Trips in gas turbines supplying steam via HRSGs to industrial processes are more problematic than those involving electrical generation alone because of the logistical and financial implications of extended downtimes.

In process steam applications, a supplementary fresh air duct firing system can be used to allow HRSGs to remain in operation whilst the GT is not running, by replacing (at least) the total waste heat produced by the turbine. This process of changing from GT exhaust gas to conventional duct firing using fresh air is usually referred to as ‘change over’.

Boiler concept

The firing system for the boiler in use at the Voestalpine plant consists of two levels, each with two burners (Table 1 and Figure 1). Each burner can burn fuel using either the GT’s exhaust gas or the supplementary fresh air supply. These two combustion air supplies are separated using dampers: fresh air and primary GT dampers feed the burner core, while the secondary GT damper manages excess exhaust gases (Figure 2). This arrangement ensures that the ratio of air to fuel mass flow in the burner core is sufficient for combustion.

The plant is also equipped with a GT exhaust gas bypass stack to ensure that the turbine’s exhaust gases do not enter the boiler when the system is running in fresh air mode. This separation is carried out by a diverter damper which leads the GT exhaust gases either towards the boiler (‘diverter damper open’) or via the bypass stack (‘diverter damper closed’).

Problems with manual procedure

Because of interruptions in the process steam supply and the occurrence of several incidents at comparable plants in which equipment was damaged and personnel were endangered during manual change overs, it was decided that a new change over procedure was required at the Voestalpine plant.

In the event of a GT trip, a boiler trip often occurs because of too-slow replacement of the GT’s exhaust gas within the timeframe of the gas turbine’s rapid rundown. When the GT falls to 50% of its nominal speed, the exhaust gas flow is not sufficient to fire the boiler. Because the existing change over procedure was entirely manual at Voestalpine, a fast response, ideally less than a few seconds, was not possible. An automatic method was therefore envisaged as the best approach to shorten the change over timeframe and prevent an HRSG shutdown.

Seeking an automatic solution

There are two key challenges that must be faced in developing a new change over procedure. On one hand, switching off the supplementary firing (which would require a necessary purging of the HRSG) must be avoided. On the other hand, a build up of backpressure in the turbine (resulting from the rapid closure of primary and secondary dampers required to maintain a safe fresh air supply) must be averted to prevent damage to the turbine’s exhaust system.

A further problem associated with the procedure employed at the Voestalpine plant was that in switching off one burner level during change over, there was a considerable sag in the boiler’s steam production. To respond to this, a test change over was carried out in which all primary dampers were closed, thereby ensuring that the burner was fed with fresh air without risk of air loss. The test involved running the GT at a reduced load whilst observing the turbine’s backpressure. This allowed the maximum possible GT load for change overs involving the firing of both burner levels, to be determined. In the event of a GT trip the turbine load is always lower than this maximum level, therefore the test conditions were representative of those arising following a trip during normal operation.

Another important investigation was to determine which diverter damper positions resulted in the back flow of hot flue gases from the HRSG’s furnace. This test involved observing the pressure fluctuations in the GT exhaust gas duct in order to detect which diverter damper positions could be problematic. Following analysis of these results, it was decided that the installation of an additional pressure loss orifice in the bypass stack would prevent backflow. The critical diverter damper positions were avoided by separating the boiler from the GT exhaust system using the primary and secondary dampers, thus allowing the diverter damper to occupy a safe intermediate position. The diverter damper ensures that backflow is significantly reduced, yet allows sufficient passage of GT exhaust gases through the bypass stack should the primary and secondary dampers be closed.

A software program was used to simulate change over scenarios which, if carried out, may have been too dangerous and problematic. These included instances such as change over involving the malfunction of the fresh air firing supply – where the boiler flame is supplied exclusively by the GT’s rundown exhaust gas.

 Computational analysis was employed to generate HRSG steam parameters during change over. These simulated data were compared with actual calculations (Figure 3). The computer-generated time dependence of the steam flow shows a good correlation with the measured data. There are some discrepancies between the individual calculated and measured data points owing to an insufficient evaluation process, but the two time dependence curves show a close correlation. Austrian Energy aims to carry out further investigations to determine the efficiency and feasibility of using predictive software programs in other projects.

Following successful auomatic change overs (Figure 4) during the commissioning phase, and detailed checks by the independent technical safety company, TÜV Österreich, Voestalpine’s Linz plant went into commercial operation in January 2006 with the new change over procedure installed. Since then, approximately 15 change overs have been carried out successfully, all of them following planned GT shutdowns. Only one change over has so far failed, but this was attributed to vibrations in the boiler burner, which destroyed damper switches, and not to the automatic change over procedure.

Figure 1. Arrangement of the boiler at block 1 of the Voestalpine steel mill

1. Boiler furnace
2. Burner
3. Bypass stack
4. Diverter damper
5. GT exhaust gas Figure 2. Schematic showing damper configurations Figure 3. Comparison of measured and calculated live steam flow data Figure 4. Successful testing of the automatic change over