District Energy - 2007 Fourth Quarter

support to assemble plant and process technical information and to monitor the installed ODST system. (There was no direct cost to UI for use of the system or setup.)

The ODST software has proven to be an effective tool to help plant operators and engineers understand equipment operation and get advance warning of potential problems. The beta testing at UI continues, with the plant providing feedback to ETS on how to make ODST knowledge part of routine plant operations. (The system has now also been installed at other plants and facilities.)

An early experience with the ODST system involved one of the UI’s turbine generators. This is a four-bearing machine, with two bearings on the turbine and two bearings on the generator; bearings number two and number three share a common housing. The control system and HMI monitors and displays many parameters including lube oil pressure, the temperature of each bearing, and generator load. These data are also archived in the process data historian. Alarms are set for bearing temperatures and other process parameters. As the generator operates, historical relationships between the temperature of each bearing relative to the other bearings develop during normal operation.

Shortly after installation, in early 2006, the ODST system alerted operators to an abnormal relationship between machine

bearing temperatures. Individual bearing temperatures were below their individual alarm limits and did not appear to have changed significantly. A maintenance shutdown was scheduled to investigate an oil leak from the bearing housing oil seal. With the bearing housing open, rolling out the bearings for inspection was relatively easy to do. Two wiped bearings were discovered when this action was taken. Without the information from the ODST system, there would not have been a compelling reason for plant staff to roll out and inspect the bearings. Had the wiped bearing condition not been corrected, and machine operations continued, a much more serious bearing failure could have occurred in the future.

The process relationships monitored by the ODST software are not confined to variables such as pressure, temperature and flow. Calculations can be also performed to determine additional variables, including heat flux. As illustrated in figure 1, the UI system did just that. This sample ODST display illustrates a probability density function plot of two derived process variables in UI’s stoker coal boiler. Boiler economizer flue-gas flow is plotted on the y-axis, and flue-gas heat content is plotted on the x-axis. The plot resembles a topographical map, and the darker the yellow color, the more likely the two variables will operate in that range. Recent data are

Figure 1. Process Variable Relationship Plot: Sample Display From the University of Iowa’s Operator Decision Support Tool Software. This example illustrates probability density of the heat content in boiler exit gas for a given exit gas flow. High probability areas are in the center of the yellow area. Like a topographical map, higher elevations correlate to higher probabilities.

Source: Operator Decision Support Tool software, Energy Technology Solutions LLC, as supplied by the University of Iowa.

plotted in red and the latest data point identified. This display indicates that the process is operating outside its normal limits and this condition may be worthy of investigation.

In this instance, the boiler feed water top heater had not been started, and abnormally cold feed water was being supplied to the boiler. Once the feed water heater was placed in service, the process returned to the higher probability operating area. The system did not tell why this abnormal condition existed, just that it did.

Investigating alerts and determining why process relationships have shifted is an excellent way for plant staff to gain a greater understanding of the process. Figure 1 shows that for a given boiler gas flow through the economizer, the boiler exit gas has a higher-than-normal heat content, as shown by the red data points; without the feed water heater in service, more heat is carried out of the boiler with the flue gas for a given boiler load. As operators learned the effects of this piece of equipment being out of service, they added to their knowledge of how the boiler operates

There are many reasons why process relationships change. Most of these changes are caused by transients and evolutions, such as startup and shutdown; some provide information useful for preventing further degradation of systems and equipment. For example, UI recently performed an infrequently done stoker boiler fireside cleaning, as part of a scheduled maintenance outage. On boiler startup, ODST produced alerts related to abnormal heat transfer and temperature relationships. Investigation determined the historical data used to formulate the process relationships was based on a fouled furnace. Process data from a clean furnace was not available in the historian, since the historian had not been operational the last time the unit was extensively cleaned. When the cleanliness condition was corrected, the system needed to relearn the new ‘normal’ relationships.