Document K60Kxd0MKdvyeaDZQY8GKdDLX

1 During winter teuton condtnaor octt et Heot exchongor to produco hot water lor district heoting. In summer It functions as condenser using circulating water How We Cjired Condenser Tube Corrosion Sudden tube leakage on a special heater-eondensor demanded quick solution to avert catastrophe. Here is an account of how some simple methods pinned down trouble and controlled it By H WIIDINGH, Orfortte Lighting Dept, Oenmerlr A hew 3500-kw steam turbine storied in January 1947 at the Odense Power Station. This plant supplies both electric energy and hot water to a dis trict heating system. Turbine discharges to a special condenser with 5400 aq (t heating surface. Steel tube plates hold aluminum-brass tubes (77% Co, 20.2% Zn, 2.7% Al) in metallic pack ings. During the heoting season the con denser cooling water circulates through the district healing system. Circulating water passes through Ihe condenser at the rate of 1,400,000 lb per hr under a pressure o( 70 psig. Condensing steuin raises water temperature from 122 to 175 K. The vacuum, naturally, is quite low. During the summer, the turbine runs as an ordinary condensing unit, using river water for cooling In the condenser. This requires two sets of circulating-, water connections, one for ihe river water end tits other for the district heal ing system. At full condensing the tur bine develops 4000 kw, hut its steam role is poorer than oilier units in the station--hence during the summer it Is used mostly for standby. Trouble. In the summer of 1949, con denser tubes storied leaking, and 113 had to he renewed. These were 5% of the total 2250 tubes. Chemical analyses and microscopic examination showed the materia) was good. Fig. 3. The corrosion involved sine removal from the alloy. A thin scale coated the tube interiors, caused by sewage pollu tion of the river water.' The tine re moval might have been earned by (I) the scale (2) district-heating water (3) stray-current electrolysis. Electrolysis. A 2-wire 220-v dc system supplies the central part of Odense." In addition, a 500-v dc system feeds the tramway. Uninsulated copper cables serve as neutrals for the dc grid and feeders. In the station, the neutrals are grounded, while foeders to the tramway are insulaicd and not grounded. Fig. I shows, schematically, the 500-v ayatem and the turbine Installation. The district-heating system consists of six miles of 16-in. mains. Feeder and return pipes lie side by aide in a tunnel of reinforced concrete. '`Cellular1' con crete fills the space between pipe and tunnel walls. Since the tunnel is not quite waterproof, it is placed above ground-water level. Measuring voltage drop between cer tain points in the power station started the search fur possible electrolytic cor rosion cause. Measuring points ineluded outgoing and return heating pipes, cooling-water pipes, backpressure turbine, condenser, other turbines, boil ers, steam and water pipes, water tanks. The portable millivoltmeter showed some of these points were negative and some positive, relative to the dc neutrals and tramway systems. Readings were taken with (1) tur bine-condenser heat Ing circulating water for distribution (2) turbine-condenser using river water (3) turbine out of commission and district-heating pipes disconnected. These showed potential differences (j>d) at all points, (1) and (2) in particular had the highest pd with one exception, a water pipe in the ash cellar. Apart from the latter, the pds were oil leas than 200 millivolts. On the district-heating pipes there 2 Electrical potential difference between condenser tube day exocily the some as the electrical load. Voltage drop plore ond stotien dc neulrol cable varied throughout the reoched a maximum during peok loads in morning and ofternoon 78 ENGINEERING AND MANAGEMENT SECTION POWER 5 Pd l " d,rK"n,110; * "d`n'" w" `"""T * *" , M diitrlct-henlinn pip- & ' . ,,,,, ,h, current in the dc neutrals . Apsrto I' current left th. bs.c ncutrsls td ur.H through the ground to tho d x'*' _ J ilirAitirli these 10 t Microphotographs Gave Clue to Cause of Corrosion Trouble \ J 5 j I 5' i '> i i. "j i ' Annlytlt. These dste suggested the lesion '(or tho sudden tube corrnelon during the (all if 1W. During the sum mer ol IMS the turbine rnn lor many hours, with riser witter flowing through Ihe condenser aggroveling the CondiUon. The running hours exceeded thoee for the two preceding summer* com- bineo. ,, In Fig. 1, A, D and C arc the pd measuring points. A is about 100 yards from C. The pds In millivolts for the three operating conditions were: A-0 A-C (1) Kesiing operation 100 (2) Condensing operation 100 (3) Disconnected pipes -- 38 35 95 We measured the voltoge variation .for-several days from A to B with a recording millivoltmeter. Records were made with the, turbine in service for heating and condensing generation. In both cases the voltage drop varied in the same pattern. Fig. 2 shows that the pd varies almost proportionally with the dc load. It's a maximum of 0.4 volts at 8 am snd 4 pm. The lowest values occur from midnight to 5 am. Cure. The condenser tubes were con stantly acting as an anode, hence, be came badly corroded. The remedy waa to provide a path for these currents around the condenser to the de neutral. We'did this in 1950 by connecting tho condenur tube plate to the dc neutral with a copper fagr 60 ft long, having a drosa section 30x5 aq mm and a re finance of 0.002 ohm. A 0.0002-ohm shunt in the bar permitted taking am meter .readings of current flow. Ammsler readings hove been token at different hours of the day for the last couple years. Current in the bar haa slwayi been from the condenser plote to the neutral. The minimum current was *ero and the maximum 85 amp. Normally it varies .fpmt 25 to 50 amp throughout a day. . Since this connection was installed, here .has not been a. single failure in condenser tube, even though the tubes were badly corroded at the time conMetion was made and the turbine run- .8 * high number of hour* dur`,nB summer. *,'8' 4 shows microphotographs of a 3M*crophotogropbs of corroded tube ot 24x ond 180x enlargement showed the trouble was line removal from olloy. Tubes hod thin scale cooling on inferior 4 Microphotographs of portly corroded tube ot 24x ond 180x enlargement ofter installation ot clectrieol bypoii showed the rote of westing wos greatly reduced tube section Installed in tbe condenser in the summer of 1949. It was exam ined in 1950. This lube had only 6 months' service before the copper bar connection was made. Although it had not failed ot time of removal from con denser, the corrosion haa penetrated about two-thirds of the wall thickness. Zinc removal is involved in this cor rosion too. We expect the tube corrosion will con tinue despite the electrical connection. The deterioration rate, howover, will be very much slower, and chemical in nature rather than electrical. Since no further tube failures have occurred, it would aeem that condenser conditions are now under control to the maximum possible extent. Without the electrical bus, it would have been necessary to replace all con denser tubes in a very short time. The prospective life of the replacement tubes would have been much too short. ; 1 JANUARY 1953 ENGINEERING AND MANAGEMENT SECTION 79