Calgary
is a city of almost 900,000 people and one of the fastest-growing in North
America ; we expect to reach 1 million before 2005. We have 4000km of
water mains, the majority of them relatively young, since half our growth
has been since 1970. An unusual feature is the proportion that
is PVC, now the single largest component of our system. Another is that
the majority of our ductile iron pipe is YDI, yellow-jacket pipe that is especially
well protected against corrosion, both with coating and cathodic protection.
In
consequence, our break-rate has now been stable for some years at only 500
breaks/year, 400 of them main breaks. Roughly 60-70% of them are related to
external corrosion. We have reduced our replacement program repeatedly
in recent years, and now feel that we have come through the other side of
a 25-year problem that started in the early 70's when our main break rate
started to climb alarmingly. After a few years of it, we started to
increase our replacement program, and continued those increases until just
a few years ago, with total added costs in this period in the high tens
of millions - a lot of money even in Canadian dollars. We had
a serious concern that the break rate would spiral out of control as we had
a great deal of pipe all installed during short "boom" periods that could
conceivably have reached the end of its life all at once. A water utility similar in size and age to ours, East Bay M.U.D., had a similar increase in breaks during the early 70's, and a consequent decrease after they also began cathodic protection strategies. This similar-looking chart was presented at the AWWA Infrastructure Conference in Baltimore in March 2000. It had to be removed from the presentation due to time limitations. It does indicate that our experience is not unique.
The easy thing to get out of my database is not the full history of what pipe
was installed originally, but what is still in the system. In the case
of the PVC and YDI lines, it's the same thing, because almost none has ever
been replaced. I've left CI out of this picture because I want to focus
on our decisions in the 70's that caused our evolution through different pipe
materials, and CI predates all that. The installation histogram of PDI,
poly-wrapped ductile iron, would have originally been twice as tall; we've
pulled out a couple of hundred kilometres of the original 700 or more that
went in. Most of it is PDI, the bare DI is the much smaller curve.
We started off installing PDI and DI in about equal amounts, but shifted entirely
to PDI in the early 70's as corrosion became a concern, while also trying
out AC, asbestos cement, which gave us a lot of trouble with "workmanship-related"
breaks, and various other DI coatings, like Taped-Urethane DI, and YUDI, yellow-jacketed
urethane. But the protection we switched to aggressively was YDI, Yellow-Jacket
Ductile Iron, because the great spike in main break rates in the early 70's
was largely because of corrosion problems and Calgary was expanding into areas
that had much more corrosive soil. We still use YDI, in any location
where there is a concern about hydrocarbons ever having been in the soil.
But after just 5 years of YDI, we switched again to PVC as the material of
choice, and that continues today. 
This is a
map of Calgary from my point of view: just the water mains, and colour-coded
by material type. It shows what I call the growth rings of our city
very plainly, with the dark blue as the thick-wall cast iron from 1900 to
1955, thin-wall cast in magenta from 55-68, and the light blue of DI from
68 to 78. I lump all the DI together except the YDI, which is
orange, and then the PVC is the green in the outermost ring. The Grey
is the colour for "other" much of which is concrete feedermains, and our brief
flirtation with AC down in the south end in the early 70's. Calgary is kind of the Canadian Denver, with metric instead of US units: if you are the mile-high city, we are the kilometre high city. We are right at the start of the foothills of the Rockies, varying by over 1000 feet in elevation from the nearly-level plains on the east side, to the two hills that dominate the northwest. The major growth area in the 70's was in the flat area in the northeast, where there had formerly been a lot of ponds and sloughs and the soil was far more corrosive than anything we'd encountered previously. I'd like to zoom in on that area.
The northeast is a good showcase for the colour-code system because it has a little of everything, from the thick and thin-wall cast of the previous growth ring to all the materials we went through in the 70's. My apologies to those with trouble distinguishing blue from green. The blue DI and the green PVC are very intermingled in this area, not only because of original installation, but because so much of the DI and PDI had to be replaced. Where you see little bits of blue DI colour in the centre here, understand that those blocks are the survivors of when it was all DI. Some of it even had to be replaced during our YDI years in the late 70's, when the DI was just 7 to 10 years old.
You can see why when I superimpose the map of all the mainbreaks the area
has had since 1960, as red dots. All of the dots on the green PVC happened
to the CI or DI pipe that the PVC replaced. Some of the YDI is
replacements, but most is original. As you can see, YDI does have some
breaks, and a few blocks out here in the far east end are becoming a concern.
This version highlights where the replacement has been by turning from red
to white any break that is on pipe now replaced. A lot of that expensive
15 and 20 mile per year replacement program in the late 80's and early 90's
was in this area. But note also, the blocks of DI that not only
survive, but to this day, have very few breaks. We've recently run a
Hydroscope through some of those mains, and a few barely even have pits, though
they are just a few hundred metres from blocks that had to be replaced very
young. Corrosion is highly dependent on local soils. To look at
the very worst locality, I'd like to zoom in further to the south end of this
picture, about six square miles of the system that was a mixture of CI and
DI. 
I've
also turned all the breaks back to red. I'd like to use this area
to show a time-series of what so alarmed us in the 1970's into frantically
researching and developing cathodic protection and new mains materials. 
First, I'll remove the visual distraction of the pipes, I just want to show
the break patterns. 
Next,
I'll remove all the breaks except those that happened in the 1970's.
In 1970, the northeastern-most quarter of this area was not yet built.
The south-most block or two was older CI, the rest was DI. 
In 1970 and
1971, there were a few more breaks in this area than the city average, but
the pattern was not especially unusual for the city norm. There were
a few blocks that had a couple of breaks per block per year. 
Adding in
the 1972-73 breaks in orange, there is already a trend of increase visible,
and a concentration of much of the increase onto a few blocks that were getting
too many complaints to be left in the ground. We also started to see
some repeat breaks in the DI blocks further north, mains that were just four
and five years old. 
I'm just
going to switch colours at random with each slide to highlight the added breaks.
1974-75 was the turning point for us. The trend continued to worsen.
It wasn't just that our break rate had gone up 50% in five years, it
was that much of the increase was concentrated in one area where the
public was understandably up in arms - and that a lot of the main affected
was very young. We did understand it was the corrosive soil and not
anything different about the DI from CI, but still it was plain that Something
Had To Be Done. The Chief Engineer of the time, Jim Bouck, eventually became the manager and my first boss, and I interviewed him in retirement. The first-blush judgment was that they were having about as much trouble with PDI as with bare DI, and though they continued to use polywrap, they were almost frantic to find something better. The PVC industry and the developers were already pushing for PVC, but there was no AWWA spec for it at the time, and Jim wanted to wait for that blessing before committing to it. In 1973, that was still five years away, so we turned to YDI as a stopgap.
We could
at least take solace by 1976 that we had started installing YDI and begun
a cathodic protection program requiring an anode install at every repair site,
and for 1976 and 1977 the break rate overall held about even, but it continued
to concentrate in this area, with new breaks appearing in the DI areas at
the north end of it that were only a few years old. 
And 1978-79
were the worst years of all, as the break rate, helped along by some cold
winters, skyrocketed from 800 to 1200 per year. It was around this time that
we pulled out a few pipes that were less than 10 years old. If
there had been any doubt that we had to get a grip on corrosion control, these
years put an end to it. 
It was also at this time that the AWWA C900 spec was approved and we began
the switchover from YDI to PVC that offered the hope of an end to metallic
corrosion altogether. We were tentative with PVC at first, gradually ramping
up it's use after a few years of experiments and some lessons learned about
installation technique. YDI was installed in at least equal amounts for four
more years. 
A few words about YDI for those blessed with such non-corrosive soils that
you aren't familiar with it. In contrast to PDI, which is wrapped in an 8-mil
poly sheet out in the field and not bonded, The Yellow-Jacket of YDI is a
40-mil high-density polyethylene that is strongly bonded to the pipe. 
As
this photo of some removed jacket beside a pipe sample shows, it is stiff,
tough stuff. This pipe coating had been used in the oil industry since the
sixties. The coating was so complete they cathodically protect kilometres
of pipe with each anode bed. We get about 100m of protection out of each magnesium
anode in contrast to 10m with bare DI, because there are so few spots that
can leak electrons. 
As
this closeup shows, the pipe is also protected against internal corrosion with
a cementitious lining. It was 1973, with our corrosion problems just becoming
apparent that Jim Bouck began lobbying - I won't say "threatening" - our pipe
suppliers about giving us YDI. The problem was that the pipe used by the oil
industry had no bells, it was all welded. A way to pull the coating over a bell
of water pipe had never been developed at the factories in Western Canada. 
But pressed directly by our Chief Engineer, developers, and CANRON, the DI
supplier, Calgary's Shaw Pipe redeveloped their coating manufacture to pull
the yellow jacket over a bell, and to install one of these straps with a bubble
of gel coating over the strap weld. 
After manufacture and deployment began in Calgary in 1975, Canron and
Shaw also sold a lot of YDI to other Canadian municipalities. However, Jim
Bouck actually counselled many smaller towns to not bother with the extra
expense of YDI - not if they couldn't install it properly, or couldn't run
a consistent cathodic protection maintenance program afterwards. The YDI shown here is ready to go into the trench. You'll note it's resting on tires to prevent any coating scratches. Before it goes in, we jeep it again in the field - use a high voltage to check it for any breaks in the coating. We use sand, pea gravel, granulite or other clean fills for bedding and the first half-metre of the pipe zone. Then we check the current flow at the anode test points forever, reinstalling the anodes every 20 years. It's a lot of trouble.
We find it, however, well worth it -- the break rates are a fraction of those
on any other metallic pipe - just over a tenth. So far, of 600 km, we've never
replaced any for corrosion-related causes, and only a few hundred metres for
any reason. You'll note that I've just lumped together PDI and DI together on one line of this chart. In 1978, Jim Bouck and other Calgary Engineers attended this same conference and described their problems and their YDI approach, and said that PDI had no discernably different break rate from bare DI. That got them a lot of hot replies and raised voices. DIPRA, the ductile iron pipe research association, was not pleased.
Well, a quarter century later, I'm prepared to retract some of that assertion
on behalf of my predecessors. We not only have data from another 25 years of break rates, but data from sandblasting thousands of metres of recovered mains and measuring pit depths. This graph follows the break rates for the DI and PDI that is still in the system today, so the breaks are all from the same cohort of pipe.
The pattern is clearer if you smooth out the yearly ups & downs with
a 5-year moving average. The PDI has a somewhat lesser break rate, though
the degree seems to vary quite a bit. Then I realized its curve was offset
sideways from the DI curve - and naturally, for most of it was years younger.
So I requeried the database for the break rates by age of pipe at the time
of break... 
...and again had to smooth it with a 5-year moving average. Now the difference
can be seen to be consistent over time. It is in very good agreement with
an independent number from a completely different study. We measured corrosion
rates on all kinds of our pipe by recovering samples and measuring maximum
pit depths per metre. From the two studies, we can say PDI offered us
about a 30% average reduction in corrosion rate, and in break rate. 
So this DI line could be split into two lines a bit apart. 30% is nice, but
I'll take the near-90% reduction we get with YDI, even at the expense and
trouble.
The
breaks we do get on YDI are mostly fitting and service leaks; corrosion has
little to do with a rubber gasket coming loose or leakage around a service
connection that grows from erosion-corrosion. But some YDI has had simple
corrosion holes and corrosion-related circular breaks. This colourful graph
I invented to show the results of Hydroscope runs - that's a so-called "smart
pig" that you can pull through your mains with a cable, detecting holes
and pits with electromagnetic fields. The Yellow line at the bottom is the
pipe itself, and the ticks on it are the several breaks it has had, this being
one of the blocks I mentioned above. The corresponding graph in blue is the
pipe wall, the blue bands indicating the 20%, 40%, and so on remaining wall
thicknesses. The things that look like mine shafts are individual pits the
Hydroscope detected. We believe the problem on this main was that electical
shorts occured across from the main to the copper services, letting the current
from the anodes drain away so it wasn't protected. Which goes to prove to
me that absolutely no coating - not even one five times as thick as polywrap
and glued to the pipe - will protect iron unless accompanied by anodes. Jim
Bouck was perfectly correct about that.You'll note that so far, I haven't mentioned PVC breaks. If we're getting about 200 breaks per year on 500km of DI, and 30 on 600km of YDI, what's the rate on PVC?
In 1990, we had about a third the PVC we have now; we've gone from 500km then
to 1400 now. And we've had 17 breaks on all of it over those ten years. They
all fit on this one slide. The top eight, in blue, are actual repairs to PVC that had no visible interference, though construction errors like bending it or placing it on a large rock were usually involved.
The middle one was a leak between PVC and PDI where the clamp between them was not on right.
And the bottom eight are all related to service taps; either they pulled out, or a crack in the PVC spread out from the tap site. You can never have perfect workmanship. The number 17 glosses over that we've probably had at least 17 more breaks on PVC during construction of it or near it. There's no question that it's lower strength does render it more vulnerable to construction breakage from tamping too hard, or from shifting the soil it is in during nearby excavations. But, really, that's quibbling about whether it has 1% or 2% of the breaks that we've had from metallic mains.
In Baltimore, a couple of American engineers expressed surprise to me at our
success with PVC; they had thought that trouble with taps was very common.
But as you see, we have less than one PVC problem at services per year when
we have 100,000 installed. I'll tell you our secret in three words: training,
training, training. This is a real source of pride for Calgary Waterworks,
our pipeman training facility. It occupies two large truck bays. The
structure in the background was built for it, to allow simulation of the above
ground parts of the system. 
On the floor, are the underground parts: examples of virtually every
combination of materials and fittings in our system. 
In an intensive two-week course, every pipeman works repeatedly with
all of them, practicing procedures for construction and repair. 
Of
particular interest, are problems involving mixtures of materials, since we
have so many. 
But, back to the PVC, here's the heart of the operation: veteran pipeman trainer
John McConnell, one of the most exacting detail men you could hope to learn
from. He's setting up for my lesson in tapping PVC under pressure on a practice
pipe that really is under 100 psi. 
PVC takes at least five minutes longer to tap than DI, and John teaches great
care in getting out every thread of plastic from the hole when you remove
the coupon. 
After the maincock has been inserted a few threads, the whole tapping machine
can be removed and the maincock turned home with a torque wrench. It's important
for long life of the plastic to not wrench it in too hard. 
With use of teflon tape around the maincock threads, however, it rarely comes
to that. A few of the stronger guys in the field do the final turns with their
bare hands to show off - it takes about 10 foot-pounds. 
But John is a bear on the subject, checking every tap with a ruler to see
that the pipe bulges out around the tap by only a few millimetres, and he
worries about that, though it is well within the spec.
| Cost of bare 6" DI, 18ft: | $207 |
| Final Cost of that as YDI: | $308 |
| 2000 Cost of 6" PVC, 20ft: | $138 ($88 bulk) |
| 2001 Cost of 6" PVC, 20ft: | $138 ($125 bulk) |
I can imagine people blanching at all this care, taking over 10 minutes to do a tap. But PVC has paid off for us long before we get the payback of almost no breaks.
It's cheaper as a material cost, though much of that gain over bare DI may be lost depending on your costs for clean sand bedding and pea gravel in the pipe-zone.
| Pipes Replaced per Crew-Day | YDI | 5-8 |
| PVC | 8-12 |
But even then, the fact that it is so much lighter to move around and so much easier to cut means that on the average, we can put it in nearly twice as fast as YDI.
In short, we've never looked back from going PVC.