Affects of Ultra-Pure Water on Mechanical Seals

Presented By:Pat Prom

Prepared By:James SaucermanArt Olson

FlowserveFlow Solutions Group

Summary:

• Background Information

• Current Test Programs:

-Reactor Recirculation Pumps

-Reactor Feedwater Pumps

The Solution

Background

The Phenomena – Field Observations

•Under certain operating conditions and environment a specific kind of damage occurs to both silicon carbide and tungsten carbide.

•Certain areas on a silicon carbide seal face experience pitting/chipping and material loss. In tungsten carbide the result also exhibits pitting and flow channels.

BACKGROUND INFORMATION

BACKGROUND INFORMATION(Field Results – Spectrum Analysis of SiC)

BACKGROUND INFORMATION(Field Results –Cross-Section Tungsten Carbide)

Nickel binder removed leaving voids in material

Original condition of tungsten carbide with nickel binder present

BACKGROUND INFORMATION

Where do we encounter this problem?

1. Nuclear primary reactor recirculation pumps Increasing number of incidents in BWR plants worldwide

2. Reactor Feed Water Pumps

3. Fossil fueled plants worldwide using COT (Combined Oxygen

Treatment) More conventional plants going to ultra-pure water are increasing

the number of seals experiencing this problem

BACKGROUND INFORMATION

What has been identified as a contributing factor?

•Purity of process fluid (Ultra-pure water)•Friction•Piezoelectric properties of silicon carbide or tungsten carbide•High circumferential speeds

BACKGROUND INFORMATION

How these factors interact…. Theoretical Explanation

•Seal face friction generates heat and charged particles on seal surfaces.

•In the presence of low conductivity water the electrical charges can not easily dissipate and charge builds on SiC or Wc seal face.

•When sufficient charge accumulates the stored energy jumps or travels to a conducting surface; resulting in the damage to the outer diameter of the rotating face.

•Discharging results in vaporized regions at the sub-micron level which grow over time (SiC) or loss of binder in Wc

Common denominator in all these applications:

Extremely low conductivity of the water

Electrical conductivity of water types:

Ultra pure water: 0.055 µS/cm

Distilled water: 0.3 – 5.0 µS/cm

Potable water: 30 – 1500 µS/cm

Sea water: 50000 µS/cm

BACKGROUND INFORMATION

Current Test Programs

BACKGROUND INFORMATION

CURRENT TEST PROGRAM: (Nuclear)

Operating Conditions:

•Seal: 9” balance diameter•Material Selections:

•Tungsten Carbide vs. Carbon• SiC w/laser treatment vs. Carbon

•Test Fluid: •Distilled water (20 – 65 µS/cm) •Ultra-pure water (< 0.15 µS/cm)

•Fluid Temperature: 75 - 160° F•Speed: 430 – 1780 rpm•Test Duration: 200 hours / test

Test Plan:

•3 Rapid Pressure Transient Cycles:

15 – 1025 psig

(2 hours)•2 Rapid Temperature Transients:

60 – 160° F

(1 hour) •Various operating speeds:

•Slow Roll Speed: 430 rpm

(1st 100 hours of test)•Full Speed 1780 rpm

(2nd 100 hours of test)

CURRENT TEST PROGRAM: (Nuclear)

CURRENT TEST PROGRAM: (Nuclear)

Results (Tungsten Carbide vs. Carbon):

Process fluid:

20 – 65 µS/cm

Process fluid:

0.065 – 0.12 µS/cm

Electro-corrosion

Test Results: 200 hours Process fluid: <0.15 µS/cm

Field Inspection

CURRENT TEST PROGRAM: (Nuclear)

CURRENT TEST PROGRAM: (Reactor Feedwater Pump)Operating Conditions:

•Test Fluid: •Distilled water (20 – 30 µS/cm) •Ultra-pure water (< 0.15 µS/cm)

•Fluid Temperature: 125° F•Speed: 5500 rpm•Test Duration: 168 hours•Flow Rate: 5 gpm

SiC

WC

AISI 316 Sleeves

CURRENT TEST PROGRAM: (Reactor Feedwater)

The Solution

THE SOLUTION: Precision Face Topography w/Laser Treatment

Low amplitude waves:

Improves hydrodynamic stability and increases load support

Self-Cleaning Design:

Expels suspended solids in process fluid

Seal dam:

Maintains low leakage

Full Laser Treatment:

Changes material and electrical properties

THE SOLUTION: Precision Face Topography w/Laser Treatment

Standard lapped SiC face Lasered processed SiC face

TEST RESULTS: (Fossil Fuel)

Results (Silicon Carbide w/Laser Treatment vs. Carbon):

Original Design after 168 hours

Proposed Design after 430 hours

TEST RESULTS: (Nuclear)

Results (Silicon Carbide w/Laser Treatment vs. Carbon):

Process fluid:

0.075 – 0.14 µS/cm

Process fluid:

20 – 73 µS/cm

CONCLUSIONS: Precision Face Topography w/Laser Treatment

•Electro-corrosion (EC) does not occur above 20 µS/cm

process fluid; field experience shows above 15 µS/cm •Field experience and tests show that the purer the water

the more aggressive the damage becomes•The use of hydro-dynamic features + post laser

processing of surface eliminates EC in ultra-pure water

applications•Combination of hard-soft seal faces are forgiving for wider

range of operating conditions and allows incidental contact

at smaller fluid film thickness (lower leakage)

CONCLUSIONS: Precision Face Topography w/Laser Treatment

•Matching electrical conductivity of sealing surface pair will

minimize the build-up of electrical charge, minimizing the

potential between the two surfaces•Precision Face Topography w/laser treatment provides

cooler running face temperatures which minimizes

likelihood of electro-corrosion

FUTURE WORK: Precision Face Topography w/Laser Treatment

•Perform long term testing on steady state nominal

operation tests for both reactor recirculation and Reactor

Feed Water pump seals