<img alt="" src="https://secure.hims1nice.com/151009.png" style="display:none;">
  • Advanced Analytics Group Case Studies
LGnE Logo 200px

LG&E OSISoft Sensor Equipment Data Extraction
(Power Gen)

The Louisville Gas and Electric company was unable to effectively analyze terabytes of power plant sensor data.  LG&E utilizes a third-party vendor software (OSISoft) which contains a large network of sensors that are installed throughout their power plant facilities. The power plants have multiple generation units with different equipment configurations. (Power plant equipment includes coal-fired boilers that generate high-pressure steam, turbines that rotate with high-pressure steam, generators that produce electricity, Air Quality Control Systems (AQCS), pumps, cooling systems, and other pipeline equipment). In total, the number of unique equipment sensors was over 70K.

After careful analysis of the vendor system and utilizing the vendor’s .NET source code, KiZAN was able to extract thousands of records every twenty minutes. This was then used to perform three operations:

  1. A backfill process to load five years of data
  2. A forward process to keep loading every twenty minutes
  3. A final reconciliation process to run two weeks behind and update changes

After weeks of processing, a dataset of billions of records totaling 5 terabytes was loaded onto an IBM Netezza appliance.

Challenges

  • Unable to query multiple sensors greater than one day
  • Unable to view sensors across the fleet (all plants together)
  • Unable to extract data in a performant way
  • Unable to view five years of history

Solution

  • The extraction process loads all values into a single location
  • Data quality is verified and values match vendor software
  • The solution is fast enough to perform multi-year analysis
  • Data types are appropriately selected in order to limit row size and keep data as compact as possible.

Results

  • Business team members freed up to work on other projects
  • Automates selection of strikes to outage for easier identification
  • Presents spatial technology in new ways
  • Fully managed and supported by the KiZAN analytics team

 

Case Studies