensored devices on the factory floor and elsewhere, but with little regard to what happened to the resultant data. Today, with IoT platforms collecting and analyzing the data to improve processes and the customer experience, there is a need for more and more IoT devices to be connected. In addition to basic performance data, data is needed about weather, street traffic, events, customer preferences, social media trends, etc.
- Correct security – your network needs to be hack-proof
Any internet-connected device is vulnerable to hacking. A determined criminal could interrupt manufacturing processes by hacking into the electricity grid, or could spy on a competitor’s new product. The more connected devices a manufacturer has, the higher the risk.
- Analytics to include condition monitoring and prediction at both the edge and the back-end data center
Some analytics, for example for remote data, can be performed on gateways on the edge of the cloud, whereas some more sophisticated ones need to be happening in the data center.
- Automated actions to unlock value
Data is useless unless you can take action on it. You will need rules and alerts to enable you to act when something goes wrong – or when someone wants to buy.
- Visual analytics
Because there is so much data to analyze and act upon, a dashboard which gives you a visual representation of what is happening and where is critical.
It is a simple formula, but the differences in technologies focus on three things.
- Speed to implementation; hooking up IoT devices and getting the data flowing.
- Complexity and types of algorithms supported; the algorithms deployed will be a competitive differentiator.
- Business rules to provide the ability to make decisions automatically based on varying needs – i.e. supplier A needs three weeks’ lead time to deliver while supplier B only requires two.
This simple formula delivers consistency and reliability, enabling manufacturers and logistics managers to get predictive maintenance right.