The Industry 4.0 digitalisation trend, together with developments in computer and sensor technology, has brought new possibilities to harvest from machine learning in operational- and maintenance management.
The hype seems to cause some people believe that now; as long as you get access to any, and a lot of sensor data, you just drag and drop it into Azure or it’s likes, then fantastic novel insight will arise like magic.
Well you may strike luck, but the key to success still involves applying a large portion of knowledge and hard work!
MainTech experience is that the key to success is to apply domain knowledge up front: The input being knowledge based diagnostic techniques, as well as data driven statistical diagnostics.
- Define and structure the context of the problem(s) to be solved
- Apply domain knowledge to define the context of the problem and to how to apply the results
- Define the speed of each problem process changes
- Pick the right physical- or chemical measurement method for the data harvest
- Select sensors or measurements with the appropriate range
- Select data architecture and common transport protocol
- Determine signal conditioning, Analogue to digital data acquisition at the right sampling frequency and common time stamp
- Apply hardware and connect software to facilitate the above
Now your system is set up to collect quality data for starting the successful application of Machine Learning.
MainTech AS holds strong domain competence in:
Process control and automation systems, Lean manufacturing, Sensor technology, Maintenance management, Materials technology and deterioration mechanisms, Condition monitoring and sensor technology, and successfully applied Machine Learning.