Mining the Corporate Data Domain Safely and Securely Across the Enterprise

Part 3: Capture SCADA Data and Schema

Automation Breeds an Unforeseen Bonus

An immense amount of effort was put into Entergy’s requirement to make the system maintenance-free. One of the unforeseen results of this effort was the ability of the system to audit database changes. This feature has been enhanced by adding a set of reports that allow a user to compare the contents of any two SCADA databases that have been put into operation since the system started, providing a complete audit trail of database changes.

Entergy’s Doug Dollar, Project Lead in West Monroe, says that this feature alone has saved his team time. “The new system provides what SCADA Support Analysts were manually recording into spreadsheets,” said Dollar. “A review of the report provides a quick double check that all desired changes have been made correctly.” [Refer to Figure 1]

Figure 1: Pegasus provides a high integrity audit trail of all SCADA object actions via online screens and exported reports. The system clearly shows old vs. new values when a SCADA object is modified. In this report example, four analog data points had their Device Type changed from VCB to XFMR on 8/4/2007.

All the Data, All the Time

Every alarm, status, analog, analog limit, and accumulator in the database, including calculated values, is stored by the system. Because the data is captured periodically rather than by exception, the load on the EMS host is steady rather than a series of peaks and valleys that can wreak havoc on system performance. The stability of the system was proved during hurricanes Katrina and Rita when every data change and alarm was captured during these times of extreme system activity.

Included at each site is a fully redundant system that regularly collects the data from both the primary and secondary EMS machines. The system takes these two data feeds and decides later which one to use. [Refer to Figure 2] This approach has resulted in no known data losses since the system went into operation, with everything on-line and available for use. This data remains available for historical reference, forensics or other diagnostics long after any event occurs; i.e., the data is not subject to the “roll-off” that can sometimes happen with archival systems. The hardware was sized for five years of on-line storage, but keeping data for longer periods simply requires installing additional disk capacity and changing one system parameter.

Figure 2: SCADA measurements are tightly integrated with the EMS information, including the enabled host, limit pair(s), quality codes and system alarms. The screenshot shows a typical system failover from node WMSYSA (“System A”) to WMSYSB (“System B”), identifying data captured where neither node is enabled with the symbols “>> <<”. Also shown are the actual limit values for one of the displayed analogs.

Moreover, sampling data frequently and storing the median of the sampled data further enhances analog data reliability. This has the effect of removing erratic samples and results in a more statistically accurate representation of the measurement.

No Maintenance Required

To come as close as possible to meeting the zero-maintenance design goal, it’s necessary to keep track of devices over time, even when they have name changes. Name changes may occur, for example, if a Gas Circuit Breaker (GCB) is being replaced by a Vacuum Circuit Breaker (VCB). The solution is keeping an “invariant key” for each measurement in the system, and automatically remapping devices to the keys when a new SCADA database goes online. There is little if any staff intervention in this process.

Once a new SCADA database is put on-line, the system automatically stores the contents of the new database and all data collection and storage applications run though the update without further staff involvement, handling all necessary re-mapping on the fly. New data is then available for display, and points with database name changes automatically return an unbroken set of data across all time stored in the database. Since there are frequent name changes on the Entergy system, this feature is invaluable from both maintenance and data consistency perspectives.

In addition to the automatic mapping, the system also handles all necessary database partitioning, indexing, trimming and log deletion automatically, so that no database administration time is needed to maintain the system.

Although the system does occasionally require some maintenance by Entergy and/or Nobadeer, the zero maintenance target has resulted in a miniscule burden on the staff, especially when considering the volume of data captured.