A new method has been developed to help measure leakages through closed wicket gates. Anatoli Zakharov takes a look at the benefits of this

THE need to measure leakages through the closed wicket gates of a hydraulic turbine is an important requirement in the water power industry because it affects the economics of hydro power and pumped storage plants.

Existing techniques for measuring leakages through wicket gates are essentially volumetric. They measure leakages, with the emergency gate in place, by detecting falls in the water level at the measuring section of the water conduit or spiral case over a time period. With this method, the measuring section should have a simple shape and lie downstream of the gate sill.

There are difficulties with using this method. A large volume of water at the measuring section of the water passageway necessitates time-consuming testing and it is not applicable in all cases: it is not always possible to find a suitable measuring section in the water passageway. The necessity to maintain constant head and tail water levels during testing means shutting down all units at a pumped storage plant.

A new patent-pending method has been developed to help solve these problems. Known as the express-diagnostics of the distributor, this method is based on the use of two factors.

Firstly, it uses the vent pipe as the measuring section of drop in water level. Secondly, during testing, the time of gate closing is fixed and leakages are computed with reference to this moment.

At the moment of the emergency gate closure t=t0, the water level upstream and downstream of the gate is the same, ensuring leakages through the gate are zero. Knowing the exact moment of the emergency gate closure permits exclusion of the transients (around t0) from consideration. It is assumed that the emergency gate closes instantaneous and the lowering of water level begins at t0.

Computation of leakages through the wicket gates in the differential form for the moment timed with the emergency gate closure corresponds to the actual head, and it gives an accurate value due to the absence of approximations in the computational techniques.

Comparative measurements have been conducted at the Zagorsk pumped storage plant using the new method of express diagnostics and the method specified by the Russian Standards (108.023.16-82 – Distributor for vertical shaft turbine and pump-turbine) to confirm the effectiveness of the new method.

The reversible unit has capacity of 200MW with a 91-111m head. Both measurements were conducted simultaneously. Using the express diagnostics method, readings of the reference pressure gauge mounted at the spiral casing were recorded with a video camera.

The moment of the emergency gate closure t0 was indicated by a signal lamp which was within the shooting field of the video camera. The lamp was switched on by the contacts of the emergency gate position controller. The tests were conducted on five operating units of the pumped storage plant.

In the computation using the test curve, the test curve H(t), which by its shape falls under the category of exponential curves, can be expressed in the analytical equation, where the origin of co-ordinates is matched with point t0:

H=H0*e-at +H1

The derivative of this equation dH/dt at the moment of the emergency gate closure t0=0 gives the instantaneous rate of fall in water level in the vent pipe:

dH/dt = -a*H0

Multiply it by the cross-sectional area of the vent pipe and you get the rate of water volume reduction – water leakage through the closed wicket gates.

Q= -a*H0*S

Take the following values for parameters H0 and H1 from the first equation:

H0- head acting on the measuring gauge in test (H0=109 m for curve 1, H0= 111.8 m for curve 2);

H1- tailwater backhead affecting the measuring gauge (H1= 10.55m for curve 1, H1= 9.48 m for curve 2).

Determine exponent, which is selected by iteration: a1= .00113; a2=0.00215.

The maximum difference in head by test and by computational curves does not exced 0.3m (0.27%) of the acting head. The rate of leakages according to the third equation works out as: Q1= 0.375m3/sec, Q2= 0.731m3/sec.

The standard leakage for the tested unit is Q= 0.5m3/sec. The results of the testing confirmed the good condition of the distributor of the first unit, and increased leakages through the wicket gates of the second unit.

Selection of the exponential function e-at and values of H0, H1 for the test curve representation is not strictly obligatory – it seems to be more convenient. It is possible to represent the test curve by another analytical function or by other constants provided it corresponds accurately to the actual curve. Then such an equation should be solved in the described manner: calculate the time derivative at the moment of the emergency gate closure, multiply it by the cross-sectional area of the vent pipe.

When comparing the measurements conducted by the two methods, the method express-diagnostics method has the following advantages:

• It offers speed of measurement.

• It does not impose limitations on operations of the remaining units at the pumped storage plant.

• It allows higher accuracy of measurements.

• It offers a smooth measurement process.

The measurement Q2 for A5 is not reliable both in terms of the results and in pressure gauge readings. Measurement for A1 was suspended by the load dispatcher and requires repetition. It means that out of five measurements made according to the 108.023.16-82 procedure, two measurements require repetition. The measurements using the method of express-diagnostics have appeared to run smoothly.

As an example, the express-method has been applied in the assessment of damage caused to the wicket gates of a Kaplan turbine with a capacity of 102MW and head of 40-48m. The turbine is installed in the power house integrated with outlet works. The emergency gate is located before the forebay of the spiral case, and there is a rectangular vent pipe 2.53m2 in a cross-sectional area. Malfunction of the unit was detected in the synchronous condenser mode of operation (increased air losses from the runner chamber), and suggested increase in leakages through the closed wicket gates. Leakages were measured by express-diagnostics. Water pressure variations in the spiral case at closure of the emergency gate were recorded on a paper tape. Pressure was measured by the electrical pressure gauge (-10) and by reference pressure gauge. The water variation curve is given on figure three and time of the emergency gate closure is t0=18 sec. The record of process shows that close to the moment of the gate closure, t0 water pressure tends to change following the distinct transient process. This was caused by the construction of the sealing of the emergency gate. The sealing system is to be inflated after closure of the emergency gate by injection of water which had not been used in the test. Increased leakage through the emergency gate lead to a tangible transient near the point of gate closure.

As the test has resulted in securing the exponential curve with the initial value H0= 47.9 m and steady-state value Hss= 42.5m in the vent pipe, it can be readily represented by equation:

H = H0*e-at + H1,

where H0= 47.9-42.5=5.4 m, H1=42.5 m. The origin of coordinates should be shifted to point t=t0. Using the test curve, a= 0.109, the equation assumes the form

H= 5.4*e-0.109t +42.5.

Leakages Q through the closed wicket gates are determined from the formula:

Q = S*a*H0 = -2.53 * 0.109 * 5.4 = -1.49m3/sec.

The standard rate of leakage through the closed wicket gates for this turbine is 0.7m3/sec. Examination of the distributor after dewatering of the water passageway revealed damage of the rubber seal on the blade butt end. Further physical examination confirmed the defect which was suggested by the measurement results. Measurement of leakages through the closed wicket gates at a good unit yielded the result of 0.85m3/sec. This example demonstrates effectiveness of the method in case of large leakages through the emergency gates. It should be mentioned that the method specified by 108.023.16-82 is not applicable to the above case because the spiral case does not have a section with simple geometery downstream of the emergency gate sill.

The express-diagnostic method can monitor regularity tightness of the distributor, for example, on an annual basis during repairs. Analysis of these data permit objective scheduling of maintenance for the distributor.When refurbishing or repairing the wicket gate sealing the result of this work can be easily checked by comparing the leakages before and after repair.

Operators are also able to check leakages through the closed wicket gates when taking over newly commissioned units. In case of any damage to the turbine water passage the distributor can be easily and speedily diagnosed without dewatering of the unit. The results of testing can then be used to analyse the design solutions for use in development of new generation turbines.