Demandbase Connect

June 15, 2007

Old plant, new mission

Pages: 12345

Old dog learns new tricks

The first target of the Sim Gideon controls upgrade project was the 350-MW Unit 3, which is powered by a tangentially fired, single-furnace boiler from Combustion Engineering and a D-8 steam turbine from General Electric. During this phase of the project, engineers used an advanced control software tool to interface to the unit's existing I/A control system from Foxboro (www.foxboro.com). The hookup had two objectives: 1) to improve the unit's heat rate and reduce its NOx emissions without increasing its CO2 emissions and 2) to reconfigure the unit for rapid load-following service.

In 1995, Unit 3's original analog boiler controls and GE Mark I turbine controls were replaced by a Foxboro IA distributed control system—serial No. 1 installed on a large steam turbine. Along with the boiler and turbine controls, burner management and controls, data acquisition, motor controls, and all remaining balance-of-plant equipment were brought under the IA umbrella. The return on this investment was better control over a wider load range at higher ramp rates than was possible with the analog system. The IA system increased controllable ramp rates, both in MW/min and the allowable control range.

For example, the range for area control had been 120 MW to 290 MW, with manual operator control above and below those values. The new controllable range with the digital controls was safely set at 60 MW to 340 MW. In addition to the 110-MW expansion of load range, automatic generation control (AGC) could now be set at 40 MW/min across a large portion of the range. In 1998, a neural net–based process optimization modeling system (POMS) was interfaced with the IA system to provide NOx reduction, lower heat rates, and advanced control features such as induced flue gas recirculation (IFGR).

In IFGR, a damper regulates the amount of flue gas recirculated to the inlet duct of the forced-draft (FD) fan. For maximum NOx reduction, the amount recirculated is maximized. However, at higher unit loads the amount must be reduced to keep the power consumption of the FD fan motor below its allowable limit. The control problem is to maintain the amount of gas recirculated near its highest level without burning out the FD fan motor.

On Sim Gideon Unit 3, the benefits of the POMS retrofit—a heat rate reduction of 0.34%, an 8% cut in NOx emissions, and a CO2 reduction of 3.4%—were so significant that the project paid for itself in just six months. A subsequent upgrade of the unit's controls, in 2002, produced a further reduction in NOx emissions of nearly 25% and a heat rate improvement of 0.65%. Like the first project, this one was conceived and executed by a team of Invensys Foxboro employees and LCRA staffers at Sim Gideon Power Plant.

Since deregulation, ERCOT—rather than LCRA—has decided which units at Sim Gideon are dispatched, and at what capacity, to meet statewide system demands. In 2002, to better accommodate the emerging ERCOT markets and to make itself more competitive, LCRA implemented a new strategic efficiency plan for Sim Gideon Unit 3. The plan called for the implementation of a new energy management system to change the drivers of the AGC system from pulses to setpoints, essentially placing the unit under local operator control.
 

Stressed-out rotor

The first grid-induced test of the upgraded controls occurred in April 2004, when ERCOT mistakenly dispatched Sim Gideon Unit 3 to catch a large load change that should have been allocated across several units. Unit 3 was ramped up and down between 60 MW and 320 MW several times. Each cycle took from 7 to 9 minutes to complete, and the average rate of load change was 35 MW/min.

The Foxboro control system held all controllable variables within acceptable limits during this AGC roller-coaster ride. The increased ramp rate appeared not to adversely affect control of drum level, superheater and reheater outlet temperatures, boiler master, airflow, or automated burners. All processes seemed to work in unison, doing their part to maximize unit ramp rate. Although the unit responded to its task successfully, the venture into severe-service territory was not without consequence: the accidental discovery of an uncontrollable temperature change in the first stage of Unit 3's steam turbine.

Sim Gideon plant operations and engineering personnel reviewed the long-term risks and benefits of load following at this ramp rate. After reviewing the data, plant personnel became very concerned about what was not being controlled. The primary concern was that the rate of change of turbine first-stage shell temperature was excessive on sustained load ramps. The data files were printed and compared with starting and loading charts supplied by GE.

The comparison raised questions about potential damage to the turbine rotor. In the past, Sim Gideon Unit 3 had a more restricted load-following range. With the upgrade of its control systems, which enabled much faster ramp rates, the allowable level and rate of temperature change rose considerably. What's more, a loading cycle now could be completed three times faster—in 20 minutes instead of 1 hour.

As soon as the Sim Gideon staff began comparing data, they knew they had a problem. GE's charts for rate of rotor temperature change topped out at 600 degrees F/hr, and Unit 3 was averaging over 900 degrees/hr, with peaks as high as 1,600 degrees/hr. GE was asked to supply new curves and to confirm the extent of damage that the turbine could suffer from being operated in this way.

According to the GE "1,000 cyclic life expenditure (CLE) curve" on hand, cycling the turbine three times that day in April 2004 had used up 0.3% of its life. In other words, the unit could be operated in that way only another 997 times without risking a catastrophic crack of its high-pressure (HP) rotor. This was unacceptable to plant managers, who asked both GE and Foxboro for guidance toward a solution.

To no one's surprise, GE suggested reducing the ramp rate so the turbine would operate within the limits of its cyclic life curves. Foxboro's advanced control group replied that none of the 50 large turbine control systems it had installed since serial No. 1 at Sim Gideon had reported any problems of this nature. But because many of these systems were operating at similarly high ramp rates, Foxboro agreed to look into the situation immediately.

Following an internal consultation, Foxboro's advanced controls group decided to optimize the rate of change using the GE CLE curves. This would require both predicting future temperature changes and analyzing the cumulative changes that the turbine had already endured. Foxboro sent personnel to the site to begin studying the system with an eye toward developing a strategy for mitigating the life-expenditure effects of unit load following and cycling.
 

Pages: 12345

RSS

 

Related Stories








Subscribe to POWERnews

First Name Address Email Last Name City Company
Title
State      Zip Code




© 2012 Tradefair Group, an Access Intelligence LLC company.