Incorporate lessons learned into specifications for new units

By Brent Cosgrove, PE, HRST Inc

Next to the gas turbine (GT), the heat-recovery steam generator (HRSG) has the greatest impact on combinedcycle system performance. Failure to address industry lessons learned in the specification for a new unit can put your plant at a competitive disadvantage. In the unforgiving world of merchant power generation, this can mean a balance sheet with red ink.

For example, if your new boiler is susceptible to fouling and you have not provided a means for dealing with it, heat rate will increase—and so will gas consumption. Because fuel accounts for more than 80% of the life-cycle cost of a typical combined cycle, even a relatively slight increase in fuel use can have a significant impact on the price of power.

Additionally, if your plant is required to cycle—as most are today—and the HRSG is not designed for that service, the risk of unplanned downtime to correct avoidable damage is significant. The cost of a forced outage and necessary repairs can turn a good year into one with marginal profit.

When developing the specifications for a new plant, there is a temptation to reuse the specs from the last project. This would save time and money, but it may not be the best idea in the long run. Regarding the HRSG, there’s a good possibility that an existing spec is outdated, at least in part, and should be revised to incorporate lessons learned.

In the extremely competitive market for powerplant equipment and services, you will get from the EPC contractor and HRSG manufacturer only what you specify. Any deviation from the spec to address a legitimate concern will increase plant cost (sidebar). The intent of this article is to bring to the attention of engineers responsible for preparing and evaluating HRSG specs some historically overlooked issues.

It is important to specify a unit that will be delivered both fit-for-use and designed for long-term performance. Incorporating lessons learned that make sense for your particular situation will increase the purchase price over that for a standard offering. But the payback is in a lower lifecycle cost. Also important is to ensure that bids are proposed on a level playing field so they can be compared and evaluated objectively.

What follows are a dozen suggestions that can help you improve HRSG reliability and performance over the life of the unit. Problems identified at public meetings, such as those conducted by the HRSG User’s Group (, generally are preventable with a proper specification and a comprehensive design review before OEM drawings are approved.

1. Flow-accelerated corrosion (FAC). Even assuming high-quality waterchemistry management, it is important to take additional precautions to minimize the effects of FAC (Fig 1). This suggests a need to assess FAC risk on the components of the lowand intermediate-pressure (LP, IP) evaporators and all economizers

Pipe and tube velocities must be specified in conformance with best practices. If these velocities cannot be maintained within the guidelines for carbon steel, FAC prevention requires use of low-alloy material (P11/T11 or P22/T22) for tubes, headers, and piping. Regardless of the tube material, industry experience suggests use of piping of generous diameter and an increased number of evaporator risers to minimize velocities.

Straight tubes have advantages over bent ones, and the design of steam-drum nozzle connections also should be examined carefully. An independent expert’s review of the bidder’s circulation analysis will help you sleep more soundly.

2. Desuperheater arrangement. When not properly specified, operational and reliability issues with the desuperheater can cause expensive, progressive damage to downstream piping and superheaters. It is important for your desuperheater supplier to consider the full range of operating conditions, and for you to make sure that the design team and powerplant operators adhere to all of the vendor’s recommendations and requirements.

To be certain that the spray water evaporates completely, the recommendations should include maximum velocity, minimum straight-length both upstream and downstream of the desuperheater, and a minimum distance to the downstream temperature element (Fig 2). For added protection, consider requiring that the downstream straight-length be increased to a value above that recommended by the vendor.

Note, too, that unless precautions are taken, a desuperheater with a conventional probe is susceptible to fatigue failure (Fig 3). To prevent the thermal cycling that causes such failure, the supplier should specify a minimum spray-water flow rate. Otherwise, if the probe cracks, a flood of unatomized water can lead to catastrophic failure of the thermal sleeve, the steam piping, or downstream superheater (Fig 4).

Lastly, to avoid any issues associated with overspray, consider specifying that if the supplier has a problem maintaining a minimum of 50 deg F of superheat, it will be a make-good warranty repair item.

Applying industry lessons learned to Northland Power’s Thorold project

The power professionals at Toronto-based Northland Power Inc know their business. If they didn’t, the independent power producer (IPP) would not have a 20-yr success story in the financing, development, engineering, construction, operation, and maintenance of highly reliable gas-turbine-based, biomass, and wind power projects.

Northland’s staff, more than 150 strong, manages and operates five generating facilities in Canada totaling more than 470 MW of electricity and another 100 MW of thermal energy. In addition, it is building a 265-MW combined heat and power (CHP) plant in Thorold, Ont, on the site of an Abitibi-Consolidated Inc pulp and paper mill, and developing a 150-MW wind farm with 100 machines.

Four of Northland’s five operating plants rely on gas turbines (GTs) for at least part of their output. The durable LM2500 matches up well with demand in a couple of locations and two LM6000PDs power a 126-MW combined-cycle cogeneration plant at another Abitibi mill. The latter was Canada’s most efficient facility of its type when it began operation in fall 1996. Northland also has an active interest in other generating plants equipped with a 6FA and 7EAs.

For Northland, Thorold CoGen LP represents a step up into the world of large frame engines and triple-pressure reheat heat-recovery steam generators (HRSGs). The 1 × 1 combined-cycle, on track for a late 2009/early 2010 start, is powered by a nominal 170-MW GE Energy (Atlanta) 7FA. The 95-MW Siemens Power Generation (Orlando) steam turbine/ generator features an SST-700 highpressure cylinder and an SST-900 low-pressure module. For details on this machine, access www.combinedcyclejournal. com/archives.html, click 4Q/2006, click “Walnut” on the cover (p 67).

The IPP approached Thorold as it had previous projects. For example, Northland’s engineering group, managed by Dino Gliosca, prepared a functional specification for the HRSG. It gave the EPC contractor qualitative guidelines for procurement— such as daily cycling service. Critical items of concern—such as flow-accelerated corrosion (FAC)— were noted.

But as anyone who has been involved in a project of this magnitude knows, schedule and cost are paramount in negotiations and once the EPC contract is signed things move forward quickly. The response to some specific concerns of the owner/operator often is “best industry practice,” to keep the project on track. That’s the way the industry works; so you wait for vendors to respond to the RFQ (request for quote) to see what they’re offering.

The OEMs are challenged as well. They have had to “standardize,” to the degree possible, on HRSG component designs. In the unforgiving world of “price/schedule, price/schedule, price/schedule,” if you’re not competitive, you’re toast. And any changes to what the OEM is offering at a fixed price can be expensive.

Along with Project Manager Greg Lennox, Project Engineer Paul Kaminski and the Northland engineering department did their homework and knew of the potential problem areas associated with F-class HRSGs from conversations with colleagues, reading the HRSG Users Handbook (for more information, visit, etc.

As experienced powerplant developers and operators, Northland’s engineers also knew you never were going to get all the features/enhancements you’d like on any project. Not having direct F-class HRSG experience, Gliosca and his team called in HRST Inc, Eden Prairie, Minn, to review the boiler proposals and to help the IPP decide on upgrades worth paying for. HRST, which was familiar with the designs submitted by all vendors bidding, specializes in developing solutions for HRSGs that don’t behave the way the owners would like.

After reviewing HRST’s vendor evaluations and reflecting on Northland’s experience, Gliosca, Lennox, Kaminski, and others in the organization recommended the following enhancements, among others:

3. Drain-system design. Drains should be treated as a significant, functional part of the HRSG, not an afterthought. All drain piping should be designed to have a downward slope when the unit is hot. Stagnant water retained at low points can lead to corrosion problems. Specify that the blowdown tank be located below grade to allow every portion of the HRSG to drain completely by gravity alone; otherwise, an additional drain trough and sump may be required.

Drains from different rows (harps or panels) of superheaters, reheaters, and economizers should remain separated until after the first isolation valve; upstream and downstream harps should never be connected through drain piping during operation (Fig 5). Drains from different pressure levels must never be connected. Continuous and intermittent blowdowns and superheater drains all should have separate connections to the blowdown tank.

In addition, specify that superheater drains be designed with a capacity to handle worst-case condensate flow plus some extra margin. Even one occurrence of condensate back-up into an operating superheater can be very damaging. If the drain system scope is split between the HRSG supplier and a balance-ofplant contractor, take steps to ensure they are properly coordinated. If the supplier provides a recommended drain layout, the contractor should be compelled to follow it.

4. Economizer shock resistance.

Ensure that multi-pass economizers are designed for low-stress pass-topass temperature differences under all operating conditions; additionally, that the economizer design considers velocity criteria to maintain continuous flow stability. Under certain conditions, failure to do so can result in random vapor-locked tubes and flow reversals, which ultimately lead to thermal fatigue and stress cracking (Fig 6).

Header-to-tube connections should be designed to avoid creating stress concentrations. Weld quality must be monitored diligently, especially on bent tubes. Large tube-to-tube temperature differences (that is, thermal shocks) must be avoided under all operating conditions.

When the temperature difference between adjacent tubes becomes excessive, the tube that is cooler will want to contract; however, it will be constrained from doing so by the significantly more rigid headers, risking the possibility of stretching and/ or tensile failure. This is presented graphically in an earlier article by HRST Inc; “Avoid desuperheater problems. . .”.

Designs should either be capable of withstanding thermal shocks or systems and controls should be in place to prevent them altogether. This is particularly important for cycling units, because the most probable time for a thermal shock event is during startup.

5. Size of IP steam drum. For triplepressure HRSGs with reheat common in F-class combined cycles, the industry has learned that traditional sizing techniques can result in IP steam drums that are too small to handle large pressure excursions in the reheat system—following a steam-turbine trip, for example. In some cases, drums sized according to the usual practice of fulfilling requirements on retention time and surge volume have proved incapable of coping with large reheat pressure swings.

One may well believe that a drum sized correctly for surge volume (swell) should be able to survive any upset condition. Nonetheless, with the dynamics at play and given the relative size difference between the reheat and IP systems, even the most diligent tuning appears unable to prevent all trips.

One approach to mitigating this issue is to provide an extra margin of safety on IP drum volume by adding a foot to the diameter. Another proven method is to supply a backpressure control valve on the IP steam outlet upstream of the connection to cold-reheat. Prevention of just one plant trip should easily pay for either of these solutions.

6. Flow manipulator arrangement. The turning-vane and/or perforatedplate design should be reviewed and rigorously examined by a third-party expert. Buckling caused by restrained expansion and cracking from uneven heating are industry-wide concerns that are difficult and expensive to resolve (Fig 7). As one would expect, these problems are exacerbated by cycling service.

7. Firing-duct arrangement. For HRSGs with a duct burner, it is essential for operators to visually monitor flame condition. This requires that burner view ports be provided and that they allow visual access to all areas of the burner and firing duct—including all burner components, sidewall, floor and roof liners, and boiler panels immediately downstream of the burner.

Specify radiation shields for firingduct thermocouples (TCs) to prevent erroneously high readings. Important, too, is to specify a sufficient number of TCs, as well as an insertion length, to provide a representative picture of actual conditions inside the duct. Lastly, provide ample space between the TCs and heat-transfer surface because radiant cooling can result in low readings.

Note that peak firing temperature can be hundreds of degrees hotter than the bulk temperature. Thus the firing-duct liner material selection ( Fig 8 ) should be based on estimated peak temperature plus radiation effects (that is, significant margin should be applied to the bulk firing temperature).

8. Foundation arrangement. Specify that the HRSG foundation be allowed unrestricted thermal growth; otherwise, the foundation and/or the casing can be damaged as early as the first heat-up. Corrective action is much more difficult after installation.

9. Gas-side accessibility. To facilitate timely and effective inspections, cleaning, and repairs, the spec should provide for a means of entry to all gas-side areas of the unit, including the crawl spaces above and below the tube headers. All access lanes should have sufficient clearance to allow scaffold erection from floor to ceiling. It also is advantageous to have headers equipped with borescope ports (1-in.-diam weldolet connections) to facilitate internal inspections from inside the crawl spaces.

10. Tube-cleaning accessibility. Be sure to specify access to cold-end tubes for the removal of ammonium salts and/or corrosion products. A recommended practice is to restrict the number of rows between cleaning lanes to no more than about 10. Lanes must be sufficiently wide to manipulate equipment; specify a minimum fin tip-to-fin tip clearance to ensure effective cleaning.

11. Shutdown protection. A shut down HRSG can be more vulnerable to deterioration than an operating one, making it important to specify cycling units with features to lock-in heat and lock-out oxygen and moisture. Include a damper or balloon to seal the stack, and insulate the stack up to the height of the shutoff device. Specify as well a nitrogen blanketing system and, especially in cold climates, a well-engineered steam sparging system. Another recommendation: Request that the floor penetration seal be of a design that minimizes corrosion of the piping within.

12. Performance testing. Occasionally, an owner errs by trying to specify an unattainable level of uncertainty for HRSG performance tests. It is important to understand that zero uncertainty is impossible and that extraordinary measures are required to achieve very low levels—such as 0.75%-1%. To simplify performance testing and to prevent haggling over procedures, responsibilities, and results, it may be preferable to specify a more reasonable test uncertainty— for instance 1.5%-1.75%.

Keep in mind that performancetest uncertainty is a measure of the quality of the test, not the performance or quality of the HRSG. To help minimize uncertainty, minimize the number of influences on its determination (for example, no duct firing, fuel heating, or blowdown during the test). It is reasonable to specify an allowable tolerance from a contractual perspective and leave it up to the supplier to determine how to achieve that. However, to prevent the possibility of surprises later on, it is advisable to have confidence in the manufacturer’s strategy from the beginning. ccj