Capping IT Off

Capping IT Off

Opinions expressed on this blog reflect the writer’s views and not the position of the Capgemini Group

The next analytics frontier – Business transformation through sensor data on the fly

The next analytics frontier to break through is intelligence tied to data from machines and sensors –the combination of streaming sensor data with complex event processing and business analytics is changing the way we look at, among others, maintenance and supply chain.  
These sensors generate and transmit data in real time, allowing for analysis to both monitor failure-indications as they occur, and predict break down before it occurs. These condition based and predictive maintenance solutions could significantly improve operational safety and reduce the quantity of unplanned breakdowns. A strategy for predictive maintenance needs to be combined with preventative and corrective maintenance strategies to find the optimal cost benefit ratio. It should also be integrated with maintenance planning and scheduling applications to ensure maintenance action before “point of failure”. For asset types where failure is more random by nature, redundancy will normally be the first preventive measure. Here, sensor data is usually stored and used for root-cause analysis to identify production weaknesses, systemic issues etc.  
Already this is affecting our daily life. Car producers are already equipping their vehicles with a wide variety of sensors – crash sensors, voice and data, speed, temperature & fuel level sensors. Newer models also have driving style recognition – rendering it possible for the car to identify an altered driving style likely caused by e.g. drowsiness, then alert the driver and possibly shut down the car if needed. This is done by transferring driver health data to a central database in real time – and “next best action” is transferred back to the vehicle. We believe that this sort of “Connect me” technology will enhance the ownership experience
The aviation industry has come a long way in condition-based maintenance, driven by strict regulations and high fixed operation costs. The airlines use sensor data to become more efficient in reducing the time each aircraft must be grounded, reducing the costs associated with warehousing, optimizing maintenance schedules, and virtually all other parts of their operation.  
So, what would a business case for improved usage of sensor data look like?
Gathering, storing and interpreting all this sensor data will tell you the state of your asset, when parts should be changed – and eventually predict which parts should be changed.
This can fundamentally change the way you are doing maintenance. By moving certain assets off a pre-scheduled time or interval based maintenance regime, you can now move into a more fact-based regime, based on real time observations. Hence parts that are not worn out get a longer life (reduced cost, higher uptime and environmental benefits), and parts that are about to break down or lose performance according to sensor data and algorithms, are replaced just in time.
This could significantly reduce time spent on unnecessary interval based maintenance, allocate resources and competency to more complex tasks and operations, and increase availability of the assets. Moreover, this will also affect supply chain management and volume of parts in stock, moving into just-in-time principles on most parts – dramatically reducing money tied up in long time storage.
To wrap this up, here’s a checklist on how to ensure success with predictive asset management and sensor data:
  1. Understand how the new opportunities regarding utilization of sensors and information technology could fit into the existing asset management strategy and frameworks.
  2. Assess what data is available which you are not using for analysis today.
  1. Decide whether you need to store or simply master the data.
  1. Understand data – put them into context and use them to describe your reality.
  1. Plan how to go from analytics to action; data does not do anything by itself.
  1. Think big – act small; develop PoCs and demos that give stakeholders something tangible. Start with a minimal solution – iterate and iterate and iterate until you have found the right solution.
  1. Put the usage of data into a business context; ensure a business case is in place.
  1. Ensure that top management stands behind the initiative, and will follow through.
Big Data essentials such as Hadoop and accompanying statistics and machine-learning libraries are affordable, scalable and mature. So, with both the data and the tools available to give tremendous business value, the question is: Are you ready to start?

About the author


Leave a comment

Your email address will not be published. Required fields are marked *.