A non-technical primer on dam design.
Or why the North Spur dam design needs to be reviewed.
Guest Post by Jim Gordon. P. Eng. (Retired)
In view of the controversy over the North Spur dam design, I have prepared the following brief review of dam design over the ages, to show why the North Spur dam design needs to be reviewed.
Dams have been built for more than 5,000 years, with the first recorded being the Jawa dam in Jordan built about 3,000 BC. Another was the Sadd al-Kafara dam near Cairo, built about 2,850BC. It had a height of 11m and a crest length of 81m.
The highest ancient dam was built by the Romans in the first century at Subiaco, near Rome, having a height of 50m.
Most ancient dams were built of masonry, and most were destroyed by floods. However, a few have survived and are still used. The Proserpina dam is still suppying water to the Town of Merida, in Spain. The Romans introduced arch and buttress dams.
Ancient embankment dams were mostly small seasonal structures for storing irrigation water, usually swept away in the next flood. One significant exception was the Wadi Dhana dam built about 750BC to a height of 4m. It was rebuilt about 500BC to a height of 7m, with a crest length of 610m, and again in 115BC with the height increased to 14m, and with a 5-gated spillway in the left abutment.
The largest dam in North America in 1832 was the Jones Falls dam in the Rideau River; it is still used. The first concrete arch dam was built at Warwick in Australia, with a height of 10m.
Embankment dam design did not really start until Professor Rankine produced a paper “On the stability of loose earth” in 1857. Since then, research has concentrated on understanding the principles of embankment dam design, and is still continuing to this day.
Despite such research there have been some spectacular failures, with the most prominent being the Teton embankment Dam in Idaho which failed on 5th June 1976 due to water flowing through the rock on the right abutment eroding the clay core. It has not been rebuilt. It prompted the USA government to introduce legislation on dam safety and inspection, with a similar effect in Canada.
The Donana dam in Spain failed due to sliding on a weak clay foundation in April 1999. The Big Bay dam in Mississippi failed in 2004 due to piping (water eroded a hole through the clay core). The Hope Mills dam in North Carolina failed in 2010 due to a sinkhole, and recently the Mount Polley dam in BC failed on 4th August 2014 due to the undetected presence of a weak clay layer 10m below the foundation.
Despite our extensive knowledge about the behaviour of embankment dams, failures still occur.
Since the Rankine paper, most research has concentrated on determining the characteristics and strength of the various materials such as clay, gravel, sand and rock used in an embankment dam. What is the effect of pore pressure? What gradation is required in filters to prevent the particles migrating into coarser material? What is the water gradient through a dam? How much compaction is required? And with clay, there are many forms, over-consolidated, soft, sensitive and of course marine clay - what are their distinguishing properties and characteristics?
Clay research has concentrated on areas where more knowledge is required. Marine clay has long been known for its propensity to liquefy, and consequently has never been used as a dam material, nor has a dam been built on marine clay foundation. Hence there has been little or no research into the properties of marine clays.
Embankment dams could not be constructed on rivers with a deep permeable foundation until about 1960, when vertical cut-off walls were introduced. They are based on oil well technology where a bentonite beneficiated slurry is used to stabilize the walls. Their stability analysis was not possible until Dr. Morgenstern co-authored a paper titled “The stability of a slurry trench in cohesionless soils” in 1965. Since then, many have been built.
The early walls often had permeable holes since gravel could be sloughed off the vertical wall by the equipment to fall into the trench during the filling process, leaving a layer of gravel within the backfill concrete. The slough was undetectable since the volume of concrete was still the same, filling instead the space from the slough. One cut-off wall I worked on was found to have so many permeable holes that their equivalent diameter totaled over 6m! Obviously, seepage was far higher than expected.
It is only recently that an instrument has been developed to detect sloughs. It is dropped down the excavation just before backfilling and a sonic survey produces a three dimensional drawing of the excavation. From this, the volume of concrete can be precisely calculated for each meter depth, and checked during backfilling. Incidentally, it is not known whether such an instrument was used for the North Spur cut-off walls. If not, they likely contain permeable holes.
Over the last few decades there have been many marine clay liquefactions in Sweden, resulting in loss of life. This prompted the government to ask Skanska, the largest international construction company in Sweden, to determine the safety of the marine clay deposits. Unfortunately it was not possible to determine their safety due to a lack research into their properties. Their conclusion was not to build any permanent structures on marine clay.
At the time, Mr. Bernander was their chief engineer. When he retired, he returned to university to undertake research into the properties of marine clay and to develop a methodology for their safety assessment. After 2 years of intensive research, he developed a 6-step series of mathematical equations which can be used to assess their safety. In his thesis, which I have read twice, he states three times that the common methodology used to determine an embankment dam safety factor cannot be used on marine clays – it gives an incorrect answer.
Current practice in assessing embankment dam safety relies on a finite difference computer program called FLAC (don’t ask – Google FLAC ITASCA CONSULTING for a demo). This program has been used by the NALCOR consultant (Reference – NALCOR presentation 21 July 2014, slide 38), and by their Independent Engineer (personal communication), to determine the safety factor of the North Spur dam, containing marine clay, and built on a marine clay foundation.
In view of this, and since Dr. Bernander has warned that the common methodology for dam safety assessment cannot be used for marine clays, Roberta Benefiel and I called Dr. Bernander in mid-December to specifically ask him if FLAC could be used. Dr. Bernander replied that he was familiar with FLAC and had tried to use it in the early stages of his investigations with no success, the FLAC answer was not correct.
After he had completed his research, he returned to FLAC to determine whether it could be used, and found that with some modification, requiring a 2-step run of FLAC, that it could be used. On enquiring as to why he was not using FLAC – his reply was that his own methodology was far faster and less cumbersome.
So, down to the bottom line – the current design of the North Spur dam safety factor has been determined by the FLAC program which, according to Dr. Bernander, who has extensive experience with the behaviour of marine clays, gives incorrect results.
Hence the design of the North Spur dam needs to be reviewed by a geotechnical engineer familiar with Dr. Bernander’s methodology.
Jim Gordon has authored or co-authored 90 papers and 44 articles on a large variety of subjects ranging from submergence at intakes to powerhouse concrete volume, cavitation in turbines, generator inertia and costing of hydropower projects. He has worked on 113 hydro projects, six of which received awards "for excellence in design" by the Association of Consulting Engineers of Canada. He was also awarded the Rickey Gold Medal (1989) by the American Society of Civil Engineers "for outstanding contributions to the advancement of hydroelectric engineering...". As an independent consultant, his work assignments have ranged from investigating turbine foundation micro-movements to acting on review boards for major Canadian utilities. He has also developed software for RETScreen and HydroHelp.