eeeoecoeorvveeceoeeeeveeveeouved KK Plaka Maren, Ap; WDE DE, Coe [Reprinted from Proceedings Forty-seventh Research Session of the CBIP, Hubli-Dharwar. Karnataka, November 1978, Vol. 1V] 7 Elimination of Surge Tanks—A Critical Study J.P. Gupta ‘Superintending Engineer H.R. Sian” Executive Engineer Design Organisation, Irrigation Department, Roorkee, (U.P) SYNOPSIS Umil recently surge tank was considered an integral part of hydro-electric schemes having long pressure conduits. of such hydro power stati possibilities of reducing the eliminating the surge tanks. deciding the necessity or otherwise of a surge tank. The successful performance of a number fons with no surge tanks have, however, indicated the ‘cost and period of construction of these schemes by The present paper discusses the criterion for The likely adverse effects on the water conductor system and on the power system as a result of surge tank climination and their remedial measures have also been discussed. A few exam- ples have been given where surge affecting the stability of the systein. system for the proposed Tehrl undergroun city of 1600 MW indicates the possibility of dispensing with the surge tanks. Detaijed studies are in progress to arrive at a final conclusion. ). Introduction 0.1 A surge tank is an artificial reservoir introduced along the pressure-conduit system at a suitable point upstream and‘or downstream of a hydro power plant. Depending upon the topography of the work, area this may be designed as a chamber excavated in the mountain or as a tower Tising high above the sur- rounding terrain. ‘The surge tank has three primary functions : (A) to provide a free reservoir surface close to the terminal discharge mechanism as a q ree of compensating water hammer reflections 10 limit the pressures in the penstock and materially reduce them inthe main conduit, the initial reflec. tions being negative in the case of closure and positive in the case of opening, (B) to supply the Additional water required by the turbine during load demand until the conduit velocity has accelerated to the new steady state value ; and (C) to store water during load rejection until tHe conduit velocity has been decelerated to the new. steady state value. 0.2 ‘The conventional surge tanks include ‘simple urge tank, throttled orrestricted orifice surge tank, differential surge tank ‘and surge tanks having upper tanks have been A preliminary siudy of the water conductor id power house with an installed capa dispensed with without andjor lower expansion chambers or galleries or spilling arrangements, etc. A conventional surge tank is usually located between the almost horizontal head race tunnel and the steeply sloping pressure shafi(s) with its top either open to the atmosphere or connected fo it through a suitable conduit. Such an arrangement necessitates an adit at the top of the surge shaft the access to which is generally difficult and costly. This can be avoided in many cases by the adoption of a closed surge chamber with air i which the mass oscillations are absorbed min the pressure of the compressed, the val air (Figure 1). 3. Until recently the surge tank was considered an jntegral part of hydro-electric schemes having long ressure conduits. For the first time the surge tank Nas eliminated at the Paotina(?) hydro power plant in Tasmania, Australia which is fed through a 8150 m Tong pressure conduit and has a gross head of 833 m_ and jnstalled capacity of 5x60 MW. The elimina tion of the surge tank did increase the pressure in the’ penstock but not enough to justify the cost of surge Yank. The successful operation of Paotina fed to the elimination of surge tanks at the Rowallan, Wilmot,Aa = 7 te As C lem Poe) ko apes eh ve SRN P Ne 2c Canam, GUPTA AND SHARMA te pew nates ee 1 = SPECIFIC WEGHT OF WATER Peg PRESSURE. OF AIR CUBHOH int — conventional svace TAN Te Kiaceusien suace: rank FIGURE 1: Se through power plant with ‘surge tank. Lemonthyme and Fisher developments in Australia, ‘Similarly, the surge tanks were eliminated on the head race tuanels of the first and second line of the Nurek(?) hydro power station the stal increase slightly the in U.8.S.R. without affecting iy of the system, ft was found necessary to Fesponse time and introduce minor power restrictions. The conduits of the first line of the station are }020 m long und of the second line 1265 m, the maximum head being 375m and 386 m respectively. another interesting example in which’ been eli Jordan Rives wer Plant(*) is airge tok hus fated... ‘The water conductor, system com- prises 5750 m D-shaped, mostly unlined, pressure funnel with a equivalent diameter of 4.9m and 1400 m of steel penst cock reducing in diameter from 3.2m at the upstream end to 2.7m at the power house which has a single 153.68 MW Francis turbine near the power"house, working under a net head of 265.5 m. Because top. graphy did not permit an economical surge tank ‘a pressure regulating valve was Provided to keep the transient pressures within Gesiga limits and to ensure stable transient operation. 0.4 The elimination of surge tank does not only feduce the petiod and cost of construction of the water conductor system of a hydro power project but fiso increases annual production of electrical “energy by eliminating the hea: tunnel with the surge t 'd losses at the junction of the ‘ank. Therefore, for hydro elect- tie schemes having long pressure conduits itis advisa- bie to carefully exam ‘ne the possibility of eliminating surge tanks, The elimination of surge tank would no doubt increase the pi tend to make the syster esse in the conduit, would m unstable and put more strain fn the governor. The system can, however, be made feasonably stable by increasing the response time, the fiywheel moment, constant inertia of the pipe lines ans by introducing minor power restrictions. The present paper discusses the role of the above para- Meters and the possi Yanik. ibility of eliminating the surge J. Ceiterlon for Providing Surge Tank: 1.L Whenever there is a change in the power demand ‘of the electric system, a dilfereat level of power fi Absorbed. by the goterater and an unbalanced torque xistson the turbine Which gives se to a speed change. The basic equation for speed change eat be written at Tail i ‘ae Tur Tonto wal) where, J=Polar moment of inertia ‘of rotating parts, Tyaenet torque being applied to unlt, and Werangulur acceleration, ‘The polar moment of Inectia includes the rotating mechanical parts and fluid, although ina turbines Fonerator combination the Muld f& generally x Ammall tion of the total mars and {s usually neglected, jeretore, wR 1 y 9) where, Waewoight 4 Restadius of gyration, The term WR" is known as dywheel effect and is defined as the capacity of a rotating mass to store of supply energy. It will, therefore, have a tendency to maintain constant speed of the turbine while the governor adjusts the flow of water to limit speed variation after a change of load. 1.2. Tho regulation of the machines causes sudden positive and negative pressure variations ot water Jammer in the pressure conduit which has, thereforey to be designed to resist the bursting effect duc to positive pressures and the collapsing, effect due to Thegative pressures. The interisities of those pressures depend upon the length of the pressure condult (1), the velocity of flow in the conduit (V), and the timo fof closure or opening of the gitide vanes (7,), | ‘The excess pressures can be reduced considernbly by increasing the time of closure (Figure 2). Howeve!, the time (Te) is governed by the allowable maximum speed. variation and is also dependent upon the pata- theters of generator design, eg, its flywheel, effect, tte ‘The transient irregular operation of the unit With load shedding ic expressed as, Pinas~Me F mae 3) where, no is the normal speed of rotation, and Minas js the maximum speed of rotation during the transient. The value of mer is usually taken(*) as 0.40.5. ‘The need to limit speed and pressure variations thus ee eeeBLIMINATION OF SURGE TANKS—A CRITICAL STUDY 3 VELOCITY OF PROPAGATION OF PRESSURE WAVE fen TIME OF COMPLETE CLOSURE OF VALVE. = Lenatn oF conouit VELOCITY OF FLOW WW CONDUIT toute INTAKE vive Weve dargeue ante nase db FIGURE 2: Variation of maximum pressure rise along the Demntock for “Bnite time of clovere Une pre- avare rise ds expressed im height of a ‘water ‘colema). r imposes a corresponding limit upon the unregulated length of the pressure conduit, namely the equivalent length ftom the machines to the first free water sur- face. To avoid undesirable speed-rise a governor is fitted 10 the machine. The mechanism is sensitive to speed variations and operates automaticelly to re- duce the ater supply Zon redocton of toad aad vice versa. fe closure or-opening timed equaey ‘Valles from Ito Wie or a given tbo geherator (i.¢, given WR*) ‘the shotter the governor time, the smaller the speed rise. On the other hand, the more rapidly the governor adjusts the guide vanes, the greater will be the resulting pressure varie ation or water hammer for which the conduit has to de designed. 1.3 It can be shown that the speed and pressure veri- ations are to some extent dependent upon the hydrau- lic inertia of the water mass enclosed in the pressure sonduit (VL/H). This inertia is characterised by the fime constant (Tw) of the inertia of the penstock, i Zhi Vi Twn ao) where, Li and Vi are the length and velocity of the flow at different sections of the high pressure stretchi, e.g., penstock, scroll casing, draft tube, etc., and Ho is the net head on the turbine. The elimination of ‘surge tank results in an increase in the inertia of the water “mass enclosed in the high pressure. stretchy Other things being equai shorter length of penstock means better speed contro! and smatler pressure rise. Formerly the value of Tw over 2 seconds was seldom. used, The curreht norms permit the value of Tw Up to 5 and above for hydro power plants whose capa- ity does not exceed 20 percent of the power system(*). king Tw as 2.5 seconds and the maximum wumriel velocity as $ m/sec Equation (4) yields, L=5 Ho 65) ‘Therefore, if the length of pressure tunnel is less thay five times the net head on the machines, normally no. surge tank is needed. Further, if the contribution of the power station in the total grid is less than 20 per- cent, the surge tank can be eliminated from the water conductor system having equivalent length upto 10 times the head on the machines, 2. Design Modifications to Eliminate Surge Tank 2.1 One of the important functions of the surge tank is to provide a free reservoir surface close to the power house and thereby limit the maximum over Pressures in the penstock by reflected water hammer waves and prevent the tunnel upstream of the surge tank from being subjected to high over pressures, In the absence of a surge tank, the water hammer pres- sures can be substantially reduced by increasing the closure and opening time of the turbine guide vanes and adopting a high value for the permissible speed variation of the unit during the transients... - This solution was adopted for four Kuban hydro. | electric stations by the Gidroprockt Eperimental” Design Office on the basis of a detailed investigation of transients on the units and_penstocks in addition to frequency regulation and active power on, the system. The total saving on these power stations, according to the Gidroprockt data, was more-than 11 percent of their total cost (*). 2.2, Mosonyi(?) recommends that the goyerngr-vand the inlet valve ahead of the turbine should prefetably be so adjusted as to prevent over: pressu‘es from exceeding , : up to 50 m head, 50 percent ' up to 150 m head, 25 percent over 250 m head, 15 percent i Should this require a closure impermissibly long. the fpotaliion offessare rele valves become necsbsay. “These, however, have no effect on the pressure drop ‘occurring in the penstock. &4 GUPTA AND SHARMA 2.3. Besides, the elimination of surge tank can also be ade possible by increasing the diameter of the pres- sure conduit and thereby reducing the time constant Of the inertia of the penstock, Tw This would, how- ever, increase the capital outlay and has to be decided after comparing the same with the cost of the surge chamber. 2.4 The second and third functions of the surge tank {para 0.1) enable the turbines to pick up their new oad quickly and thereby to control speed variation, and voltage and frequency change in the system. The latter is not important if the contribution of the power station in the total grid is not significant. In the absence of the surge tank these functions can be performed toa great extent by the increased inertia of the rotating mass if found economical. Besides, minor power restrictions may also be necessary. 3, Examining the Possibility of Surge Tank Elimina- tion at Tehri Dam Project 3.1 Tehri Dam Project envisages the construction of a 260.5 m high earth and rockfill dam across River Bhagirathi about 1-5 km downstream of its confluence with River Bhillangana near Tehri town. The design flood has been adopted as 15,300 m?/sec and a chute spillway with a designed capacity of 13,150 m*jsec is proposed to be located on the right bank. According to the latest plans an undergcound power house with an installed capacity of, $200 MW (likely to be modified to 8x 250 MW) is proposed to be located on the left bank. Two open intakes will supply water to the power house through two 12'S m diameter head race tunnels, each feeding four tinits. According to the present location of the power house, the overall Jength of each head race tunnel including penstocks works out to approximately 900 m. The: gross: he on the machines varies from a maximum of 231% at normal maximum reservoir level of 830.0 to a minimum of#T41'3'm at the dead storage level of 40:00. ‘The maximum discharge required to generate | 1600 MW at the minimum reservoir level works out to 1472 m'/see which gives a maximum velocity of 6 m/sec through each head race tunnel The average head losses work out. to approximately 12%. The two diversion tunnels on the left bank will later serve as tail-race tunnels. The proposed layout of the scheme - is shown in Figure 3. a 3.2 The peaking capacity of the power grid in Uttar Pradesh is likely to be of the order of 11,750 MW and the -total installed capacity of more than 14,000 MW by the time Tehri dam Project is completed. Thus, the contribution of Tehri power house to the (tal grid will be of the order of 11 to 12 precent. SCALE Remenals METRES coe SHEL bo abo so 0 gf ‘SURGE TANKS. O25m a) POWER HOUSE (ax 200Mw) Se FIGURE 3: Layout of works of Tehrl Dam Scheme. SRM la l Pelt DAD PD. BP PADPRPRPPHDEDHHDPDEH DOLD=~ wT ewer rer weer wwe ewe ewT ewe eweewe wwe we w~Twwerwwwvwvs we wo . cs! k io ELIMINATION OF SURGE TANKS—A CRITICAL STUDY 3.3, According to the’ present design of water con- ductor system for the Tehri power house, the time constant of the inertia of the penstock works out to, oe ELV 9006 eGR, Two Ho ~ F81K (aT 3 ) Since according to the current norms a value of Tw upto 5 and above is permitted(*) for hydro power plauts whose capacity does not exceed 20 percent of the power system, there appears to be a possibility of inating the two 32°5.m diameter surge tanks at Tehri dam project. With a view to find a suitable site for the location of the power house cavity further upstream, goological investigations are in progress, IP it is found feasible to. locate the power house further upstream, the possibility of the elimination of surge tank will further increase. Besides, by adopting certain modifications in the design of the water con- ductor system and of the turbo-generator as indicated in para 2, the system can be made steble without the surge tanks. However, detailed analysis of the system vis-a-vis economic considerations has to be carried out before arriving at a conclusion. These studies are in progress. 4. Conclustons 4.1 The climination of a surge tank from the water conductor system of a hydro power plant docs not only teduce the cost of its installation and the petiod of construction of the scheme but also increases. the annual production of electrical energy by curtailiag the ocal head losses at the junction of the pressure con- J duit with the surge shaft. 7 4.2 Asa general rule, if the unreguisted tength of the pressure tunnel is less than five times the net head on the machines, no surge tank is normally needed. 4.3 The value of the time constant of the inertia of {the penstock [ Equation (4) } upto 5 and above can be permitted if the capacity of the hydro power plant in question does not exceed 20 percent of the capacity of the grid to which the station is connected. The time constant of the inertia of the penstock can be reduced by increasing the diameter of the conduit within, eco- s 4 In the absence of a surge tank the pressure variations in the pressure conduit can be reduced to a great extent by increasing the response time of the governor. The extent of over pressures can be further Teduced by installing a pressure relief valve. The system ean be made stable by increasing the inertia of the rotating mass’ and putting some minor power res- trictions, if needed. 4.6 _A preliminary study of the water conductor system for the proposed Tehri hydro power plant shows that there is a possibility of eliminating the surge tanks. However, detailed analysis of the system needs to be carried out before coming to a definite conclusion. 5. References () SHARMA, H.R. and PETER, J.C: “Air Cushion Surge Tank". Proceedings, Voll, Alf india Symposium on “the Economic and Civil Engineering Aspects of Hydro-etectric Schemes, Roorkee, 14-16 Aprit, 1978 (2) GRIFFITHS, P.7.A.:_ “Surge Problems of the Hydro- electric Commission ‘of Tasmania, International Con- ference, on Pressure Surges held’ at the University of Kent, Canterbury, England, 68 September, 1972. (3) POKROVSKII, BLM. etal : “Regulation of the Hydra. lic Units of the Nurek. Hydro-electric Station without ‘Translated 11, pp. 13-14, Surge Shafts in the Head Race. Tunnels’ from’ Gidro-teknicheskoe Stoltel'stevo N November 1971, (8) PORTFORS, B.A‘and CHAUDHRY, MAHL: “Analysis and Prototype Verification of Hydraulic. Transients in Jordan River Power Plan”. International Conference ‘on Pressure Surges held at the University of Kent, Can: terbury, England, 6-8 September 1972. () STREETER, V.L, and WYLE, E.B. : “Hydraulic Transl- ento™, McGraw-Hill Book Company, New York, ‘St. Louis, San Francisco, Toropto, London, Sydney, 1967 (6) ARSHENEVSKIL, N.N. and TRUBITSYN, YUN. “Significance of Dispensing with Surge, Chambers to Lower the Cost of Hydro-electric Stations". ‘Transl from Gidro-technicheskoe Stroitel'stevo’ No. 9, 10-14, September 1971 (7) MOSONY!, EMIL : “Water Power Development Vol, 2". Publishing House of the Hungarian Academy of Sciences, Budapest, 1960. PP. @ BROWN, J. GUTHRIE: tric Engineering nomical range. o : Pracice-Vol, 1; Civil Eagieering’ lacke& Son 4.4 In many cases it 8 possible’-and advisable to dis- Mala eae Eusen | pense with the surge tank. Therefore, the possibility (10) DAVIS. CV. and SORENSEN, K-E. | “Hand Book of Of surge tank elimination should be thoroughly exa. — PUlled, Macrawlee oat ee MgntoT end mined in each case. Kogakusha Company Ltd., Toky9. 4) YW