Dear David,

 

The first phase of any index test is a thorough checkout of the unit’s control system operation. Kaplan turbine gates and blades must be positioned accurately and robustly in order to maximize operating efficiency and service life. The Index Test Box was developed at Woodward Governor Company in 1985 as a dual-purpose test instrument; the primary function was to analyze a Kaplan turbine’s efficiency performance in order to optimize or tune-up the 3-D cam surface profile and the secondary use is diagnostic analysis of the turbine control system (aka governor) behavior. One of the ITB design objectives was remote testing to allow an engineer back in the office to observe and diagnose turbine control systems in the field, in real-time over the Internet - without leaving the office.

This capability has been demonstrated on 3 occasions.

 

Due to fears of malware, viruses and hackers powerplant personnel have always objected to any new equipment from an unknown supplier getting connected to their governors or SCADA system, and direct connection to the Internet of any kind is rejected outright. Low cost CDs, thumb-drives and high-speed broadband Internet transfer of data files recorded on the normal powerplant data logger provide suitable alternatives to a direct connection.

 

A proper index test  includes a dewatered inspection of the unit so that the surface condition of water passageways and turbine working surfaces can be checked and the full range of gate and blade motion can be verified from squeeze to full open while the normal powerplant instrumentation and index testing equipment are adjusted and calibrated.

 

In lieu of such a comprehensive inspection, with a few modifications to the datalogger programming a relatively modest review of the time-response behavior of the unit can be made to show if gate and blade movements do not conform to ASME and IEEE performance standards. Using this new method, several problems were found with the Kaplan unit control system at Dorena Dam.

Synchronizing a unit to the grid before closing the breakers on startup is a tedious and precise requirement of the turbine control system. In this case the ITB’s stripchart playback function shows that the efficiency bounces very hard when the breaker is closed on startup, which indicates the synchronization may not adequate. This was seen in the typical operation data from the data set from 11-18-2015.  

StartUp

Figure 1 ITB Cartesian coordinate display during normal startup

 

The green traces on the left side show that the rotor is not spinning at the exact speed as the stator’s magnetic field when the breakers were closed. The unit should be at “speed-no-load” with flow but no power generation i.e. a slightly negative number. What is happening is when the breakers close the rotor bounces against the magnetic field until it is pulled into synchronization. 

 

This may or may not cause damage to the unit, but the control system should synchronize the unit to the grid better than this for reliable long-term operation.  More investigation will be able figure this out.

 

Another problem was found by looking at the stripchart traces from the datalogger files using Microsoft Excel®. 

Blade_Dither

Figure 2 Excel® Stripchart Display Showing Blade Instability

 

At first glance the blade trace in the lower chart above was noticeably thicker than the others, so it was magnified at each test point blade value used for the index test to investigate further. The Y Axis on these traces is Blade Angle in percent and the X Axis is number of scans where each scan is 0.51 seconds.   This representation of the data was then magnified even further to take a closer look.

Blade_Dither_Zoom 

Figure 3 Excel® Stripchart Display Showing a Magnified View of the Blade Instability

 

This blade trace display is magnified to show the amplitude and frequency of the instability.

The blades are always moving about 0.6 with about an 8-second period.

This could be caused by a positive lapped pilot valve on the blade servo, but more instrumentation will be needed to know for sure.

A third problem showed up with the Cartesian coordinate display of the gate vs blade motions during startup of the unit.

 

Figure 4 ITB Cartesian Coordinate Display showing 3-D Cam Output Problem

 

This chart shows a problem with the blade to gate relationship. The two red lines immediately above and below the heavy blue "On-Cam" line are the ASME prescribed deadband for proper blade positioning. This chart shows that there is unacceptable wandering of the blade vs gate positioning.

 

Another problem was seen during the index test. The blades bobbled at the high flow points while they should have been holding steady. 

Blade_Bobble

Figure 5 ITB Cartesian Coordinate Display showing Blade Position Bobble at High Flow and Power

 

It was relatively small movement and did not overly affect the index test analysis, but if the cause of it is that the Hydraulic Pressure Unit (HPU) is undersized for the blade loading at high flows, the resulting unwanted motion will wear out the blade trunion bearings prematurely and shorten the turbine's working life.

 

Conclusion

 

The index test results showed that the unit has significant efficiency losses due to its hydraulic setting (friction and turbulence from the suction system and a long penstock with several direction changes to water flow). The ITB works on gross-head so the overall efficiency performance of the unit from this analysis (76.9% at test head of 69ft) does not meet the best-case efficiency predictions of the model test data (93.61% at design head of 92.8ft). Testing the unit at its design head and isolating the turbine’s efficiency from the setting losses will be a better comparison. The ITB can do this if net-head transducers are used instead of gross head from forebay and tailwater.

 

The objective of the ITB analysis is to optimize the 3-D cam to maximize operating efficiency under the existing situation.

 

In this case, the four problems listed above will reduce unit operating efficiency and cause excessive wear and tear on the unit that will significantly hamper operating efficiency and shorten its operating life. Meeting the ASME and IEEE prescribed blade to gate positioning accuracy and dynamic response criteria are essential to maximizing efficiency. Unfortunately, this machine does not.

 

The control system on this Kaplan turbine needs the attention of a specialized “Governor Controls Engineer.” Kaplan turbine blade and gate control systems (aka governors) are more complex and intricate to design, program and maintain than a simple hydraulic positioner such as a Francis turbine gateshaft operator or a back-hoe.

 

Sincerely yours,

Douglas Albright

Actuation Test Equipment Company

(815) 335-1143