In a new research, a team of researchers have been able to combine artificial intelligence with sensors. The AI-based maintenance system gathers data on industrial machinery, thus helping to make sensors smart. The system features to detect damage, wear, and error states. And, it can uniquely recognize when a previous unknown state of the machine arises, thereafter, learns from them and assigns to their underlying root cause.
Using this approach, small and mid-sized enterprises can automate their machine maintenance and service operations. This allows these companies to plan more precisely and avoid unpleasant surprises.
System features Detection and Interpretation of Changes in Machine
Meanwhile, vast number of sensors constantly collect data from industrial machinery. And, these data sets can be of immense use. When a machine operates normally, the way it shakes, hums, vibrates is unique to that device. However, when machine components start to wear out, these characteristics undergo subtle changes. Slight changes in vibrational behavior, minute temperature fluctuations, and minor shifts in measurement data all act as early signals of warning that indicate when a component begins to show signs of wear. Thus, the subtle variations needs to be detected within the sea of data that is produced.
A single sensor can produce a terabyte of raw data in just few days. Nonetheless, in addition to detect changes, it is important to know how to interpret these changes.
The team has been working with partners in academia and industry to develop a system that is able to extract valuable signal data from vast quantities of data being generated. For this, the system independently assigns signal patterns for specific damage, wear or error states, thus, makes the machine’s status permanently visible.