Monday, 12 September 2016


Guest Post Written by James L. Gordon, P. Eng. (Ret'd)

Marine clay, commonly known as “quick clay” and as “sensitive clay” by geotechnical engineers is clay deposited through salt water where the particles pick up salt, which alters the properties of the clay. The clay becomes “sensitive” since it has a propensity to liquefy when disturbed or saturated. It has been avoided by dam engineers due to its sensitivity.

Wikipedia defines marine clays as – “Marine clay is a type of clay found in coastal regions around the world. In the northern, de-glaciated regions, it can sometimes be quick clay which is notorious for being involved in landslides. Clay particles can self-assemble into various configurations, each with totally different properties. When clay is deposited in the ocean the presence of excess ions in seawater causes a loose, open structure of the clay particles to form, a process known as flocculation. Once stranded and dried by ancient changing ocean levels, this open framework means that such clay is open to water infiltration. Construction in marine clays thus presents a geotechnical engineering challenge.”

This explains the controversy surrounding the North Spur dam at Muskrat Falls since it will be the first hydro dam built on a quick clay foundation.

However, I have come to the reluctant conclusion that the North Spur dam is not safe and there is no easy way to make the dam safe. This conclusion has been arrived at by applying some logic to the situation as outlined in the following.

Until recently, dam design has been based entirely on precedent, since there was no mathematical procedure available to determine the safety factor. 

Dam design began in 1857 with a paper by Professor Rankine titled “On the stability of loose earth”. Since then hundreds of scientists working at many universities have investigated the properties of soils and rocks in an effort to arrive at a methodology for calculating a dam safety factor. A breakthrough did not happen until about 1955, when the concept of a slip circle on the downstream face was conceived by A. W. Bishop and a methodology developed to determine the stability of the slip. It became known as the “limit equilibrium method”. However, it took some time before the methodology became generally accepted.

Jim Gordon, P. Eng (Retired)
During this time, in 1968, I attended a short dam design course at Berkeley, where a procedure for determining dam safety was developed by Professor Seed. The process was to build a model of the dam on a shaking table, and gently shake it if there was no earthquake, and more violently if there was. If the dam slopes remained intact, then the dam was safe. However, the geotechnical community did not follow this development since there were issues associated with calculating the reduced size of rocks, gravel, sand and clay used in the model.

It was the use of computers that helped solve the problem. Calculating the stability of a slip circle was a laborious and tedious process, since many slip circles had to be investigated to arrive at the circle with the lowest factor of safety. The computer solved this problem, and with the use of more sophisticated programs such as FLAC designed to calculate the stability of many slip circles, the minimum factor of safety of 1.5 could be easily determined.

The safety factor is simply the ratio of the failure force divided by actual force. 

Perhaps best explained is in the case of a pipe, where the bursting pressure is divided by the operating pressure to obtain the safety factor.

But how was a factor of safety of 1.5 arrived at. It was the spectacular failure of the Teton Dam in 1976, which energised the geotechnical community to determine an acceptable safety factor. The Canadian Dam Association was founded in 1986, with the prime purpose of developing safety standards. This was accomplished in 1995 after many years of work by a dedicated group of geotechnical engineers. A safety factor of 1.5 under normal loading was considered to be acceptable, and 1.3 under an unusual loading such as an earthquake.

But why 1.5, why not a higher factor of safety. At 1.5 the dam factor of safety is the lowest of any other component in a hydro development. For example, the pipe carrying the water from the intake to the powerhouse has a factor of safety of 3.0 to the bursting pressure - the allowable maximum working stress being 1/3 of the ultimate stress. The wire rope that lifts the generator has a factor of safety ranging between 4 and 8. For iron castings it is 8.

The relatively low factor of safety in a dam reflects the geotechnical community’s confidence in the calculation methodology and determination of soil/rock properties based on over 170 years of research work. However, all this research has been undertaken on non-marine clay and other materials, hence the same safety factors cannot be applied to dams founded on marine clays.

Research into the properties of marine clays has only recently been undertaken by a few scientists, with the sole objective of determining the safety of quick clay deposits – are they liable to liquefy, and hence only be suitable for farming, or are they stable, allowing the construction of permanent buildings such as housing. Absolutely no research has been undertaken on the safety of dams built on quick clay deposits.

NALCOR’s engineers have developed a design for the North Spur dam using the FLAC program which has been shown to give incorrect results both by Dr. Bernander and Dr. Locat when used on marine clays.

In Dr. Bernander’s thesis he states (Page XXII) “Landslide hazards in long natural slopes of soft sensitive clays may – on a strict structure-mechanical basis – only be reliably dealt with in terms of progressive failure analysis. There exist, for instance, no fixed relationships between safety factors based on the conventional limit equilibrium concept and those defining risk of progressive failure formation. In consequence, the safety criteria have to be redefined for landslides in soft sensitive clays”. In other words - conventional safety factors are not applicable to sensitive clays.

And from Page XX in the Bernander thesis - Considering deformations and strain-softening in the assessment of slope stability normally results in a higher computed risk of slope failure than that emerging from the conventional ideal-plastic approach, depending in particular on the nature and the location of the applied additional load. In other words the FLAC approach currently used in conventional dam safety analysis is not correct.

In the abstract for the lecture titled “Spreads in Eastern Canadian Sensitive Clays” by Dr. Locat recently  presented at Memorial University, Dr. Locat states “Based on witnesses, spreads (landslides) generally occur rapidly, without any apparent warning sign, and cover large areas (> 1 ha). In addition, conventional stability analyses give too large safety factors when applied to this landslide type. Spreads are therefore serious threats to population and infrastructures on sensitive clays and the need for tools enabling their prediction and mitigation is quite necessary”. Again, regular stability analysis in not applicable to marine clays.

So, if the FLAC stability analysis cannot be used, can the old method of precedent be invoked, neglecting for the moment that there is no precedent for a dam founded on marine clays, and assuming the North Spur is founded on a soft non-marine clay.

Why soft clay? Over a month ago, I had the opportunity to discuss drilling on the North Spur undertaken in 2013 with the mechanic operating a vibrating drill rig. What he told me is not at all reassuring. He mentioned that on several occasions, the casing would slowly descend under its own weight. On other occasions it would drop suddenly by about 4 ft. On one memorable occasion, when the casing was left protruding some 20ft above the earth at the end of the shift, on returning the next morning, it had disappeared and was found some 20ft below ground. It had descended 40ft under its own weight overnight. All this indicates a soft to very soft foundation. Samples obtained from the drilling were placed in core boxes, now stored on site. The logs are also stored in NALCOR site files.

                            Figure 1 – North Spur modified downstream slope.

SNC and MWH have developed a design for the downstream slope of the North Spur with a 1:8 slope from El. 25m to water level at about El. 2m, which requires a horizontal slope length of about 184m. Above El 25.0m to berm at El. 40.0m, the slope is at 1:3, requiring a horizontal slope length of 45.0m. Above the berm at El. 40.0m and on to top of Spur, the slope is 1:2.5. With top of spur at about El. 64m, the horizontal distance from berm to top will be about 60m. Allowing for three berm thickness at 12m each, the total horizontal distance from the shore up to the crest is then about 326m. Source – Poster presentation by “Lower Churchill Project Geotechnical delivery team” October 28-30, 2013. All as illustrated in Figure 1.

At a total horizontal distance from shore to crest of 326m, the downstream work will cover over half of the Spur’s thickness of 570m at the narrowest section. Current work on the downstream face is shown in Figure 2. Normally, the upstream face of a dam has a flatter slope than the downstream face due to the lower friction from water lubrication of the particles. With over half of the Spur thickness taken up by the downstream slope, there is insufficient room for the flatter upstream slope.

Figure 2 – re-shaping work on the downstream face. August 2016.

A Google Earth view of the North Spur is shown in Figure 3, marked up to show the location of landslides. It is interesting to note that the large 1978 landslide extended upstream to the middle of the Spur, not a very reassuring development, since it could happen again.
Figure 3 - Source – North Spur Stabilization Works Progressive Failure Study Figure 1-1.
 Aerial photo of the North Spur.  SNC-Lavalin 21 Dec. 2015.

One of the simplest measures of the dam stability using the precedent analysis, is the ratio of base thickness to dam height. For a dam founded on rock, this ratio can be as low as 3.5, but the ratio increases rapidly as the foundation material becomes softer, and more so as the dam height increases. For the North Spur, with crest at El. 64.0m, and a base thickness of 570m, the ratio is 570/64 = 8.9. The question of a precedent then becomes – what is the ratio for a dam of similar height founded on a soft clay foundation.

The base thickness is the actual thickness of the dam at the contact with the foundation from upstream to downstream, as shown in Figure 4. For example, a 10m high dam on bedrock could have an upstream slope of 2:1 for a horizontal length of 20m. The downstream slope could be 1.5:1 for a downstream horizontal length of 15m. Neglecting the crest thickness, the base thickness is then 20+15 = 35m, and the thickness/height ratio is 35/10 = 3.5. For dams on softer materials, such as deep deposits of clay overlying bedrock, the side slopes are much flatter, resulting in higher thickness/height ratios. Also, as the height of the dam increases, the weight of the dam increases, requiring the slopes to become even flatter, again increasing the thickness/height ratio. 

 At Muskrat, the problem is even more complex due to the layers of quick clay within the body of the natural dam. Theoretically, this will require some further flattening of the slopes, but the effect has been neglected in the                    precedent analysis. 

            Figure 4. Thickness to height ratio.

For Muskrat, fortunately there is a precedent in the Gardiner Dam on the South Saskatchewan River in Saskatchewan. It is also 64m high and founded on soft clay. There the base thickness is 1,500m, for a base-height ratio of 1,500/64 = 23.4, or 2.6 times the North Spur ratio. Thus precedent indicates that the North Spur dam cannot be stable when founded on soft clay.

Figure 5 - Gardiner Dam. 64m high, base thickness 1,500m. Thickness/height ratio = 23.4

A photograph of the Gardiner Dam is shown in Figure 5, where the flat downstream slope is clearly evident. In fact, the slope is so flat, that it is rented to a local farmer as a hay field!

This analysis can be criticised as being incorrect, since the North Spur dam crest could be cut down to El 45.0m, requiring a much shorter base thickness, since the thickness ratio is also a function of the height for a dam on the same soft foundation. Fortunately, there is precedent for this in the Rafferty Dam, also in Saskatchewan, which  has a height of only 20m, and a base thickness of 278m, for a thickness/height ratio = 13.9. If it is assumed that the thickness/height ratio is a linear function of the height for dams on the same type of foundation, a 45m high dam on a soft clay foundation would require a thickness/height ratio = 19.3, for a base thickness of 869m. This is considerably wider than the Spur thickness, hence there is insufficient room for a 45m high dam with sufficiently flat side slopes to be stable.
This analysis has indicated that –            

1.         Dam stability analysis using conventional liquid equilibrium methods    cannot be applied to dams on marine clays.
2.        Safety factors developed for dams on non-marine clays cannot be
applied to dams on marine clays.
3.        There has been no research into the stability of dams founded on
marine clays.
        4.     The North Spur foundation consists of soft to very soft marine clay.
        5.     There is no precedent for a dam founded on a soft clay foundation with      
                the steep slopes shown for the North Spur, where the thickness to          
                height ratio is only 8.9.
        6.     Based on precedent, the thickness to height ratio for a 45m high dam        
                on the North Spur has to be at least 19.3 for a base thickness of about       
        7.     Based on precedent, the North Spur with a 570m thickness, has is              
                insufficient thickness to construct a 45m high dam with safe side       

Conclusion – the dam design developed for the North Spur is just not acceptable, and I hope that this analysis will be proved to be incorrect by a panel of international experts which should be convened immediately to resolve this issue.

- Jim Gordon, P. Eng. (Retired)

Editor's Note:
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.