Cost effective retrofit restores LP heaters at Maanshan20 August 2001
An innovative and highly economic retrofit has restored original operating parameters in an LP heater at Taiwan Power’s Maanshan 951 MWe, unit 1. Deteriorating operation had been the cause of tube vibration and leakage, contributing to a decline in overall plant performance. Instead of replacing the entire tube bundle, a retrofit strategy was developed which was dramatically cheaper. Adam H C Shih, Taiwan Power Company, Maanshan nuclear power station, Taiwan; Abraham L Yarden and Jay Wu, Thermal Engineering International, Los Angeles, California, USA
When the Taiwan Power Company’s Maanshan unit 1, a 951 MWe pressurised water reactor (Figure 1, above), was commissioned in 1984, it was equipped with two strings of LP feedwater heaters, one string in each of the two condenser necks. This configuration saved valuable space, but had no thermal effect on feedwater heater operation.
Four of the low-pressure feedwater heaters (numbers 3 to 6) in the necks of the two parallel condensers (designated A and B) operate in series. These straight-tube closed heaters use LP turbine exhaust extraction steam in sequential pressure–temperature shell-side stages to heat tube-side feedwater. All eight LP feedwater heaters employ shell expansion joints to eliminate thermal stress differentials between the carbon-steel shells and the stainless-steel tubing.
The 4A feedwater heater in this arrangement has a tubesheet face-to-face length of 52 ft 6 in (16 m) and a diameter of about 4 ft 8 in (1.8 m). Operating conditions before the retrofit were:
Feedwater – Inlet 229°F (109°C), outlet 294°F (313°C).
Extraction – 311 000lb/h (141 t/h) at 69 psia (4.7 bara) and 318°F (158°C).
Condensate – Saturated 307°F (153°C), subcooled 241°F (116°C).
All of the LP feedwater heaters in each string are equipped with sub-cooling zones, Figure 2, designed to prevent the cascading of saturated condensate to the next lower-pressure heater. This saturated condensate would quickly become erosive 2-phase flow and damage condensate piping in the next LP heater.
In each case, these 15 ft 4 in (4.5 m) sub-cooling zones are baffled to increase the condensate sub-cooling flow path, Figures 3a and 3b, to increase sub-cooling effectiveness. This sub-cooling zone isolates the incoming feedwater tubing from the remaining feedwater heating areas, thus sub-cooling the condensate to about 66°F (36°C) below the saturation temperature corresponding to the extraction pressure within the heater. It is isolated from the condensed extraction steam that collects in the bottom of the heater shell by a loop seal permitting condensate to fill the sub-cooling section up to the top row of the incoming feedwater tubing.
Over its 15 years of operation, a lower-than-normal saturated condensate level in the 4A LP heater caused an excessive ?P across the 11/4 in (32 mm) thick end plate which separates the liquid-filled condensate sub-cooling zone from the extraction-steam condensing area. Since this adverse ?P is greatest near the top of the sub-cooling zone, condensation of a small amount of extraction steam flowing through the circumferential clearance between the end plate and the upper row of tubes did not occur as designed. Rather, a two-phase flow through this clearance exited into the sub-cooled condensate zone. This 2-phase steam flow, especially around the top few tube rows, into the condensate sub-cooling zone created excessive tube vibration which caused tube failures in this area, Figure 4. These failures, of course, further reduced heater performance with an adverse effect on overall plant heat rate. Also, during operation the leakage adversely affected plant operation.
As a result, the sub-cooled condensate temperature increased to 241°F (116°C), and the 11/4 in (32 mm) sub-cooling zone end plate became warped due to the high ?P across it.
Instead of replacing the entire tube bundle, it was decided to take a more economic approach. This involved replacing the drain-cooler end plate and all the internal parts of the drain sub-cooling zone inside the condenser neck.
During the evaluation process, it was recognised that this approach would cost only about one-third of the cost of replacing the entire tube bundle. This dramatic cost difference was due to the constricted space involved and the typically limited accessibility around the condenser neck. Therefore the majority of the cost of entire tube-bundle replacement would have been in its installation.
In order physically to get into the sub-cooling section of the 4A feedwater heater near the damaged end plate, it was necessary to cut a large “window” into the heater shell for access purposes, Figure 5. All inlet feedwater tubes were removed along with the bottom three rows of return tubes, allowing removal and replacement of the end plate, baffles, and other parts within the drain-cooling zone.
A new 3 in (75 mm) end plate was installed, replacing the original 11/4 in (32 mm) plate. The added thickness lengthens the extraction-steam path to ensure that the steam flowing through these circumferential areas fully condenses before entering the sub-cooling zone. This eliminated the possibility of the previous two-phase flow causing tube vibration damage and ultimately breakage.
This retrofit not only improved thermal performance (drain cooling approach improved from 11.9°F (6.6°C) to 5.2°F (2.9°C) by replacing the plugged tubes and optimising the liquid-level setting), it also increased longevity of the feedwater heater.
Now that feedwater heater 4A has been successfully retrofitted, work is underway to similarly retrofit the No 5 LP heater in Maanshan unit 2.