Wouldn’t it be nice if there was an
Automatic Index Test Box for Kaplan Turbines
By Douglas J. Albright
This article describes how index testing and then optimization, or tuning of Kaplan turbine variable geometries achieves higher efficiency over a wide range of power levels. Explanations are provided about how the new generation of computer-based Kaplan governors and 3-D cams facilitate the new “Constant Power” index testing method. This Constant Power method is delineated, and then the new Actuation Test Equipment Company (ATECo) Index Test Box (ITB) is presented with a description of how it conducts this test.
Index testing has been available as a means to improve Kaplan turbine performance for many years1,2&3 but due to the expensive, tedious and time-consuming requirements of index testing, unfortunately turbines are not often index tested. Industry organizations (ASME & IEEE) have gathered to create standards for measuring turbine control system performance6 and index testing5 to determine absolute and relative operating efficiency. A specialized tutorial8 has been prepared to assist in specifying turbine governor and blade controlling equipment. Many articles have been written to present methods to determine the optimum blade to head and gate relationships of Kaplan turbines using traditional methods1,2,3&4
In this time of heightened global concern over energy, an instrument that can automatically perform an index-test, or “tune-up” on Kaplan hydroelectric turbines to improve efficiency up to 5%, while reducing the environmental impact, has great appeal to the power industry. Index testing identifies the head and gate to blade relationship for optimum efficiency at a given water flow rate. This profile is then programmed into the turbine blade control system to realize this improved efficiency performance. A new Automatic Index Test Box for Kaplan Turbines has been developed that utilizes a “Constant Power” method of index testing that takes advantage of the ability that new computer-based Kaplan control systems have to position the blades independent of the gate stroke.
The ITB is intended only for use in index testing and optimization in order to get maximum return on investment from Kaplan turbines.
It is not intended for acceptance tests to verify that contract guarantees are met.
The Constant Power method has three primary benefits over traditional index testing methods:
1. Power level remains constant during the automatic testing sequence; the traditional index testing method requires changing power level every 5 minutes or so. These power changes must then be coordinated with the control room and dispatcher.
2. Flow level remains constant within a few percent during the automatic testing sequence; the traditional index testing method requires wide swings in flow and power and then waiting for water levels to stabilize after each flow and power level change. In most instances the waiting interval is too short which makes the collected data much noisier.
3. The new test instrument and methods presented herein allow unattended, automatic index testing, thereby alleviating the expense and tedium of this task and avoiding the problems of both #1 and #2 above.
Kaplan turbine variable-pitch blades are adjustable to achieve maximum operating efficiency across a wide power range. In early designs, the blade angle tracked the gate stroke along a 2-Dimensional cam profile; whenever the gates moved, the blades would follow as defined by the two dimensional cam response; but no automatic correction was made for head changes, Whenever water level changed enough to warrant it, the 2-D cams were swapped manually by the operator to one for the new existing head. This was time consuming and often didn’t get done - and even when it did the head-tracking resolution of this old system was still not very good.
Even greater efficiencies were made possible by the development of computer-based Kaplan 3-D cams that trim runner blade angle for minute changes in water level automatically – with up to 16 bit resolution. This provides precise, automatic high-resolution trimming of the blade to gate relationship as head varies, but in order to achieve these gains the turbines must be individually index tested and setup properly in order to achieve the highest possible operating efficiencies.
Figure 1 Kaplan runner on display outside Bonneville Dam
Figure 2 Cutaway view of wicket gates and runner section of Kaplan turbine
This cross section shows the arrangement of the wicket gates and the runner blades. At lower flow rates, the gates are mostly closed; water flow squirts through the gates into the runner section with a significant whirl imparted on it. The optimum blade angle is a function of both the volume of water flowing through the unit and the vector angle that is caused by the velocity of the whirling. When turbine efficiency is maximized, the whirl is completely removed from the water column in the discharge; the water flowing out of the turbine has no spin left.
Data from index testing is used to optimize, or tune-up the 3-D cam surface to maximize power and efficiency. This discussion compares real-world data from two index tests run one week apart on a 28MW Kaplan turbine at Clarence Cannon Dam in Missouri. One test used traditional index testing methods, and the other used the new Constant Power index testing method. Results of these two tests were virtually identical; both showing an approximate 3/4% efficiency increase at 28MW output would result from optimizing this turbine.
The design head of this unit is 75 feet “net-head.” Because the conventional test by USACE was an official acceptance test, the operating head for the traditional fixed-blades test was held at 75 feet net head utilizing the re-regulation dam 10 miles downriver. For the ITB Constant Power index test, head was at the seasonal norm of 82 feet. (The downstream neighbors didn’t like the higher water level that resulted from raising the tailwater to maintain the turbine design head, so we took what we could get.) As the tests were conducted, timing of the test points of the traditional method was made more difficult by the sloshing water levels that resulted from the large flow swings that are typical with the traditional test method. The Constant Power technique does not suffer this problem.
As different head levels become available as the seasons change, the index test should be repeated to define a new optimum cam line at the new head – but this is rarely, if ever done; index test results from a single head are typically extrapolated across the
entire family of heads shown in figure 3 and that’s the end of it. The ITB is an automatic instrument that is comprised of a low-cost COTS Personal Computer with a few software programs. Cost of an ITB installation is low, so it can be installed and left on the unit as a continuous performance monitor of turbine performance, collecting data year-round.
Figure 3 Standard Display of 28MW Kaplan 3-D Cam Surface Data
The American Society of Mechanical Engineers (ASME) Power Test Code #18 (PTC-18) defines index testing procedures to determine efficiency at a single operating point. Test definitions range from absolute-flow “code acceptable” tests to verify contract guarantees to the relative-flow “index test” that simply determines the optimum blade to gate relationship without measuring flow in absolute terms.
PTC-18 defines the method to collect and reduce a single data point. This procedure is repeated at several different gate strokes at a single blade angle to define a single point on the “best cam” line. This process is then repeated at a number of fixed blade angles across the full power spectrum of the turbine at a single head. The 3-D cam surface is then fully defined by repeating this process at a number of heads.
This Cartesian coordinate relationship of blade angle to head and gate stroke for this turbine are shown as the standard display of a Kaplan 3-D cam data surface in Cartesian coordinates in figure 3. This graph consists of a family of curves for seven different operating head levels with gate stroke as the X-axis, and blade angle as the Y-axis.
Wicket gate opening is the X-axis across the bottom, which may be expressed in several ways.
Gate Stroke is the linear extension of the servo piston that moves the bullring, which is the easiest to measure and control.
Wicket Gate Angle is the rotary angle of the gate shaft(s).
Gate Opening is the area of the opening between the wicket gates through which the water flows.
Wicket gates are designed so that the area of the gate opening increases linearly with piston stroke and gate rotation as closely as possible. In addition to flow increasing, the vector-angle of the water flowing into the runner section becomes more vertical as the gates are opened. To accommodate this steepening vector angle of the water flow, the blades are rotated to more steep angles as power level increases.
The purpose of index testing is to perfect and “fine tune” this angular relationship of blade angle to the flow vector angle.
Errors in the initial 3-D cam data surface for a Kaplan turbine arise from many sources. For model tests, data is taken from measurements on a scale model of the prototype (or full sized turbine) and scaled-up to predict the initial best-cam surface for the prototype. For model tests “net head” is generally used to determine the 3D-cam data map; prototype units generally use “gross head” as the input to the 3D Cam function. Model test rigs lack a trashrack and have smooth surfaces in the water passageways; the prototype does have a trashrack, and usually has a rougher surface finish, which increases losses due to the orifice drop of the trashrack and friction & drag from the passageways. The optimum cam profile for Kaplan turbines can also be altered by errors in scaling-up the model to the prototype in fabrication, manufacturing tolerance stack-ups and variations in the turbine setting at the dam.
Kaplan turbines should be immediately index tested when commissioned in order to correct for scaling errors and variations in the turbine setting. The model test surface that is initially installed in the turbine’s 3-D cam is only an approximation of the optimum cam surface for the full-sized prototype turbine. Only index testing can delineate the actual optimum surface and yield maximum turbine efficiency. It also behooves the powerplant to periodically retest their Kaplan turbines every 5 years or so because years of wear and tear on the machine will alter the actual optimum cam profile due to strain relief on the metal and cavitation & erosion on the gates & runner blade surfaces.
Index testing produces three main benefits:
1. The optimal 3-D cam surface profile is defined for the turbine blade control system to maximize operating efficiency from the unit; and
2. Benchmarks are created for subsequent index tests to be compared against, providing ‘trending’ information to help determine proper inspection and maintenance intervals - but only for an individual turbine; index test data from one turbine is not applicable to another.
3. Operating limits to define cavitation regions of operation are identified; staying out of these areas prolongs the interval between time consuming and expensive cavitation repairs.
When exact differences between the model test and prototype are defined and the prototype is optimized; the resulting performance increases can be significant. Index test results for one turbine are meaningless for another turbine, except only to note similarities.
For budgetary considerations, planning numbers for index testing a turbine that has never been previously index tested are for a 2%-3% increase in turbine efficiency. Actual results often achieve more than this range, and some less - but without actually testing every turbine it’s impossible to know for sure. Typical costs of index testing run in the $30k to $50k ballpark, depending on the degree of difficulty and extensiveness of the testing procedures. This is minimal considering the payback possible from optimization of hydroelectric turbines.
For example, here are planning numbers worked out to budgetary figures for index testing a Kaplan turbine. Assume a turbine is operated at 12MW output, 24 hours/day and 365 days/year. And then assume a 10-cents/kWh retail price for the added power output; know that all overhead expenses are already paid - so the increase from index testing will be pure profit.
· Starting with the most conservative estimate – a 1% efficiency increase will bring in an additional $105,120/year.
· A more ambitious estimate of a 5% efficiency increase would return $525,600/year.
· Not too shabby for a $50,000 tune-up.
The six following graphs present the data and results from a traditional “fixed-blades, moving-gates” index test first, and then for comparison the data and results from a “Constant Power swept-gates and blades” index test on the same 28MW Kaplan turbine run a week later.
(Notes: To allow a direct comparison of the results from these two test procedures, the X-axis scale of all three of these graphs are the same and to allow a more direct comparison, the X-axis for all plots is gate stroke, instead of power, which is typically used in the traditional method.)
Figure 4 Traditional Index Test points sweep across the On-Cam line
The traditional index testing method blocks the turbine runner blades at a constant angle while the gates are stepped or “indexed” to several positions across the On-Cam line with efficiency measurements at each step as shown in figure 4. Efficiency data is collected at each step to locate the optimum gate to blade relationship for optimum efficiency. This method is time consuming and required coordination of wide power output swings with the dispatcher and delays in order to wait for forebay and tailwater levels to settle out. Power and flow swings with gate stroke are shown in figure 5.
Figure 5 Traditional Index Test relative flow and generator output power
Using the PTC-18 data reduction techniques for each test point, a relative efficiency plot for a number of fixed-blade to gate pairs is created to determine the fixed-angle turbine propeller efficiency curve. The procedure is repeated for different blade settings to obtain sufficient efficiency data to define the optimum cam curve for a family of fixed-angle turbine propellers.
Field test data are dependant on existing head and tailwater availability. To fully map the 3-D cam surface, this procedure should be repeated at three or more different heads across the turbine design head range. But this is rarely possible because the range of heads doesn't normally occur within a relatively short time period.
Figure 6 Traditional Index Test relative efficiency profile
The classical method of index testing turbines is made more difficult, tedious, time-consuming and expensive due to repeated manual positioning of gates & blades, coordinating the resulting power level changes with the dispatcher or control room, and then waiting for the sloshing that results from varying the flow rates on forebay and tailwater levels.
Woodward Governor Company acquired a patent in 1987 (U.S. Patent 4,794,544) for the Automatic Index Test Box for Kaplan turbines with George H. Mittendorf Jr. and this author named as co-inventors, but was never utilized effectively – until now -the patent expired and the invention has come into the public domain.
The patent text describes a “Method and apparatus for automatically determining the set of optimal operating angles for the variable pitch blades of a Kaplan-type turbine which has moveable gates that are controlled by a governor, and an electronic 3-D cam” that works like this:
1. The unit is loaded up to the desired power level for testing.
(Note: All unit safety features and interlocks remain in force during this entire procedure.)
2. The blades are decoupled from the setpoint output from the head & gate to blade 3-D cam, allowing the ITB to provide the blade angle setpoint while the governor controls the gates to maintain the turbine at the power setpoint. (Decoupling the blades provides the benefit of more stable unit operation for the Constant Power test; the cascaded control algorithm typical in Kaplan governor system has a tendency to make the units more “wobbly”& unstable - and more difficult to get good steady state data.)
3. An initial “on cam” data set is captured; all data values are normalized to a common head using the ASME PTC-18 technique (ref: ASME PTC 18-2002, Hydraulic Turbines and Pump turbines, Performance Test Codes, paragraph 5,2,1, page 70). Both raw and normalized values are stored.
4. 2 The ITB moves the blades slightly off-cam, and then the governor resets power back to the load-setpoint by moving the gates. The ITB software SteadyState routine waits for the unit to settle out at each new operating point before capturing another data point.
5. Using these two degrees of freedom (shown in fig 7) the unit’s gates and blades transcribes a curve along the Constant Power axis of the turbine when the blades are swept across the “best cam” line and a load-feedback governor moves the gates to hold the power constant.
6. The power (red lines) and flow (blue lines) in figure 8 are relatively flat across the test range of gates (and blades) when compared to the traditional test methods plot shown in figure 5. Flow dips noticeably at peak turbine operating efficiency while power remains constant. Efficiency values of the collected data points are evaluated for a peak and two decreasing efficiency values on both sides of the on-cam line as shown in figure 9. When enough data is collected, the ITB releases the blades to normal operation.
Figure 7 Best-Cam lines for 80 and 85 feet and Constant Power Curve for 28MW Kaplan
Figure 7 shows the parabolic Constant Power (brown) curve for 55% power and the "80 & 85 ft Best Cam" lines (violet & orange) from the 3-D cam. These are all mechanical control parameters (blade vs. gate) of the turbine. The test is initiated with the unit operated at a point where the blades are open enough to provide room under the blade angle starting point for the below-cam test points.
The operating head at the start of this test was 82 feet, so the on-cam point would be approximately ½ way between the two best-cam lines shown on Fig 7. This example test started at 45% gate and 8% blade. After measuring efficiency data at the "on-cam" test point, the blades are stepped (or indexed) off cam in 3% increments on both sides while the load-feedback governor holds power and flow constant, or nearly so.
Figure 8 Relative Flow and Power Output as percentage
Figure 8 shows how closely the unit was held at a constant 55% power output, and maintained a constant relative flow of about 35% as the blades were swept from 2% to 20% and the gates were swept from 42% to 58% as the unit was swept along the Constant Power Line. The output power and flow of the unit remains quite constant all along this curve, allowing index testing to locate the optimum blade to gate point without disrupting power output or changing flow level. The constant power level allows testing without the necessity of changing power generation level every 5 minutes or so, alleviating the need for constant coordination of changing power levels with the dispatcher to conduct an index test. The constant flow level expedites the testing; during traditional index tests, delays result from waiting for the water levels to settle out after a flow change.
Figure 9 Relative Efficiency during Constant Power sweep
The optimum efficiency point from the Constant Power test is easily seen in Figure 9 as the point of highest relative efficiency. Relative efficiency is again computed using the conventional efficiency calculation methods documented in PTC-18. Kaplan turbine variable gate and blade geometry provides an infinite combination of gate and blade positions that will yield the desired power level, which makes the Constant Power method of testing possible.
Basic requirement for an ITB test is an electronic computer-based 3-D cam and load-feedback governor controlling the gates of a Kaplan turbine. Many vendors provide suitable equipment for this; an upgrade to a state-of-the-art computer-based control system is a prerequisite for an ITB Constant Power test; this does not rule-out index testing units with older mechanical systems with the Index Test Box.
If at all possible - it is recommended to start with a dewatered inspection to verify full stroke and linearity of gates and blades with protractors on wicket gate shafts and at the oil head for blade angle. The ITB has built-in utilities to make a few checks of steady state accuracy and dynamic robustness before proceeding with the index test. The ITB instrumentation inputs are calibrated with a simple “zero & span” procedure, with a few data points across the measured parameter’s span to verify the calibrations.
To check the design and condition of turbine gate and blade actuators the servo pressures are both recorded while the unit is in operation. This is to look for hysteresis and friction, and as a safety check to make certain the unit will return to a safe operating condition in the event of a catastrophic failure such as a linkage pin breaking off or loss of oil pressure.
Strip chart recording is provided to facilitate ASME PTC-29 deadband and deadtime performance testing; this recorder function can be started and stopped by internal triggering on collected data stream levels to make these tests easier to perform.
The ITB has it’s own 3-D cam routine. By entering the 3-D cam profile that is supposed to be in the unit under test into the ITB computer, the ITB will provide a reference blade value to verify the 3-D cam’s ideal blade setpoint values. The ITB’s StripChart and XY displays will verify the accuracy and robustness of the blade control and actuator, thus both steady state accuracy and dynamic robustness can be compared with industry standards quickly and easily with the ITB turbine control system checkout routines.
The ITB is currently being interfaced to the Quantum PLC computer for a North American Hydro Kaplan governor system in order to provide an index-testing accessory for their customer’s planned testing of three Kaplan bulb turbines and a fourth vertical Kaplan turbine. Test experiences and results will be published in upcoming issues of this magazine; so stay tuned,
Actuation Test Equipment Company
2. Voaden, G. H., "Index Testing of Hydraulic turbines," Transactions of the ASME, July 1951.
4. Kirkland, J. E. & Brice, T.A.," Checking Turbine Performance By Index Testing," Hydro Review, Winter 1986, page 49.
5. Hydraulic Turbines and Pump-Turbines, Performance Test Code PTC-18, American Society of Mechanical Engineers, 2002.
6. Speed Governing Systems for Hydraulic Turbine-Generator Units, Performance Test Code PTC-29, American Society of Mechanical Engineers, 2005.
7. IEEE Std-125, IEEE Recommended Practice for Preparation of Equipment Specifications for Speed-Governing of Hydraulic Turbines Intended to Drive Electric Generators, IEEE Power Engineering Society, 2007.
8. IEEE Std-1207, IEEE Guide for the Application of Turbine Governing Systems for Hydroelectric Generating Units, IEEE Power Engineering Society, 2004.
End of file