Repowering Florida

The utility industry has a substantial number of older generating stations, with a 2400 psi steam cycle. Repowering these older stations with a combined cycle, combustion turbine and heat recovery steam generator (HRSG) results in a highly efficient and environmentally friendly power plant. Florida Power & Light (FPL) is using this combination for repowering projects at two sites, Fort Myers and Sanford. In both projects, existing plant, land and equipment is being utilised to improve project economics.

The environmental benefit to the community of these repowerings includes substantial reductions in NOx, PM, SOx , CO, CO2, and solid waste (ash), but with a slight increase in VOC emissions.

In the Fort Myers case the HRSG is designed to produce main steam output conditions of 1050 °F at 2400 psig, a first for HRSG technology.

Fort Myers is located in southwest Florida, which is growing 40 per cent faster than most other areas of the state, while Sanford is in central Florida, also a high-growth region.

The Fort Myers project

The Fort Myers repowering project will expand generation capacity from around 550 MW, based on 1950s era heavy-oil-fired steam generators, to some 1550 MW, based on natural gas-fired, combined cycle technology. Six combined cycle units will replace the two existing heavy oil-fired steam generators. Natural gas will be piped to the site by underground pipeline, eliminating heavy fuel oil deliveries by barge and tanker, via Boca Grande up the Caloosahatchee River.

The new plant configuration, which includes six combustion turbines (GE Frame 7FA) and six Foster Wheeler HRSGs, represents the current state of the art in both cycle efficiency and HRSG design. The project represents a $450 million power plant upgrade phased for completion in time to meet the 2002 summer peak load.

January this year saw hydrotesting of the first of the new HRSGs and the remaining units are being hydrotested at the rate of one per month until June. Each HRSG is scheduled to be mechanically completed around one month after the hydrotest. Work on the balance of plant will start in mid-July and in late August the existing oil-fired units will be shut down and conversion work will start on the steam turbines. Commercial operation in full combined cycle mode is scheduled for the end of May 2002.

Meanwhile, all six of the new gas turbines at Fort Myers are now operating in simple cycle mode. This is possible because the Fort Myers units are equipped with bypass stacks. The first of the Fort Myers gas turbines started up in autumn 2000 and the sixth in early May.

FPL is planning to further augment generating capacity at Fort Myers with two additional 170 MW natural-gas-fired gas turbines operating in simple cycle mode, for peaking duty. These are due to come on line in April/May 2003.

The Sanford project

The repowering project for the Sanford plant will include installation of eight GE Frame 7FA gas turbines with eight Foster Wheeler HRSGs. These will operate in combined cycle mode, making use of the two existing steam turbines.

  As at Fort Myers, the new combined cycle units will run on gas rather than oil, resulting in substantially lower emissions and eliminating the need to barge oil from Jacksonville along the St Johns River. The generating capacity at the site will rise from about 950 MW of 1970s-vintage oil-fired plant to some 2050 MW of gas-fired combined cycle capacity.

Pressures at Sanford are similar to Fort Myers, but the final steam temperature is slightly lower, 1005 °F instead of 1050 °F. The first of the eight new HRSGs is due for hydrotesting in June and the remaining seven will be hydrotested at monthly intervals thereafter. The aim is to also have Sanford in full commercial combined cycle operation in early 2003. All eight of the new gas turbines for Sanford have been delivered to site, but cannot be operated in simple cycle mode because there are no bypass stacks.

HRSG design

Designing and manufacturing HRSGs for reliable steam supply to Fort Myers’ 2400 psig turbines was particularly challenging, requiring extensive natural circulation analysis, dynamic thermal transient analysis, and fatigue life investigation.

The Fort Myers design consists of three pressure levels. The side section elevation of the HRSG is presented in the diagram, right.

The HRSG cools the GE Frame 7FA combustion turbine exhaust gas from 1125° F to 211° F. LP economiser feedwater is regulated to 140° F by recirculating a portion of the outlet LP economiser feedwater back to the economiser inlet. Final steam conditions and flows for the design case are as follows:

-------------Temp--Pressure--Flow

-------------(°F)--(psia)--(lb/h)

HP outlet--1054--2486--407600

RH outlet--1052--555--474000

LP outlet--612--86--54000

The Fort Myers plant is designed as a “wet bypass” unit. During start-up and following a steam turbine trip, the bypass system routes HP steam through a pressure reducing valve and attemperation station to the reheater. At the reheater exit, steam is then either vented to atmosphere – in the case where the condenser is not on-line – or is routed through a hot reheat bypass pressure reduction valve and attemperation station to the condenser steam dump. This system provides operating flexibility in accommodating plant upset conditions and also reduces the thermal transients at the reheat headers.

A key plant design criterion was the maximum allowable instantaneous loss in MW output due to a steam turbine trip. As already noted, the gross output of the station is 1550 MW. FPL stipulated that the plant be designed to lose no more than 910 MW over a 30-minute period. The 30-minute period corresponds to the time necessary to reconfigure the grid generation capacity to ensure no loss of power to FPL customers. This is accomplished by running the combustion turbines with the HRSGs dumping steam to the condenser through the bypass system and/or through vents to atmosphere.

The air permit for the combustion turbines stipulates the maximum start-up duration to reach a load of 55 per cent for both a cold and a hot start. Once the combustion turbine reaches a load of 55 per cent, emissions will be within compliance. The stresses induced during the start-up of an HRSG are inversely related to the duration of the start-up. A shorter start-up duration for the combustion turbine leads to higher stresses in the HRSG, which leads to shorter component life. Foster Wheeler, Black & Veatch, and FPL jointly developed the start-up approaches and system designs to ensure a reliable plant for the 30-year design life

Dynamic thermal transient analysis

The determination of anticipated life and clarification of operating constraints necessary to enhance the life of the HRSG were critical to success in the 2400 psig repowering effort. Typically combined cycle plants are dispatched quickly, thus leading to thermal transients that exceed those seen in traditional utility-type boilers. The design of HRSGs includes many features to limit stresses resulting from these thermal transients. However, as the boiler operating pressure increases, the component thickness increases, and stresses from thermal transients increase.

The first task in determining the fatigue life of the HRSG was to define the anticipated thermal transients. A dynamic simulation of the HRSG operation during the worst condition of a cold HRSG start-up was conducted to determine HRSG outlet conditions and perform cyclic structural analysis of critical components. This work was divided into the following three subtasks:

• Overall system simulation including HRSG, combustion turbine, and steam turbine.

• High-pressure evaporator detailed model simulation. And

• Fatigue life analysis.

A ProTRAX® dynamic model of the Fort Myers HRSG was constructed including representations of all heat transfer surfaces, feedwater systems and the combustion and steam turbine. The model has the capability to run start-up, shutdown and load-change transients, explore the system response to upset conditions, and evaluate full- and part-load process performances. The model also allows tuning of control elements, such as three-element drum level controls, and can easily demonstrate the effect of transient operation on drum level or the consequences of using various vent-valve characteristics.

The ProTRAX model was used to run a number of transient start-up and shutdown conditions. It was used to identify temperature ramp rates for critical headers in the superheater and the reheater and to assess drum temperature ramp rates both with and without venting.

The input boundary condition to the model consists of the combustion turbine from the point of firing through to spinning reserve conditions at 7.5 per cent load. The combustion turbine is held at this load for the first 80 minutes of operation, after which the turbine is ramped over the next 30 minutes to the 55 per cent load condition. The highest ramp rate for the HRSG HP drum fluid temperature occurs from minute 30 to 70 in the start-up process.

During the start-up process, high-pressure steam is bypassed through the reheater either to atmosphere or to the condenser dump. The vent flow is dependent on pressure, and hence, the initial ramp from ambient to first steam generation (212°F) is dependent on heat input, mass of metal, mass of water, and heat transfer surface. Following initial steam generation, pressure is developed in the high-pressure drum in accordance with the venting and bypass capacity of the system. The unit will stabilise at a pressure where the bypass/vent flow balances the steam generation. At this point, a 30-minute hold is instituted to allow the drum metal temperature to approach equilibrium prior to further heating. Following the hold period, the combustion turbine is ramped to full load at a controlled rate, and the rate of pressure rise in the high-pressure drum is controlled to minimise the thermal stresses.

The ProTRAX model was used to predict overall dynamic thermal performance and assess impact of various start-up scenarios, including vent flow, initial drum level, operation of intermittent drains, and combustion turbine start-up constraints.

In conjunction with the ProTRAX investigation and as a result of the preliminary findings in the fatigue life investigation, examination of the high-pressure evaporator section was refined. This was accomplished using a proprietary Foster Wheeler code THTNET (Thermal Hydraulic Transient Network Analyzer). THTNET performs one-dimensional transient thermal hydraulic analysis of a flow heat transfer network through finite difference techniques. The program applies finite difference techniques to solve one-dimensional transient conservation equations of energy, mass, and momentum.

The application of this methodology allowed examination of temperature gradients on a tube-by-tube basis throughout the bundle, and identification of detailed thermal gradients that exist at the nozzle attachments to the high-pressure drum. The main objective of the use of the THTNET program was to accurately predict thermal gradients for use in the life-cycle fatigue analysis investigation. A secondary objective was to determine the sensitivity of the drum thermal gradients to drum water level.

Fort Myers life-cycle fatigue analysis

The Fort Myers HRSGs will operate under cyclic conditions during start-up, shutdown and load changes. These conditions will induce thermal stress particularly where transitions occur in thick-walled pressure part components. The thermal stress events lead to fatigue damage.

An extensive investigation of the fatigue life was carried out based upon ASME Section VIII, Division 2 criteria and finite element stress analysis. The objective of this analysis was to provide feedback to the customer regarding the impact of combustion turbine start-up scenarios and to provide parametric curves showing the cycle life for different ramp rates and temperature differentials.

The analysis procedure was based upon the general purpose program, ALGOR®. The temperature and thermal stress calculations obtained by the finite element analysis are post processed for determining the fatigue life of the selected components. The pressure stresses are added to the thermal stresses. The Mises equivalent of the total stresses is used in the fatigue calculations. A stress concentration factor of 2 is assumed to account for local effects.

A screening analysis of the headers and drums determined the most critical in terms of fatigue during a start-up transient. The components consisted of headers and high-pressure drums and their connection to thin-walled tubing and piping. Components analysed included superheater and reheater outlet headers and manifolds, high-pressure drum, downcomers, and risers. The critical area was identified as the riser connection to the high-pressure steam drum. The consequences of drum differential temperatures were also examined with respect to displacement of the downcomer pipes due to drum “humping.” This was found to be a secondary concern, however the investigation led to refinement in the design for evaporator feeders from downcomers.

The end result from the investigation into cycle life is presented in the figure left. This shows the relationship between life expenditure per hundred cycles, ramp rate in °F/h and the total temperature change in °F. This set of curves is based upon ASME Section VIII, Division 2 rules that include a safety factor of 20 on fatigue life and a design margin of 4 on ultimate tensile strength. The total temperature change is additive.

As an example, consider the following cycle. The heat recovery steam generator is started from a cold condition at 80 °F to a final saturation temperature in the high-pressure evaporator of 680 °F. The unit must undergo a total temperature change of 600 °F. The first 400 °F of that change is completed at a ramp rate of 500°F/h. The life expenditure reading is 0.015/100 = 0.00015.

The next 200 degrees temperature change is done at a ramp rate of 200 °F/h for a life expenditure of approximately 0.002/100 = 0.00002.

Total life expenditure for the cycle is therefore 0.00002 + 0.00015 = 0.00017.

Conversely, it would be anticipated that the unit could withstand 1/0.00017 = 5882 cycles. Note that, in practice, a much greater service life than this prediction would be anticipated due to the safety factors inherent in ASME Section VIII, Division 2 analysis techniques.