Why We Need Historians

As we've seen, Industrial Data is used to help industrial firms understand and control their equipment and their processes so that they can produce the maximum output with the fewest possible defects. To perform historical analysis, firms store industrial data in Historians, which are databases that specialize in time series data.

So, why do we need specialized databases? Great question. A normal database is designed to store complex relationships between data elements. They are optimized to allow for data to be queried in ways that often leverage the relationships, to show meaning or value. An example might be how many stocks were above historical highs at the end of a market correction.

The needs for industrial data are slightly different. Here, we are looking to be able to accurately recreate a waveform from sampled data. It’s true that we may want to compare waveforms once recreated, but the focus of an Historian is to efficiently store sampled data to allow high-fidelity wave reconstruction.

There are a number of different techniques to compress time series data as well. Some look for efficiency of storage over quality of reproduction, while others look for quality of reproduction over efficient storage. For example, if you got a sampled reading for three seconds and the value didn't change, one compression technique is to store the value once, and only add a new value when the value changes. If on the fourth second the value increases, you only had to store the value at first reading and the fourth, i.e. you can skip storing the same reading during the second and third sample.

Another technique might be to average all samples together that are read during a second, so if you read 100 samples per second, you could average these together and store one value. Here, you wouldn't be able to recreate the waveform during the second time interval, but you may not need that for historical analysis.

Of course, Historians do many more things than just store data. They provide collection tools, sampling tools, compression techniques, data management, query interfaces and data distribution tools so other applications can use the historical data.

How do you use your Historian?

Proficy Historian

Brian Courtney

A recent transplant to the MidWest, Brian thinks Big Data “rocks.” He’s recently taken the Analytics piece of GE’s business under his wing, so if you have thoughts on any of these – MidWest, Big Data or Predictive Analytics – even rocks – follow Brian on Twitter @brianscourtney.

More Posts