Guiding seismic safety

Gibe 3 RCC gravity dam in Ethiopia

The Committee on Seismic Aspects of Dam Design is one of ICOLD’s oldest technical committees, which at present comprises dam and earthquake experts from 34 countries. The Committee was created in 1968 and celebrated its 50th anniversary in 2018. Guidelines on (i) different seismic hazards affecting storage dams, such as fault movements in the footprint of dams, and reservoir-triggered seismicity, (ii) seismic design criteria, (iii) dynamic analysis of dams, (iv) conceptual guidelines for earthquake-resistant design of dams and design of appurtenant structures, and (v) seismic monitoring and inspection of dams after earthquakes, have been published. These guidelines, listed below, represent the international state-of-the-art in the seismic design, construction and safety assessment of large storage dams, i.e. 

Among these guidelines, Bulletin 137 provides information on reservoir-triggered seismicity (RTS), a hazard unique to large storage dams, which is often a key dam safety argument brought forward by NGOs and people opposing new dams. For properly designed and constructed dams, RTS is not a new safety concern. 

However, the publication with the greatest long-term impact on the seismic design of new dams and the safety assessment of existing dams is Bulletin 148. Two earthquake ground motion levels are specified for dams and safety-critical elements, which are the Operating Basis Earthquake (OBE) and the Safety Evaluation Earthquake (SEE). The safety-critical elements are the spillway and low level outlets, which are needed to control the reservoir level after the SEE or for lowering the reservoir for repair works or for increasing the safety of a dam. This is new. Valley slopes and dam abutments may also be classified as safety-critical, if landslides, rockslides or rockfalls can block intakes of spillways or low-level outlets, or damage equipment needed for the operation of spillway gates or low-level outlets. In addition, abutment movements may damage arch dams and earthquake-triggered mass movements may create impulse waves in reservoirs. Therefore, these slopes must also be checked for the SEE. 

Bulletin 120 complements Bulletin 148 as it includes conceptual features for the seismic design of dams, which are extremely important, as it is well known that it will be difficult to have a structure to perform well during an earthquake, when the basic seismic design concepts are not followed.

Hirakud earthfill dam in India

Designing against earthquakes

Dams were the first structures designed against earthquakes, on a worldwide basis, starting in the 1930s. At that time, the ground shaking was the main seismic hazard and was represented by a seismic coefficient of typically 0.1, almost irrespective of the seismic hazard at the dam site, which was often unknown. The seismic analysis was done with the pseudostatic method, ignoring the dynamic characteristics of dams. Because of its simplicity, this method is still in use today, although it has become clear that this method is obsolete following the observations made during the 1971 San Fernando earthquake. The pseudostatic method is also not compatible with current seismic guidelines (Bulletin 148) and, therefore, this obsolete method shall no longer be used for the seismic safety checks of large storage dams. Using the pseudostatic concept, the seismic load case was rarely controlling the design of dams. This has changed by using today’s rational concepts for seismic hazard analyses and dynamic analyses of dams. The earthquake load case has become the dominant one for most dams.

Since the formation of the seismic committee, the magnitude 8 Wenchuan earthquake of May 12, 2008 in Sichuan province, China, was the most important earthquake for dam engineers as it damaged some 1580 dams. Most of them were small earth dams, but also some large dams were damaged. The main lesson from this earthquake was that the seismic hazard is a multi-hazard. Thousands of mass movements occurred in the epicentral region. Mass movements that can be triggered by strong earthquakes are often ignored or the hazard is assessed using criteria, which are different from those used by dam engineers. Based on the past experience, it is obvious that dams are not inherently safe and can be damaged by strong earthquakes. The most vulnerable dams are those, which are poorly constructed and/or designed. Still a lot of work is required in order to ensure that all dams comply with modern seismic safety criteria, which is the main concern of the seismic committee.

Ituango rockfill dam in Colombia

Fifty years

The main developments may be summarised by the following conceptual changes in the seismic design of new dams and safety evaluation of existing dams:

• From pseudostatic analysis to dynamic seismic analysis of dams.
• From the representation of the earthquake ground shaking by a seismic coefficient to safety evaluation earthquake ground motion parameters.
• From consideration of single hazard to multiple seismic hazards including ground shaking, mass movements, and faulting. 
• From stability safety factors and allowable stresses to rational seismic performance criteria, characterised by dam deformations and seismic failure modes of dams.

There are many dams that have been built without taking into account earthquakes or which were designed against earthquakes using the pseudostatic analysis method. Therefore, it is not known if these existing dams satisfy today‘s seismic design and safety criteria, which are the same for old and new dams. As there should not be any difference in the safety of people living below an old or new dam, all dams must ultimately the same seismic safety criteria.

Neelum Jhelum hydropower project in Kashmir, Pakistan

Modern performance criteria?

If we use modern seismic design criteria for large dams (ICOLD, 2016), the following, very general, performance criteria apply for the effects of the strongest ground motion at a dam site:

• Retain the reservoir and to protect people from the catastrophic release of water from the reservoir.
• Control the reservoir level after an earthquake as a dam could be overtopped and destroyed if the inflowing water into the reservoir cannot be released through damaged spillways or low-level outlets.
• Lower the reservoir level after an earthquake (i) for repair works or (ii) for increasing the safety of a damaged dam or when there are doubts about the safety of a dam.

These seismic performance criteria are different from those used in the past, when a dam was declared safe, when for different load combinations including static and seismic loads, the stresses were within the allowable stresses, the deformations were within allowable deformations and the safety factors against sliding, overturning and others were larger than the safety factors specified in design guidelines. This concept has been used in the past and is still being used by some engineers today.

These new seismic performance criteria have far-reaching consequences, which go beyond the tasks of dam engineers – mainly civil engineers -, because functionality of gates of spillways and low-level outlets is the main task of hydro-mechanical and electro-mechanical engineers, who may not be aware of these new requirements. Therefore, in future, there is a need to have a broader look at the seismic safety of dams and to include the functionality of safety-critical hydro-mechanical and electro-mechanical components. There is a need to adjust the design guidelines for hydro-mechanical and electro-mechanical components of spillways and low-level outlets. These components must be designed for the SEE ground motions at the support of these components.

Rather than looking backwards, which is done above where the transition from old concepts and methods towards new ones is listed, it is better to look into the future. Anyway, the present state is described by the current ICOLD Bulletins, which include the new developments.

Punt dal Gall arch dam at Swiss-Italian border 

What has been achieved?

At the Annual ICOLD Meeting in Antalya, Turkey, in September 1999, I was appointed Chairman of the Committee on Seismic Aspects of Dam Design, and just a few weeks ago I celebrated my 20th year as Chairman of this important ICOLD committee. During this long period I also had the chance to work with eight ICOLD presidents. Evenin my first meeting I stressed the importance of the seismic safety of existing dams as most seismic design guidelines are concerned with new dams. The seismic safety of the existing dams is still a top priority as there is a need to know if these dams satisfy today‘s safety criteria, so appropriate action can be taken, if this is not the case. As the economic life of any dam is only as long as it is technically safe and functional, it is obvious that we must know if it is safe.

The seismic safety assessment of a dam includes the following: 

• Seismic hazard analysis.
• Selection of seismic design criteria.
• Modelling of dam-reservoir-foundation systems.
• Determination of material models and dynamic material properties.
• Methods of nonlinear seismic analysis.
• Definition of seismic performance criteria. 

All of the above vary with time. Therefore, it is necessary to review the seismic safety assessment periodically. It is obvious that in a comprehensive safety review all other hazards must be included as well. This is the main task for the future. 

This periodic review concept is not new, but it is very useful and should be implemented by all dam owners and dam safety authorities. By this concept, effects of the widely discussed climate change on dam safety can also be assessed. If the safety criteria are not satisfied, then remedial actions have to be taken. Such detailed review should be done every 5 years, as for example in Switzerland. In practice, a reanalysis of the seismic safety may only be needed when the bases of the analysis have changed significantly, which may be in the time frame of 20 to 40 years – not every five years.

Zarema May Day asphalt core rockfill dam in Ethiopia

What is expected in the future?

 In future new developments may be expected in different areas that will affect the seismic analysis and design of new dams and the safety assessment of existing dams. In the assessment of future developments, the current ICOLD guidelines documented in several bulletins serve as a benchmark. As these guidelines, which represent the state-of-the-practice have not yet been implemented by all dam owners or dam safety authorities, the first steps in the future will be to follow the recommendations made in the current ICOLD guidelines. Moreover, the seismic safety standards used in some countries may be ahead of that of ICOLD and what is considered as new or future development may not be the case for everybody. It is also important to note that the future development does not mean new research results but new methods and guidelines that are suitable for practical application. Accordingly, the following developments may be expected in the future:

• Seismic hazard evaluation of dam site: There are four aspects: (i) besides ground shaking the earthquake hazard includes faulting, mass movements and others; (ii) the dam engineer does not need real earthquake records as analysis input but models of the earthquake ground motion, (iii) for the safety check of dams spectrum-matched acceleration time histories of the safety evaluation earthquake are required, and improvements in ground motion prediction models, especially for ground motion parameters with very long return periods.
• Seismic design criteria: Changes are related to (i) the seismic design of hydro-mechanical end electro-mechanical components of spillways and low-level outlets, (ii) the seismic load combinations, and (iii) the design criteria for dam cascades along rivers and very large reservoirs. 
• Seismic performance criteria: The general criteria may remain, but there is a need for low-level outlets.
• Dynamic material properties: New material models are expected for embankment dam materials, estimates of the deformational characteristics of rockfill is required as input for advanced deformation analyses of embankment dams. 
• Methods of seismic analyses of dams: Nonlinear seismic analysis methods need further development. New types of embankment dams need reliable estimates of inelastic seismic deformations, e.g. asphalt core rockfill dams.
• Seismic instrumentation of dams: Seismic instrumentation should be installed in all large dams.

Dr. Martin Wieland, Chairman, ICOLD Committee on Seismic Aspects of Dam Design, Poyry Switzerland Ltd