FLOW

METERING

HYDROELECTRIC

SYSTEM

AT THE

PLYAVINYAS

STATION

Go I. Prokhorov, Ao Do Taukach, and Vo So Serkov

UDC 621211.21 : 627.815

The Lenin Plyavinyas hydroelectric station is cm'rently the uppermost unit in the cascade of hydroelectric stations on the Daugava River. The headwork structures (overfa11 power-station building and earth dams) impound a reservoir 57 km long, with a total storage of 546 million ms at normal retention level (NRL), and a surface area of 34.9 km z. The design regulation prism has a height of 7 m and a volume of 184.6 million m3. The reservoir effects a weekly flow regulation. The power station building houses 10 vertically mounted rurbogenerator sets powered by radial/axial hydraulic turbines, type RO 984-VB-600, with a working head range of 29 to 40 m. At the design head of 34 m the units have a capacity of 80.000 kW, with a turbine discharge of 320 mS/sec. The spillway part of the power station building has 10 bays 14.m wide/ the maximum crest head is 10 m. The maximum capacity of each spillway opening at NRL is q90 mS/sec, according to full-scale measurements [1]. Operation of the hydroelectric station during a sharply peaked load regime with a deep diurnal flow regulation, presents special requirements with respect to the accuracy of computation of the weekly and diurnal water balances, which depends on the necessity to obtain accurate data on storage and inflow, aimed at planning the power-station operation as a part of the power system. Therefore, considerable significance is attached to flow metering and water-balance characteristics. In accordance with the prescribed operating ins~uctions [2, 3], a continual monitoring of flow through the Plyavinyas hydroelectric station is carried out, i.e., the mean daily discharge passing downstream each day is deterrnined. The equation for computing the mean daffy discharge has the following form: Qaes=Q~ ~-Qnl+Qs 4-Q/+Q3r, ma/sec,

(1)

where Qhes is the mean daily discharge past the hydroelectric station; the mean daily discharges are, respectively: QT-- through the turbines, with the turbogenerator units operating in the generating mode; Q n l - the same, with the units operating in the no-load mode; Qs - through the spillway openings~ Ql - leakage and seepage; Qsr - station's internal reqniremeuts (usage). The components of Eq. (1) differ greatly in their absolute values, and influence the overall accuracy of the mean daily discharge to different degrees. At the Plyavinyas hydroelectric station, which has a flow-utilization coefficient of 0.95 over the full year. the major part of the flow passes through the uuhines. This compouem QT represents the principal segment of the total, and the errors incurred in its determination in practice determine the total error in the monitored flow. Because at the Plyavi-nyas power plant, as at most other hydroelectric stations, there are no turbine-flow meters, the turbine discharges are calculated, using the operating flow characteristics of the turbogenerator units, by the following equatiom Qr=

Qam St 24 , ms/sec' (2)

where QT is the mean daily turbine discharge, with the turbogenerator in the generating mod~ Qam is the mean turbine discharge during the operating period of the turbogenerators~ Et is the total period of Operation of all the turbogenerators in the generating mode, during the days under consideration, h.

Translated from Gidrotekhnicheskoe Stroitel'stvo, No. 9, pp. 19-23, September, 19q5.

842

FLOW METERING SYSTEM AT THE PLYAVINYAS HYD ROELEGTRIC STATION

843

A practicany identical equation is used for calculating the flow through the spillway bays during the passage of floods. Use of the operating characteristics, and computation of the mean daily discharges using Eq. (2), require analysis of the following two forms of error: that camed by insufficient accuracy of the operating characteristics used, and that due to methodological errors incurred in determining the mean daily flows. These errors introduce systematic inaccuracies into the computatior~ whose identification and elimination present the greatest difficulties. With regard to a third form of error, namely that due to accidental errors incurred during the direct measurements of power, head, and other quantities, these have a considerably smaller effect, due to the combination of a large number of quantities, even when examining the daily balance. During the initial period of service, the mean dailydischargesof the turbines were determined by using the design characteristics of the turbogenerator unit, for the averaged load of all the units and the mean daily head; the averaged turbogenerator load was determined by dividing the daily energy output of the power station by the total number of working hours of all the turbogenerators in the generating mode. The mean daily head was computed as the arithmetic mean of the hourly measurements, without allowing for losses at the intake screens. This, the simplest computation method, involves also the maximum errors;, an analysis carried out subsequently [4] showed that merely the use of the weighted mean value of the head instead of the arithmetic mean, effects a correction of the mean daily discharge, increasing it by up to 2~ even with the turbogenerators working continuously under a varying load. If the hydroelectric station operates, as at Plyavinyas, in a sharply-peaked regime with intermittent stoppages of the turbogenerators, the correction is increased a further 1 to 2%. An even greater error was introduced by insufficiently accurate values of the design characteristics. The large errors in flow monitoring did not meet service requirements and the service personnel were faced with the problem of improving the monitoring methods. With this objective a number of measures were undertaken, namely: a) fuU-scale power tests on the turbogenerators with a more a c c ~ a t e evaluation of their characteristics in service (two units tested by current-meter measurements of the t~bine discharge, the remainder by the comparative method); b) full-scale tests of the waterway openings, with a more accurate evaluation of their discharge characteristics [1_]; c) a more accurate determination of the live storage in the reservoir by surveys, employing an echo-sounder type IR~L-1; d) based on the accumulated service experience, an analysis of the method used for flow metering and a review of the local flow-metering instructions. The State Trust for Organization and Rationalization of Regional Electric Power Plants and Networks (ORGR~S) participated in all of the above-mentioned work. It was established that the turbogenerators have characteristics which substantially differ from one another, owing to differences in the dimensions of the water passages. For instance, for heads of 36 to 38 m and a load of 75,000 to 85,000 kW the divergence of the discharge characteristics ranges up to ~ o . The full-scale tests also showed that the actual service characteristic differs from the design value; thus, for an 80,000 kW load the turbine discharge exceeds the design value by 4 to 6%. The more accurate determination of the discharge characteristics of the water-passage areas made it possible to establish that at partial gate openings of up to 2 to 5 m, the model characteristics overestimated the discharge by 6 to 13%; at full gate openings the discharge capacity of the passages was found to be 2 to 3% less than given by the model-test data. The surveys of the reservoirs established that the volume of the regulating prism was in fact 2~ less than the volume given in design data. Thus, the errors in the service characteristics used in flow metering were substantially reduced as the result of the full-scale investigations. As was indicated above, the major part of the flow passes through the turbines and the errors permitted in its determination establish, in the main, the accuracy of flow measurement at the power station. With a hydropower plant operating in the sharply peaked regime, the smallest errors in computing the mean daily discharges occur when making a detailed determination of the flow individually for each of the turbines, according to its mean power and mean head during the operating period each day, using the equation Q

Qztl + Q2/~ + 9 . 9 + Qlot~o --

24

mS/sec ' '

(3)

where Q1, Qz . . . . . Q10 are the average discharges of each turbine during its operating period, m3/sec; and ti,t~ . . . . . ti0 are the operating times of the different turbogenerators in the generating mode, h. For the Plyavinya~ hydroelectric station with its relatively large number of turbogenerator units and a sharply peaked operating regime, the detailed computation of the turbine discharges is very time-consuming or requires the

844

G. I. PROKHOROV ET. AL. ~.o

!

i

;

',

!

I H.=4er~ J - - - - ~

T

I//S/,2~'f'/oo

b,'~ =36m

r

lo

r,.9

f:o

~

i

O

50

[

too

! t50

I Z6#

i .

1 ~

{

Z,~O

300

350

Discharge through turbine, m3/sec Fig. 1. Operating discharge characteristic curves of turbogenerator (No. 6 at station) of Plyavtnyas hydroelectric station, obtained from full-scale tests. 1) Lines of turbine head; 2) lines of distributor openings.

76

I I t/

IM..-..~--I I

,a

~

iill

o

o

I

I ;

g

3 a

q

1 o

3

$

, h

b

Fig. 2. Chronograrns of turbogenerator load. a) Turbogenerator No. 6, December 15, 1971; b) turbogenerator No. 7, December 17, 1971. application of a special computation technique. In the absence of the latter an examination was made of the possibility of simplifying the computation of the average daffy turbine discharge, using Eq. (2). This computation method involves a systematic underestimation of the actual average daffy discharge of the turbines owing to the errors arising as the result of averaging the power and the head [5]. Let us examine these errors in greater detail. Errors due to Averaging of Turbogenerator Power. Presented tn Fig. I are the operating discharge characteristic curves of one of the turbogenerators (No. 6 at station) plotted from the results of full-scale tests. For turbogenerator loads exceeding 60,000 kW the relationship between output and discharge assumes a curvilinear form. The P1yavinyas power-station turbogenerators nsually carry a load of nearly 82,000 kW, but for several reasons it persistently diverges by 10 to 20,000 kW below th~ maximum (Fig. 2); sometimes the load fluctuations increase and the divergences range up to 30 or 40,000 kW. Presented in Table I for purposes of evaluating the effect of these fluctuations, are typical results of the computed average daffy discharges of hydraulic turbines Nos. 6 and 7. The computations were carried out by two method~- using the averaged values of the load and the head over the whole period of turbogenerator operation; and, in detail, using the values of the load and the head over short intervais (15 min) these being obtained from the graphs plotted by automatic recording instruments at the station (see Fig. 2). From Table i it follows that the computation based on the averaged figures underestimates the average daffy water discharge by up to 3%. Further, the underestimation of the average daffy discharge occurs also when averaging the output of all the turbogenerators, if computed by Eq. (2). A comparison of these discharges with the sum of the discharges of the operating turbogenerators, determined separately for each, shows that the underestimation of the discharge ranges up to 2% (Table 2). A more detailed analysis, taking into account the regime characteristics of the operation of the Plyavinyas hydroelectric station, showed that on the average and for all regimes, the average daily turbine discharges are, as the results of averaging the turbogenerator outputs, underestimated by up to 2%. Errors due to Averaging of Head. The actual value of the turbine discharge can be determined most accurately by using the weighted mean value of the head dining turbogenerator operation, averaged according to discharge or time, using the following equations:

FLOW METERING SYSTEM AT THE PLYAVINYAS HYDROELECTRIC STATION

845

TABLE 1 t.t., t"~ 0

,~..~ ~ 0

Discrepancy

Z,

D ate

[.~ ~'=

,~ o

25.3 270

ec.

~

~..~

260 27'9

aq

%

+7 +9

2,8 3,2

TABLE 2 O

O~"

Z

.o.,~ ,

o r "=' ~ '~

"fi

.E-, 2

3 4 5 6 7 8 9 lO

=

.-,,--, . ~ Z

~ .q

~ ~

589 380

6,43 4,45

80,9

570

7,37

77.3

550 600 170 170 49O 570

~E

= ..=

6.87 7,07 2,05 4,85 6,03 7,15

90,2

80.0 84,9 82,9 92,5 81,3 79,7

37.0

38,0 38,0 38,0 88, 0 38,0 38,0 38,0 38,0

'~5

250 i 235 245 267 258 332 252

245

Sum, for turbogenerator computation

Total + 4343 52,27 83,10I 38,0 I 258 avgra~e~

1939,71 1112,50 1731,95 1683,15 1887,69 528.90 1610,20 1519,58 1751,75 13735,41

572,3

13485,66

561,9

va~ueff

Discrepancy between the two methods of computation: A Q = I0• mS/see, or 2~o.

/-/wt.mn=-VH Qhes/XOhes,

(4)

or

l~t~

(4 ')

For the Plyavinyas hydroelectric station, the weighted mean values of the head, eomp~ed according to discharge of time respectively, are practically e q u a l An analysis of the arithmetic and weighted mean values of the head, determined from the load graphs and head- and tail-water levels, showed that the mean value of the discrepancy between these heads reaches 2%; also, the arithmetic mean of the head is found to be always greater than the weighted mean value. In other words, the mean daily discharge of the turbines, determined from the mean arithmetic value of the head is likewise about 2% less than the actual. Thus, the determination of the average daily turbine discharges by Ect. (2) and from the mean arithmetic value of the head for the peak-load graph, under the conditions at the Plyavinyas hydroelectric station, leads to an underestimation of the computed discharges compared with the actual values. The overall value of the error in " determining the turbine discharge, arising from averaging the loads and heads, can be evaluated by summing these errors and amounts to about 4%. A plant analysis made it possible to take into account the systematic errors when computing the flow, by introducing a constant in Eq. (2):

846

G. I. PROKHOROV ET. AL. TABLE 3 Dis!crepancy ~11:l 0 ~ii I ~l ill ~.~ =I 0 o 3 ~ =i 0 ..i.i

Year

0

1966 1967 1968 1969 1970 [971 M~7~

~ iod

Discrepancy

~'~

'=' 0 "~

610 455 518 372 605 461 370 424

582 442 494 369 577 451 355 409

~

sec

--28 --13 --24 -.-3 --28 --I0 --15 --15

^

4,6 2,9 4,6 1,0 4,7 2,2 4, I 3,5

~< sec i

605 459 513 383

--5 +4 --5 +II --5 +8 --I +I

6,90

469 369 425

I

0,8 0,9 1,0 2,9 0,8 1,7 0.3 0,2

19664

TABLE 4 / I ogo~Q~

I~

~-,-,oi~

i ,~, i,..#

t~

NN"

~

~.

_1

~pancy

Ims/

%

ii,,i ,~i i....7

2O4 444 525 371 195

242 5O2 191 408 552

:~ , . . l ( P l v a v i n -

~ 2 ~lvas'-Ke-

~l.,.., ~

+38 94-58 ~

+47

IDiscrep.

l

Dis-

l~l.., ~J,

I 18,6l 13,0 [

243

6,2 2,9

1~8~8

9,3 |

575

509

QT=~1.04(QmnEt124),m s / s e o .

+I +7 +

0,4 1,5 1,3 2,9 4,2

14,2 (2')

In computing the mean head for the operating period of the turbogenerators, no allowance was made for the losses at the trash screens. An estimate of the trash-sereen losses for each turbine when operating in the peak-load regime, would be very complex and time cor~uming as the losses are not constant throughout the day, and depend on the condition of the screens and the load on the generators. Head losses at the screens were measured during comprehensive power tests on two of the turbogenerators; it was established that the overestimation of the turbine heads resulting from ignoring the trash screens amounted to about 1%, therefore the turbine discharge was underestimated to about the same r As the Plyavinyas turbogenerators carry a load generally close to their maximhm output, the screen losses for which are about 0 2 m, it was decided to simplify the estimate.of these losses, when computing the flow, by assuming them. Another simplification made in estimating the flow was that, in the computations, the characteristics of turbogenerator No. 6 was used for all the sets, as a special analysis had shown that this provided mean daily powerstation flows which lay closest to the values determined from the characteristics of the individual turbogenerators. The above method of estimating the discharge was used at the hydroelectric station from the beginning of 19q4. The results of applying this method can be judged from the annual water-balance computations compiled for the reach of the river between the Ekabpils gaugingstadonlocated at the point of its entry into the reservoir, and the power station site. The relevant data are presented in Table 3. The mean annual fl0wspast the hydroelectric station, determined during the preceding years by the old method, are less than the mean annual flow at the gauging station site, i.e., the discrepancy is systematic and its maximum value ranges up to 5%, which exceeds the ~allowable figure for the yearly balance by 2 to 3% [6]. The discrepancies in the mean annual discharges when using the new method are both positive and negative, i.e., they are random, and lie within a range which does not exceed the recommended maximum values.

FLOW METERING SYSTEM AT THE PLYAVINYAS HYDROELECTRICSTATION

847

TABLE 5

Month, ,~

~ 9

~Froms hydroelec.-flow t adatat IFr~ i oHydr~ n OT, | WTz I-, ~ m~f/'cse'c~'O6mSlies Im3Z~ed~"~ I~es

Feb. 49,370 38.30 " 1109'720 37,37 ] 77,170 38 67 Oct. I 92,690 38,19

232 492 336 398

561,3 84 1317,8 82 870,9 I067,9

194" 469,3 I lO! 434 I I182,4 93 361 935,7 78 340 910,5 98

* A discharge o f l 0 m3/sec is deducted from the mean daily flow at the Plyavinyas station and Ekabpiis gauging-station site, representing leaks through the stator guides of the turbines, when they are on standby or operating as synchronous condensers, and including the seepage through the various hydraulic structures. The information presented in Table 4 gives the mean monthly flow past the Plyavinyas hydroelectric station during 1974, computed by the plant personnel in accordance with the new method, analogous quantities obtained from the gauging-station data, and also the mean monthly flow past the Kegum hydroelectric station, which lies below Plyavinyas. As can be seen from Table 4, the data computed by the new method agree well with the data obtained at the Kegum station. In only two instances does the discrepancy reach 4.2%, which, nevertheless, does not exceed the recommended allowable values (5 to 10%) for the computation interval of one month. The discrepancy between the mean monthly flow past the Plyavinyas station, determined from the Hydrometeorological Service information, and those determined at the station in particular months exceed the allowable values and reaches 16 to 19%. The reason for such significant discrepancies is evidently due to an insufficiently accurate determination of the discharges at the gauging-station site during the winter months, and to an insufficiently clear evaluation of the hydrologic regime of the Plyavinyas reservoir by the hydrometeorological measurements, e.g., later inflows, rainfall, etc. These assumptions are confirmed by computations of the mean monthly values of the efficiency of the Plyavinyas hydroelectric station, using the following relationship: n = (367,2 E/WH),

(5)

where E is the monthly output of electric energy by the turbogenerators, kWl~ W is the monthly runoff through the turbines dt~ng the operating period of the turbogenerators in the generating mode, m3~ H is the mean monthly head during the same period, m~ The results of these computations are presented in Table 5. According to the flow determined from the gauging-station data, the mean monthly efficiency of the hydroelectric station would have amounted to 93 to 101% drying the months with the greatest discrepancies; such operational factors are unrealistic and indicate an underestimation of the flows through the turbines~ From flow data obtained at the hydroelectric statior~ the mean monthly values of its efficiency for the same months are 82 to 84~ which are in accord with the operational conditions at the Plyavinyas plant. The magnitude of the discrepancy between the daily balances at the gauging-station and power station sites, respectively, is still significant, and for specific days ranges up to 20 or 25% o However, this discrepancy is not characteristic of the exrors in flow estimates at the station site. but is explained by errors in determinations of the daily drawdown or reservoir-filling volumes resulting from wave action, windbtide (rise or fall) phenomena, and dynamic variation of the regulating-volume prism with sharp changes in inflow during floods. It should be noted that in compiling daily balances a smaller discrepancy has not been achieved so far at any major reservoir. The installation of reliable automatic inslzuments for measuring discharge through hydraulic turbines could greatly reduce the volume of work involved in estimating flow, particularly at power plants with several turbogenerators which operate under conditions of diurnal regulation.

848

G. I. P R O K H O R O V ET. AL.

CONCLUSIONS 1. When computing the average daily discharges through hydraulic turbines, using their operatS.ng characteristics, systematic errors can arise, caused by the use of imperfect methods of determining the mean loads and heads. The large volume of effort involved in a detailed computation of flow can be avoided by introducing correction factors into the simplified equations used, whose values can be determined by making a detailed analysis of the available flow-estimating data. 2. A substantial increase in the accuracy of the flow estimates at a hydroelectric station is achieved by fullscale tests made to determine the operating characteristics of hydraulic turbines and spillways. LITERATURE I.

2. 3. 4. 5. 6.

CITED

V.S. Serkov, A. S. Vorob'ev, A. P. Gur'ev, and L. N. Baichikov, Discharge Capacity of Spillways at Hydroelectric Stations [in Russian], ~nergiya, Moscow ('1974). Basic Conditions for Estimating Flow at Hydroelectric Stations [in Russian], Min~nergo SSSR. STsNTI ORGR~S (1974) o

Instruction for Estimating Water Flow at Hydroelectric Stations [in Russian], Minerergo SSSR, STsNTI 01~ RES (1975). Method of Estimating Water Flow at Hydroelectric Stations [in Russian], Minenergo SSSR, STsNTI ORGRES (1974).. R.A. Shestakova, Accuracy of Determining Flow through Turbines at Large Peak-Load Hydroelectric Stations [in Russian]. Trudy GTI, No. 185 (1974). A . M . Gavrilov, Fundamentals of Flow Estimates for Hydroelectric Stations [in Russian], Gidrometeoizdat, Leningrad (1965).