The safety of Aotearoa New Zealand’s hydro scheme dams in the face of a large Alpine Fault earthquake is a recurring concern raised at many AF8 Roadshow events. We asked Brent Wilson*, chief engineer at Meridian Energy, to have a talk with us about dam safety.
Dams are an important part of how we produce electricity. The energy generated by water moving quickly downhill can be harnessed and converted into electricity. Electricity is hard to store, but by storing water behind dams in large lakes or reservoirs above a power station, we are effectively storing electricity until we need it.
All dam owners need to abide by specific legislation for dam safety under various acts like the Building Act (2004) and the Resource Management Act (1991). Requirements are based on a classification system that assesses potential damage and loss of life if the dam were to fail.
Many of Meridian’s structures have a high potential impact classification because of the volume of water they retain and the height of the dam. Therefore, there are rigorous safety measures to be met. Two key requirements are:
An “extreme load event” is the probable maximum flood or the maximum credible earthquake for the location. For many of Meridian’s dam sites, the maximum credible earthquake is a 1 in 10,000-year event. An “unusual load” is equivalent to a 1 in 150-year earthquake. After this kind of earthquake, dams should be able to continue operating without any damage.
As a business, we’re looking to increase that unusual load case to a 1 in 500-year earthquake which would include a magnitude 8 on the Alpine Fault. That’s aspirational but it’s nice to put a peg in the sand as to where we want to get to. Anything new that we do, we’ll be looking to that higher criterion.
Certainly, our business is making renewable power to de-carbonise the economy. We make returns for our shareholders by selling it. But under a seismic event or a flood event, our primary purpose for the business is to keep populations downstream safe.
Maximum earthquakes are location-specific so let’s talk about the Waitaki Valley where we own and operate six hydro power stations. Seismic assessment work began decades ago to determine the maximum credible earthquakes that hydro structures would have to withstand.
Active faults were identified and quantified to determine the likely peak ground accelerations (PGAs) their earthquakes would produce at each site. Because of the early dam building activity, the Waitaki Valley became one of the most seismically understood regions in the country.
What is clear, is that local faults present more of a hazard to our structures than the Alpine Fault. Although earthquakes on local faults occur much less often than Alpine Fault earthquakes, they produce greater ground shaking at our sites.
For example, for the dams and canals in the upper Waitaki (the Pukaki-Ōhau chain), an earthquake on the Ostler Fault could produce PGAs of 1.4g. For the Benmore dam, the Fern Gully Fault is estimated to produce a PGA of 0.9g. For the Waitaki and Aviemore dams, release of the Waitangi-Awahokomo Fault could produce PGAs of 1.4g. By comparison, the Alpine Fault is estimated to produce PGAs of 0.22g at our closest sites.
Because of its distance, a magnitude 8 earthquake on the Alpine Fault is going to result in an order of magnitude less ground shaking than what we’ve been planning for from the local faults.
We do detailed characterisation of the dams. We look at construction records, the compatibility of different construction materials, what standards they were designed and built to. Often we find dams have been built to higher standards than their original design. Then we go through a complex modelling exercise to look at the behaviour of the dam under different load cases.
We really get to know the behaviour of our dams. The Waitaki Dam, for example, is a concrete arch dam. Under normal conditions the weight of the dam holds the water back. When an earthquake happens, the dam behaves as an arch with blocks tending to be pushed downstream and load transferring into the side abutments.
I’ll make the point, considering the case of an extreme load event, that we’re aiming to avoid catastrophic dam failure. There would be damage, we would have to remediate it, but there wouldn’t be an uncontrolled release of water.
Overall, we’re subjecting our structures to much greater load cases than what they would see in an Alpine Fault earthquake. That’s why we have confidence that we won’t be getting a wall of water coming down the valley.
At Meridian we’re really fortunate that most of our assets were built at a time when understanding was really good, and they were built to a high standard. However, we’re always learning more and figuring out ways of reducing pressure on the structures themselves.
The Waitaki dam is now 90 years old, and we’ve discovered that the ground beneath the dam is a lot more fractured than initially thought. After an earthquake there’s the possibility of increased ground water pressure under the dam so we’ve drilled hundreds of additional drain holes to enable water to be pumped out and reduce pressure on the dam.
Aviemore is part earth dam and part concrete. The Waitangi-Awahokomo Fault goes right through the earth part of the dam so if that fault released, the malleable earth material would deform into a new position and retain the integrity of the dam. But we were concerned about the possibility of seiche waves on the lake overtopping and eroding the earth dam, so we’ve put up a wave wall to protect it.
There are many measures like these that can be taken to support the structures. We can also do an additional release of water to take pressure off dams during extreme events – that’s what the spillway capacity is for, to help us manage large floods and earthquakes. We work collaboratively with Genesis who operate schemes upstream of ours to ensure that water flows remain within the prescribed rates and levels.
The simple rules are that we will respond in the same way for any large seismic event – firstly, for our people to “drop, cover, and hold” and then to follow the advice, “if it’s long or strong, get gone”. We’re talking with our staff about the importance of knowing they’re in a comfortable environment before helping with recovery – they need to know where their kids are, where their partner is, that the rest of their family is safe before they come back to work.
Then we’ll set up a response team with a CIMS (coordinated incident management system) structure. There’s a whole chain of events that would transpire post-event depending on what has happened, and we’ve practiced many scenarios.
We have generation controllers sitting in Wellington and Christchurch who can operate from there if communications are intact. We have a lot of our primary protection and control sitting right at the power stations which can be controlled remotely via computer or operated manually by people onsite – which includes operating flood gates and releasing water to take pressure off structures if necessary.
We also have a fully automated backup control system that can make decisions if no one can access the power stations. A general rule of thumb is that if the dam could be overtopped within 24 hours, then that system will kick in to activate spillways and release water in a controlled manner.
Ideally, we would get inspections underway as soon as possible. We have a comprehensive dam safety and surveillance team who regularly walk the canals and dams and outflows, so they know the infrastructure thoroughly. These “route marchers” can pick up minor leaks by looking at piezometer measurements of groundwater and vegetation growth on dam surfaces. Such information will be crucial for getting back to the intelligence team to determine how to respond to any major event.
For an AF8 scenario earthquake there will undoubtedly be damage to infrastructure and extensive business interruption. We expect large civil infrastructure such as penstocks and canals to be impacted. As far as the mechanical and electrical plant is concerned, the actual generators themselves, they’ll probably be ok. The key problem to solve will be how to get water via canals to the power stations.
Given the extensive nature of the Waitaki Valley hydro system, we have a variety of options to divert water and provide some contingency. Structures further downstream are more likely to be resilient because they’re further from the Alpine Fault. So, if power stations and canals upstream are out of action, we can use alternative flow paths like the old Pukaki and Tekapo riverbeds, to get water into Lake Benmore. Then, all of our power generation post-event could be out of the mid-Waitaki Valley. Benmore is where the high voltage DC link goes to the North Island so, if we’re up and running at Benmore, we can get power in from the North Island and still put power out to the North Island.
Meridian have been participating fully in the AF8-focussed working groups and working closely with district councils to think about how we would coordinate an emergency response and post-event recovery.
One thing is clear: recovery will be from an NZ Inc perspective – we will not be fighting with Genesis over access to heavy machinery or worrying about each other’s balance sheets at that point. There will be a strategic and collaborative approach to do what’s best for New Zealand.
We’ve been thinking about how to recover quickly but there’s no doubt there will be huge earthworks needed. Already, we’re going out and identifying the original borrow pits where clay lining material came from and looking at sourcing more material if necessary so we can be ready to make repairs.
Every time there’s a large earthquake we review the implications for our sites. So, after the Darfield and Christchurch earthquakes, after Tōhoku in 2011, and Kaikōura in 2016, we reassessed and updated our load cases.
We’ve learnt that we need to keep doing fault studies. New information has shown that the Ōhau A power station could end up higher than its current position after an earthquake on the Ostler Fault zone. And that’s got big implications for getting water to the power station. So, we’re continuing to undertake fault trenching work to better characterise the Ostler Fault.
We’ve also learnt that, for example after the Christchurch earthquake, there can be a significant drop in demand for electricity. So, if we experience a South Island-wide event, we could expect that a drop in demand might match the drop in our capacity to generate electricity. This means there could be time to repair damaged infrastructure.
All large hydroelectric dam owners are really prudent operators of their assets. We all have an obligation to ensure that populations downstream are safe. Certainly, for Meridian structures, we do not expect any uncontrolled release of water. That’s what we hold ourselves to. If we found a deficiency that meant the structure wouldn’t meet safety criteria, then we would put plans in place for remediation. But in thirty years of investigations, there’s been no significant deficiencies identified.
* Brent Wilson is an electrical engineer by training. He has worked in hydro schemes for 20 years and with Meridian Energy for 16 years. In his current role as Engineering Authority for Meridian he acts as the “voice of the assets” providing impartial advice to the Generation Leadership Team about how best to look after the infrastructure that generates electricity. He also facilitates obtaining insurance for these assets. Previously, he was the Engineering Strategy Manager with accountability for asset management strategies and risk mitigation covering dam safety, electrical, mechanical, civil, and automation. Brent is passionate about getting people and property prepared for an AF8 scenario earthquake because, as he says, “it’s likely to occur in your lifetime”.
Thanks to Brent Wilson and Meridian Energy for supporting this article with their expertise, and Resilience to Nature's Challenges for funding the development of this online article.