Case Study
Case Study
Case Study
Dominion Bridge and Sulzer partnered to introduce a low head horizontal Kaplan “S”
type turbine into the North American market in the 1970’s. These units were intended
for generating capacities up to about 10 MW. It appears that there are somewhere in
the vicinity of 150 units in service in Canada and the United States of America.
These units perform well generally, but many owners have experienced short life spans
and unreliable performance of the turbine runner water seal. This is a costly problem as
the unit needs to be de-watered in order to change out the seal. As well, some of the
components are long lead-time items creating extra costs associated with inventory of
on-hand spare parts.
Hydro Tech was called to the Dolgeville Generating Station in January 2018 to review a
history of seal failures that occurred from 1987 through 2017. The objective was to see
if there was an engineering solution available to keep seals functioning for a reasonable
time frame before needing to be replaced.
The S-Turbine turbine currently operating at the Dolgeville site was selected in 1984
and commissioned in 1987. The sacrificial seal element for the runner hub seal was
asbestos based. The original seal material had a modest life expectancy that needed
improvements. When health and environmental concerns that were followed by strict
government regulations surrounding use of asbestos were introduced, it was easy to
make the decision to change materials to see if life expectancy could be improved. This
horizontal water S-t Turbine has seen numerous resin-composite style seals installed
only to fail prematurely due to overheating and extensive pre-mature wear. Asbestos
has high temperature resistance and on start-up it is tolerant of the lubrication system
breaking down temporarily, so it was not unreasonable to expect a reduced life when
introducing alternative composite products. These compound changes to address the
asbestos issue did not provide the desired improvement on life expectancy.
Underlying Issues
The designers relied upon lubricating and cooling the seal face by extracting penstock
water. While economical in the first instance, any blockage of flow would cause the seal
element to contact the thrust face, heat up and wear prematurely. River silt would create
grinding action between the contact faces and could plug up the passageways.
Figure 2 - Original
Water ports to positively lubricate the seal were not designed into the stationary bearing
housing as space for this was not readily available. The original engineering process
relied on large clearances between the runner shaft and the outer seal ring, thus
allowing river water to pass over the runner shaft flange, under the seal ring, across the
seal face, and into the drain. Under start-up conditions, and during episodes with high
particle counts in the water, one can see how quickly the seal contact face can lose
lubrication and cooling.
Another factor in designing hydraulically actuated seals of this nature is the
maintenance of consistent contact pressure at the seal face. Seal pressure in this
design was achieved by tapping into the penstock to utilize the river head pressure.
This pressurized water activates the piston that applies pressure to the seal element
This penstock water was passing through a small duplex strainer. It is worth noting that
air entrained water is a compressible fluid and is normally not used for hydraulic
Construction Faults
Imagine this part having water sit stagnant in the O-ring groove and on the outer
surfaces where the piston slides. Corrosion will always be the enemy here, eating away
at the O-ring lands and binding up the movement of the piston.
Dowel
Also, noteworthy, once the seal material started to wear, the material started to
gall the seal steel face material. One should note that the damaged originally
polished steel is now acting as a grinder on the seal.
Photo 5 – Gap and sharp edges act as a scraper on the seal element
The unit had been overhauled recently and experienced two seal failures during
commissioning. The metal seal components showed evidence of damage sufficiently
great enough that repairs and replacement of these components were in order.
The original flange was made from a German stainless steel. The flange material had
been changed over time and was found to be 316L stainless steel. The piston was
originally carbon steel and had been changed to 316 stainless steel. The outer seal
cover was a cast steel of an unknown European grade.
L/D material is basically a cotton type yarn passed through resin and wound on a
mandrel.
In addition, in 2017 another manufacturer provided one or two seals made from a
polymer base substance that failed during commissioning.
CUSTOMER REQUIREMENTS
As the repairs to this power plant were being closely monitored, some restrictions were
placed on Hydro Tech in coming up with a solution. Also, some historical
documentation from the seal manufacturer, the previous repair team and an alternate
seal manufacturer were presented to aid us in coming up with a workable solution.
The supplied solution shall be reversible so the unit can be put back to original.
A vacuum seems to be present on start-up and shut down causing the seal to run
dry.
Work with Bestobell to come up with a solution they agree with.
Inspect and witness the repairs being attempted on the piston, inner and outer
seal covers and 2-part flange with the onus on re-using these parts if possible.
Minimize any changes to the bearing housing.
Keep the chamber inside the seal from being flooded as this compromises the oil
seal on the runner stub shaft.
Use the existing control system, modifying it as needed.
Best industrial practices outside of the hydro power industry use centripetal forces to
send lubricating fluids across the seal face from the inside of the seal to the outside.
Positive spring pressure keeps the moving element in enough contact with the mating
surface to prevent excessive flow, while keeping boundary lubrication present. In the
original design this does not happen; the water comes from the outside to the inside and
relies on water contained in a hydraulic cylinder to pressurize the seal for reduction of
water flow.
Seal face theory suggests a hardness difference between materials, with the harder
material surface finish polished to work successfully. Usually the hard surface is fixed,
and the soft surface is rotating. There are exceptions to this and conditions on these
exceptions. The split line must be invisible to the seal element or the edges dig in,
shearing the seal material and pushing away the lubrication.
The cooling and lubricating fluids should not contain anything that can act as a grinding
compound as this affects the performance of the seal boundary. Silty river water can
act like sandpaper, quickly compromising the performance of the seal face.
These engineering principles were not accounted for in the original design of the seal,
and retrofits around the original assumptions became hard to accomplish.
ENGINEERED SOLUTIONS
The necessity to achieve water flow across the seal became the priority we needed to
address first. Bestobell provided the flow rate and we tapped into a city water line to
achieve this rate. We set up pressure and flow monitors on this for verification and tied
it into the original control system.
We connected the city water supply to the pressure sensing port on the bearing
housing. This port previously monitored the draft tube water pressure in the distributor
housing.
We installed a turbovent to provide air into the distributor housing which would relieve
any vacuum pressure that could occur during start-up or shut down. We set up the draft
tube pressure gauge on this piping.
The turbovent was designed as a 3” vent line. As this line may be subjected to a full
vacuum (30” HG or up to 30 psi), it was fitted with a shut-off ball valve, a sinking ball
check valve, a silencer and a strainer basket. Its purpose was to feed station air into the
water stream during start-up automatically and on demand.
Photo 11: Draft Tube PSI
It was noted that when the seal wears and the piston is extended, it closes off the
opening to the old pressure port. As we were using this port to provide water, we
modified the outer seal ring by cutting in a slot and closing the clearances by adding a
bronze band. In hindsight, it turns out that when the piston closed over this port, when in
original factory condition, a vacuum was occurring that was being read as negative
penstock pressure.
It was proposed that the original manufacturer’s seal element would be installed. The
entire set of seal components was measured. Then the clearance, fit, and finishes were
put back to the 1980’s drawing specifications. The seal piston controls were repaired,
and the float switch system put back into working condition. Some compromises were
made so components could be salvaged and avoid replacement costs.
We verified the dimensions of the seal components and made some adjustments to fit
tolerances and surface finishes in the machine shops. Then we fitted the seal
components on the unit as best we could back to original specifications. The outer seal
cover had been damaged during the last seal failure and an attempt to fix it with a weld
repair (prior to Hydro Tech’s involvement) had left the outer seal cover in a distorted
condition. We attempted a repair to this cover when we added the bronze band. No
attempts were made to change any of the additional mechanical issues as they had not
revealed themselves as causes of seal failure up to this point.
A thorough study of the startup procedure was carried out. During the life of this unit
some design changes to seal controls were made. This complex control system
provided a method to maintain water pressure to the back face of the seal piston and a
rudimentary float switch system used to monitor water discharge from the drain of the
casing. Both components had feedback to the turbine control system and alarms to
indicate functionality.
Clean water flow across the seal face was achieved. In order to verify this, we installed
flow meters and pressure gauges on the piston water supply and on the seal face water
supply.
With tremendous support from the EONY personnel, we began to rebuild the seal
components to get this unit online.
On 3 May 2018 the seal assembly work was completed, and the unit commissioned for
start-up. Within 4 days (7 May 2018), the seal had worn down to 3 mm remaining and
the turbine shut down.
.
Photo 17: Worn L/D seal element in piston
A failure Investigation began on 14 May 2018 and the resulting changes were made:
The 304 stainless steel flange ring was replaced with 400 series stainless steel,
heat treated to give better wear properties.
We replaced the Bestobell element with a natural fiber wood (Lignum Vitae), which
has much better properties for sealing on rotary faces and complements the choice
of 420 or 410 stainless steel. Lignum vitae is a wood that has natural lubricating
properties. While Lignum Vitae should always be kept wet, it is able to self-
lubricate should water lubrication and cooling conditions be marginal.
Photo 20: Segmented Lignum Vitae seal element with bronze backing rings
The turbovent provided the evidence that ruled out vacuum conditions.
We fitted the Lignum Vitae seal element into a 304 stainless steel piston.
Hydro Tech manufactured a new outer seal cover with additional split line fasteners
so the conical end would not open under pressure and changed the material to
400 series stainless steel.
Hydro Tech possesses an in-house material testing bench for measuring wear under
both wet and dry conditions. This bench was manufactured using ASTM Standards for
testing material properties and the test equipment itself meets ASTM requirements.
Results:
LD 24358
0.600000
0.500000 Dolgeville condition
Wear rate cc/24hr x10‐3
0.400000
0.300000
0.200000
0.100000
0.000000
0.0 100.0 200.0 300.0 400.0
P/V Ratio metric
0.500000
Dolgeville condition
Wear rate cc/24hr x10‐3
0.400000
0.300000
0.200000
0.100000
0.000000
0.0 50.0 100.0 150.0 200.0 250.0 300.0 350.0 400.0
P/V Ratio metric
The results indicate that as the P/V value goes up, the LV seal outperforms the L/D
material. In flooded conditions, under the P/V value the seal element is set up to run at
Dolgeville, the wear characteristics favour the LV material only marginally.
We then ran the same tests dry to see what happens under no-water conditions.
Direction of axial
profilometer readings
Axial Surface Finish Rz: (Na) Axial Surface Finish Rz: 0.29 µm – 0.37
µm
L/D Sample (Ring LV Sample (Ring L/D Sample (Ring LV Sample (Ring
#3) #2) #3) #2)
Axial Surface Finish Rz: 2.68 µm – 2.90 Axial Surface Finish Rz: 0.33 µm – 0.39
µm µm
L/D Sample (Ring LV Sample (Ring L/D Sample (Ring LV Sample (Ring
#4) #1) #4) #1)
These limited tests indicate that Lignum Vitae is much less abrasive to shaft journals
then an equivalent material under dry conditions. One possible explanation might be
that the L/D samples have been observed to produce large quantities of fine particulate,
which may result in increased abrasive wear on the collars. Another explanation is the
natural lubricating properties of Lignum Vitae.
Photo 24: Worn LD material ring #4 Photo 25: Worn L/V material under ring #1
The results of these tests supported our view that a lignum vitae seal element
would work well in this application.
These tests supported the manufacturer’s position that L/D material needs to be
continuously lubricated and that damages can occur to the stainless-steel
counter-face.
IMPLEMENTATION
This addressed most of the mechanical issues and metallurgical issues thought to
contribute to seal failure.
Reversing water flow was ruled out as this would have required changes that would not
be reversible.
Photo 28: Flange ring Water Surfaces after 1400 hours operating time.
CONCLUSIONS
This paper has outlined the process of improving a turbine shaft water seal to allow us
to offer solutions that have eluded power plant owners for decades. Lignum Vitae has
proved to be a tough and reliable selection to solve marginal lubrication problems in
horizontal Kaplan water turbines.
Start and stop episodes pose little difficulty in survival of this seal element and should
prove over time the desirable life expectancies needed.
About Us
Hydro Tech Inc. has offices at key border points between Canada and the USA,
in Sault Ste Marie, ON and Niagara Falls, ON.
US operations are conducted from Hydro Tech USA Inc. located in Brunswick,
Maine.
For the past 18 years, Hydro Tech has been designing innovative large thrust
bearings for the hydro Industry.
Hydro Tech is bringing its extensive knowledge of water lubricated bearings to
correct and improve seal life in the hydraulic turbine sector using sound fluid
dynamics and engineering principles.
Hydro Tech also specializes in turbine and generator overhauls and mechanical
design improvements for turbines and generator equipment, gate installations,
and seals as described in this paper and sources.
Authors Credentials:
The seal OD (working surface) is 33 11/16" and the ID is 31 5/16" nominal. This
is the thrust face contact area. Seal nominal size 32 5/8” Dia. x 1 7/8” x 1 3/8”