Hydropower is not a new industry. We’ve been building hydroelectric dams for over 100 years. So what have we learnt? Are we getting any better at the design and construction of dams, and is dam safety improving?

To answer these questions, let’s consider some statistics from the International Commission on Large Dams (ICOLD). ICOLD reports that 2.2 per cent of dams built before 1950 have failed so far, while the figure for dams built after 1950 is less than 0.5 per cent – a reduction by a factor of four or more (ICOLD Bulletin 99).

So there has been a significant improvement. But it’s important to examine why there has been an improvement, whether age alone is a factor, and how we can make sure that this positive trend continues.

Despite the industry’s best efforts, dams and hydropower plants still fail today. On average there are around 50 dam failures each year, with an estimated annual cost impact exceeding USD 1 billion.

Most of these incidents occur at dams less than 75ft (22.86m) tall. Over 70 per cent are during the first ten years. Failure also appears to become more likely after around 50 years of service.

Although dam failures are relatively rare, they can have very severe consequences, including considerable environmental damage and substantial loss of life.

Given that it’s not possible to remove risk entirely, how do we determine an appropriate safety benchmark? One answer is to look at societal risk tolerance. In general, society is willing to tolerate some risk in return for certain benefits.

Think of air travel. We all know that there are accidents occasionally, but as a population we are generally willing to accept this relatively low level of risk when we fly, in exchange for the convenience.

In the same way, the public needs to have confidence in our hydropower projects. If there’s no social acceptance of the projects that we’re developing then we will not, ultimately, be able to build them.

This is an issue which affects the entire industry. Think of the Fukushima power station in Japan. The failure of that plant had a very negative impact on the public perception of nuclear power, not only in Japan, but across the entire world.

Managing the risks is certainly a big challenge. We need to make sure that the people who are most capable of dealing with particular risks are the ones to whom that responsibility is assigned. This has become more difficult and blurred with newer forms of contracting in project development.

While project owners are often looking at an asset life of 50-100 years, contracting mechanisms don’t usually align with these long timescales. So, it will be increasingly important for construction contractors, consultants and operators to clearly communicate risks and distribute responsibility in an appropriate and effective way.

Building information modelling (BIM) systems provide one solution. We use this software at the planning and design stage to create very detailed 3D models which can help detect certain safety issues in advance. We then provide these models to the construction contractor. Eventually, they can be used by the owner to improve operations and maintenance.

Although this reduces the potential for miscommunication between these various groups, suitable caution must be used. We can’t assume that our digital models are infallible just because they derived from computer analysis. We have to take the possibility of human error into account.

We need to share the lessons we learn with the wider industry."

We also need to improve communication in the industry more generally. ICOLD ascribes the declining rate of dam failures at least in part to the wider dissemination of information and increased knowledge of the risks.

This is an area where the hydropower and dams sector can definitely continue to improve. So how do we learn from each other? How do we make sure that our past failures are not repeated?

One solution is the use of event reporting systems. For example, Stanford University runs the National Performance of Dams Program (NPDP).

Formally launched in 1994, the NPDP allows the dam engineering and safety community to access current and historical data on the actual performance of dams, including incidents of failure. The NPDP hasn’t been very widely adopted yet, but it does give a good sense of what we could achieve by using event reporting systems more widely.

Earlier this year, the United States’ National Hydropower Association (NHA) launched a new programme in operational excellence (OpEx). The programme includes a voluntary event reporting system which receives, distributes, and catalogues operational excellence information (OEI). We can learn from this data to constantly review what we mean by best practice.

The programme also has an educational slant. Considering that both the assets and workforce are ageing in the US context, the OpEx programme will serve as a training resource for future staff by providing a trusted database of historical and current operational issues.

To ensure consistency of safety procedures across the design, construction and operation phases, we need to start thinking more carefully about our contracting mechanisms. We need to make sure that all stakeholders – contractors, consultants, owners and operators – communicate safety information effectively at all stages of a project.

But we also need to improve our external communications. We need to share the lessons we learn with the wider industry. Event reporting systems provide one effective solution, which can have a high impact even in a short period of time.

We should now take full advantage of the increased possibilities for sharing information digitally. In this way, we can make sure that everyone can learn from each other to avoid repeating the mistakes of the past.

You can find out more about BIM software here.

You can find out more about the National Performance of Dams Programme at Stanford University here.

You can find out more about the National Hydropower Association’s OpEx programme here.