Canada

Mitigating Operational Effects of AAR at Otto Holden

When alkali-aggregate reactions began impeding the performance of equipment at Ontario Power Generation’s 240-MW Otto Holden Generating Station, the utility performed a comprehensive study to determine a solution for the operational challenges that can be associated with AAR.
Tags

By Paramjeet Khoral, Ahmed Hafez, Ed Hong and Dehai Zhao

This article has been evaluated and edited in accordance with reviews conducted by two or more professionals who have relevant expertise. These peer reviewers judge manuscripts for technical accuracy, usefulness, and overall importance within the hydroelectric industry.

The effects of concrete movement and growth, a result of alkali aggregate reaction (AAR), have been observed since the early 1970s at the 240-MW Otto Holden Generating Station. Major overhauls were conducted on all eight units from 1980 to 1988, and the effects of concrete movement were noticed during turbine and generator component alignments.

Otto Holden Generating Station – located on the Ottawa River, near Mattawa, Ontario, Canada – has been operating since 1952. The powerhouse has eight generating units with a total capacity of 240 MW under a gross head of 77 feet. All eight units have the same Westinghouse generators. Units G1 to G4 are Allis Chalmers turbines and G5 to G8 are John Inglis turbines. It is owned and operated by Ontario Power Generation Inc. (OPG).

Identifying problems

The units underwent their first set of major overhauls from 1980 to 1988. During the stationary alignment of the units it was noticed that the concrete embedded components were not aligned anymore, and in order to monitor this a program to collect turbine clearance data was begun. OPG has since collected runner clearance data for all eight units and used it to calculate the rate and direction of unit component relative movements. A concrete movement investigation study was conducted by OPG in 1997, and petrographic analysis of concrete samples collected by Material & Petrographic Research G-B Inc. in 1998 proved AAR in the concrete. Multiple instruments were also installed in 1998 at various locations in the dam to monitor concrete movement, providing data for the calculation of the concrete growth rate and strains in different parts of the dam.

Unit G8 encountered a major operational problem of sticky wicket gates in 1999. This initiated a second set of major overhauls and turbine and generator component modifications on all eight units from 2000 to 2008.The turbine bottom rings were made adjustable to compensate for relative horizontal movement between the top and bottom speed rings.

In September 2013, a technical team comprising OPG’s Engineering and Technical Services (ETS) and Eastern Operations (EO) personnel was formed to perform detailed assessment of the turbines to find the current condition of the units and to prepare a long-term operation and maintenance strategy.

Finite element analysis of stay ring and head cover

Most of the turbine and generator components can be moved for unit alignment purposes with some modifications. However, the stay ring is embedded in the concrete, and movement due to AAR distorts the ring, both in vertical and horizontal directions.

However vertical offset is difficult to measure because there are no reference points. Horizontal offset is generating large bending stresses at the joints of the stay vanes with the top and bottom speed rings. The seal weld joint between the stay vane and speed ring has failed on almost all units and was repaired during the second set of overhauls. The structural integrity of the stay vane and its joint with the speed rings is a major source of concern.

The Otto Holden Generating Station is located on the Ottawa River and has been in operation since the early 1950s.
The Otto Holden Generating Station is located on the Ottawa River and has been in operation since the early 1950s.

Ideally, the turbine head cover and bottom ring should be concentric. Or, in other words, the top speed ring and bottom speed ring should be concentric. At Otto Holden, however, the horizontal offset between the top and bottom speed rings varies between 0.150 inch to 0.479 inch. OPG compensated for this large amount of offset by installing the head cover eccentrically opposite to the AAR-induced direction of movement.

The head cover’s locating spigot was machined from 0.060 inch to 0.330 inch and its hold-down bolt holes were slotted by up to 2.25 inches during the second set of overhauls. To compensate for the horizontal offset generated since the unit overhauls, further machining of the head cover is required and the bolt holes may have to be opened to the outer edge of the flange. These modifications can impact the structural integrity of the head covers.

OPG’s Machine Dynamics and Component Integrity (MDCI) group assessed the structural integrity of the stay ring and head cover. MDCI conducted the finite element analysis (FEA) of both as per the boundary conditions and load scenarios supplied by ET engineers. Separate FEAs were performed for each of the two sets of turbine designs.

FEA of the concrete dam structure

To understand the effects of AAR and non-linear behavior of the materials, a detailed FEA model of the complete concrete dam structure for one unit and the embedded stay ring was prepared. This analysis was conducted by OPG’s Dams & Structures Department.

Original station drawings were used for preparing the concrete structure model. However, the stay ring model geometry was simplified to keep the analysis problem to a manageable size.

Non-linear material properties were used for concrete and the stay ring. Two boundary conditions were used in the concrete model:

  1. It was restrained at the foundation level both in the horizontal and vertical directions.
  2. East and west sides of the concrete structure were restrained in the horizontal direction – longitudinal direction of the dam.

In addition to gravity load on the structure, unit weights were applied at the lower bracket and stator sole plate elevations as distributed loads. Hydrostatic loads were applied in the upstream and downstream direction. Thermal load (temperature gradient) was applied to the concrete structure to simulate the concrete growth due to AAR.

Analysis of the concrete FEA model was calibrated using the total horizontal offset value of 0.400 inch between the top and bottom speed rings for G1. This horizontal offset in the FEA model was achieved by applying a temperature gradient of 500 degrees Celsius.

The calculated horizontal G1 stay ring offset by the FEA model was 0.430 inch. The calculated total strain (by extrapolation) from the concrete instrumentation data was 4545.6Με , matching nicely with the applied thermal strain of 5000μΕ to the FEA model.

Analysis of concrete FEA results

In the previous studies, only the relative movement between the top and bottom speed rings was estimated from actual measurements. However, there was no information about the absolute movement of the rings. The results of this FEA show that absolute movement of the top speed ring is in the northwest direction and of the bottom speed ring is in the southwest direction. The total relative movement of the top speed ring with respect to the bottom speed ring is 0.430 inch in the southwest direction, which matches with G1 measured results.

The FEA calculated direction of relative movement is 14 degrees in the southwest direction and it matches with the measured direction of relative movement of 17.4 degrees in the southwest direction.

Another significant revelation was that the stay vanes are in compression. This is completely opposite from all previous studies, which concluded that stay vane and anchor bolts are under high tensile stresses and may be close to failure.

Vertical displacement of the stay vane measured by the FEA was 0.193 inch. This is the result of the top speed ring being pushed downward by 0.098 inch and the bottom speed ring being pushed upward by 0.094 inch. The geometry of the stay ring was simplified in the FEA model to limit the size of the analysis problem.

The vertical deformation calculated by the FEA may not be an accurate representation of field conditions. In reality, this large vertical displacement is not possible because the stay vane will buckle and collapse under vertical displacement of 0.132 inch, as shown by the buckling analysis of stay vanes in the FEA report. This kind of vertical movement will also seize the wicket gates.

In reality, localized plastic deformation of the top and bottom speed rings close to the stay vane joints, cracking of seal welds between the stay vanes and speed rings, and cracking of the concrete re-distributes stresses and avoids any buckling failure of the stay vanes.

Plumbness and vertical height of the stay vanes should be measured in the field to determine the amount of distortion due to AAR.

Engineering a solution for the stay ring problem

From the analysis and measurements described previously, it can be seen that the component affected most by AAR-induced distortions is the concrete embedded stay ring. In 1999, the horizontal offset of 0.441 inch between the top and bottom speed rings led to wicket gate operational problems.

The stay ring offset led to off plumb wicket gates that were rubbing the head cover, and the bottom ring and seized around a 50% gate position. This led to the second set of unit overhauls.

The current stay ring horizontal offset varies from 0.370 inch to 0.479 inch for the four worst units (G1, G6, G7, and G8). However, they still have not encountered the wicket gate operational problems due to the modification done during the last set of unit overhauls.

As discussed previously, stay ring stresses are increasing due to horizontal and vertical relative movement between the top and bottom speed rings. It is hard to predict the exact amount of these stresses accurately, but it can be deduced based on horizontal offset and AAR-induced compression on the stay ring that stresses are high and are a combination of high bending stresses and compressive stresses.

The actual stresses may be lower than the stresses calculated by the stay ring and concrete structure FEA, due to plastic deformation of the stay ring, cracking of seal weld joints between the stay vane and speed rings, and cracking of the concrete behind the stay ring.

The bottom line is that with 0.400 inch or greater horizontal offset, the stresses at the joint of the stay vanes and the speed rings are high and will continue to increase with concrete growth.

This will lead to eventual failure of the joints or stay vanes and may lead to cascading failure of the stay vanes and flooding of the pit liner and the station. Stay ring condition has to be improved using one of the following methods.

Dealing with the stay vanes

Several options were considered to deal with the stay vanes, which impact the speed ring:

Option 1: Cut stay vanes in the middle, only one at any given time, to provide elastic recovery at the joint of the stay vane with the speed rings. Depending on the orientation of the stay vane with respect to the horizontal offset of the speed rings, the offset between the cut stay vane halves can be as high as the total offset of the speed rings. The offset between the cut stay vane halves will also depend on the amount of elastic recovery at the speed ring and stay vane joint. After the elastic recovery, the stay vanes can be welded back and will also require hydraulic profiling at the new welded joint`. The amount of hydraulic profiling will depend on the orientation of the stay vane with respect to the direction of the horizontal offset of the speed rings.

This method will reduce stresses at the speed ring and stay vane joint but will not reduce the horizontal offset between the top and bottom speed rings. It will also require eccentric machining of the speed rings and head cover and will also require new head covers at a certain point.

Option 2: Cut stay vanes at two places, 6 inches from their joints with speed rings, and cut one stay vane only at any given time. Short (6 inch) stay vane stubs will be left close to the top and bottom speed ring and a large portion of the stay vane will be removed. A new piece of pre-machined stay vane can be installed and welded to the stubs. This method will require less grinding for hydraulic profiling but will double the amount of work required for cutting the stay vanes and their welding.

This method will reduce stresses at the joint of the stay vane and speed rings by elastic recovery but will not reduce the horizontal offset between the top and bottom speed rings. It will also require eccentric machining of the speed rings and head cover and new head covers at a certain point.

Option 3: Cut all stay vanes in the middle so that the top and bottom halves can move freely, independent of each other. As in Options 1 and 2, the stay vanes will be welded back after the elastic recovery of stay vanes and they will also require hydraulic profiling.

The disadvantage of this method is that the horizontal offset between the top and bottom speed rings may increase once the stay ring resistance is removed. This will make the situation worse for the head covers and new head covers may be required.

Freeing up the top speed ring

The top speed ring can be freed up by removing all the concrete behind it, cutting all anchor bars and removing all the bolts from the top speed ring connection with the pit liner.

Once this ring is freed up, there will be an elastic recovery in the stay ring that will reduce the stay ring horizontal offset and stresses in the stay vanes and at the joint of the stay vanes with the speed rings. Reduced offset will eliminate the need for eccentric machining of the head covers or any need for new head covers.

This method was used in G8 in 2000 during the last overhaul. It reduced the horizontal offset from 0.441 inch to 0.185 inch. Assuming that we get similar results in the other units, technically, it is the best and recommended solution for the stay ring problem.

Conclusion and recommendations

OPG used its analysis to determine several courses of action:

This figure shows the proposed unit overhaul schedule based on a number of factors determined during the units’ analyses.
This figure shows the proposed unit overhaul schedule based on a number of factors determined during the units’ analyses.

1. An overhaul schedule (see Figure 1 on page 20) was established based on several factors:

  • Horizontal offset of the stay ring;
  • Rate of relative movement between the top and bottom speed rings;
  • Room for bottom ring adjustments;
  • Vertical setting of the runner;
  • Wicket gate clearances and operation; and
  • Any other unit operational issues.

Units with horizontal stay ring offset of higher than 0.400 inch should be overhauled to avoid any operational problems and to relieve excessive stresses in the stay vanes.

G2, G3, G4, and G5 will not require overhauls until the 2030s, and their schedule depends on the future bottom ring adjustments and the ability to correct the vertical setting of the runner.

The proposed schedule for bottom ring adjustments will assist in delaying unit overhauls, allowing them to operate in a safe manner.
The proposed schedule for bottom ring adjustments will assist in delaying unit overhauls, allowing them to operate in a safe manner.

2. The bottom ring adjustment program (see Figure 2 on page 20) will assist in delaying unit overhauls and allow them to continue operating safely.

G6, G7, and G8 do not have room for further adjustments of the bottom ring. These units will have to be realigned by unit overhauls. Bottom ring adjustment aligns the head cover and the bottom ring and improves wicket gate plumbness, clearances and operation. However, it does not reduce the horizontal offset in the stay ring or its stresses.

The schedule to adjust runner vertical settings was established to avoid damage to runners and wicket gates.
The schedule to adjust runner vertical settings was established to avoid damage to runners and wicket gates.

3. The schedule for adjusting the vertical settings on the runners (see Figure 3 on page 20) will help avoid damage to runners and wicket gates.

This work will be accomplished by placing a shim between the generator and turbine shaft coupling during annual inspection outages. During the unit overhauls, the shim can also be placed between the runner and the turbine shaft coupling.

4. To extend working life of the embedded stay ring, its stresses and horizontal offset should be reduced during the future unit overhauls. This can be achieved by freeing the top speed ring by removal of concrete behind it, cutting of all top speed ring anchor bars and removal of bolts between the pit liner and ring. Freeing the top speed ring will allow for self-adjustment/aligning of the stay ring due to elastic recovery of the offset. This ensures long operational life due to:

  • Reduces the horizontal offset between the top and bottom speed rings and improves unit alignment. It is expected that we can recover up to two-thirds of the offset by freeing up the stay ring.
  • Reduces stresses in the stay vanes and their joints and anchor bars.
  • Eliminates the requirement of any more offset machining of the head covers, hence any need for new head covers.

5. The condition of the head covers is good. By reducing the stay ring offset as recommended above, head covers will not require any further offset machining or further slotting of the head cover mounting bolt holes. Head covers can be used as is with slight machining. There is no need for new head covers.

6. The other major work during the overhauls will be as follows:

  • Head covers will have to be installed eccentrically within the top speed ring as far as physically possible in the direction opposite to the relative direction of head cover movement due to AAR. A crown seal in the head cover will set the unit centerline. The bottom ring, turbine guide bearing, generator guide bearing and stator will be statically aligned to this centerline.
  • Generator sole plates were modified during the last set of overhauls. The sole plate pucks will have to be machined for leveling and the stator will have to be moved laterally for static alignment and re- dowelled.
  • Generator lower bracket will have to leveled and moved laterally for unit alignment.
  • Head cover and bottom ring seats will have to be machined for leveling.
  • Install shims between the generator and turbine shaft coupling to adjust the vertical setting of the runner.
  • Install new wicket gate bushings.

7. Concrete distortion varies largely between summer peaks (August and September) and winter troughs (February and March) in a given year. Peak-to-peak variation can be as high as 3.5 mm, although the measured concrete growth rate is in the order of 0.023 inch/year. The large variation in the runner clearance readings based on season can be eliminated by conducting turbine inspections in the same month.

8. Long-term unit inspection schedules should be prepared to cancel out seasonal effects on concrete growth and turbine deformation.

9. A shaft vibration monitoring system should be installed and data should be analyzed on a continual basis for condition monitoring of the generator, bearings and unit alignment.

10. More ongoing attention is required for monitoring the unit conditions, on the basis of turbine clearance data, concrete instrumentation data and shaft vibration instrumentation data. All data should be collected and maintained in one location.

Unit overhauls with top speed ring concrete modifications are scheduled to begin in 2018, with an anticipated cost of C$4 million per unit.

Editor’s Note: This article was adapted from a paper submitted for the HydroVision International 2016 conference program. The full paper is available at http://bit.ly/HRNov16.

Bios: Paramjeet Khoral is a senior mechanical engineer with Ontario Power Generation’s Power Equipment department. Ahmed Hafez is a senior structural engineer with OPG’s Dams and Structures department. Ed Hong is the station manager at Otto Holden. Dehai Zhao is a senior plant engineer for OPG.