Speaker: William Cook
University of New Brunswick
Topic: UNB's Flow-Assisted Corrosion Research
Location: J.L. Gray Centre
Deep River, Ont.
Date: Thursday, November 12, 2002 (8:00 pm)
William Cook photo

Summary published in North Renfrew Times, November 13, 2002:

Researcer Probes Problem of Premature Reactor Aging

by Jeremy Whitlock

Can you trust a cook that talks about feeders getting thinner?

On Thursday, November 7 about thirty people attended an evening talk on feeder-thinning by William Cook at the J.L. Gray Centre.

The title of the presentation was “UNB’s Program for Flow-Assisted Corrosion in CANDU Outlet Feeders”.

A CANDU-6 reactor uses 380 outlet feeders: carbon steel pipes, a couple of inches in diameter, carrying heavy water coolant at 310 degrees Celsius from the reactor core to the outlet header (and thence to the steam generator).

Around 1996 an inspection at the 13-year-old Pt. Lepreau station in New Brunswick revealed these pipes to be corroding from the inside (getting thinner) much quicker than expected.

The worst effects were seen near the reactor itself, particularly around the first bends. A phenomenon known as “Flow-Assisted Corrosion” (FAC) was suspected, whereby the impact of the water itself at high velocities can greatly accelerate any corrosion activity.

William Cook works for a lab at the University of New Brunswick that specializes in reactor coolant chemistry, under the leadership of former AECL scientist (and long-time Deep Riverite) Derek Lister.

The UNB lab was commissioned to help understand the phenomenon, with the aim of slowing it down. The lab’s assets included a high-temperature apparatus for simulating CANDU conditions, plus experience with on-line measurement probes.

As the corrosion progresses, iron is dissolved out of the carbon steel outlet feeders and deposited at cooler spots downstream, such as throughout the steam generators. This thins the feeders over time - a matter of some importance to a system pressurized to 100 atmospheres.

The UNB lab conducted a number of experiments, and discovered that a small addition of titanium to the coolant can slow the corrosive process. This unexpected finding swamped any benefits from their first hypothesis of reducing the pH of the coolant.

Subsequent experiments showed the therapeutic effects of titanium to be insensitive to coolant velocity, unlike the corrosive mechanism itself.

Mr. Cook then described a series bench-top experiments with plaster of Paris molds, carried out by a visiting French graduate student with a penchant for extreme Canadian adventure. The results suggest that the scallop-shaped pockmarks in the pipe’s inner wall are initiated by defects in the metal.

In response to one of the questions from the audience, Mr. Cook described how new CANDU reactors will likely be built with feeders made from steel with higher chrome content - a modification shown by AECL to eliminate the problem. AECL is also considering fitting out its new ACR models with stainless steel feeders.

As for current reactors, the industry is investigating the use of continuous titanium injection in the primary cooling system, but the process will be a lengthy one since the effects on all other components, including fuel, have to be well understood.