NVE is currently testing a new method for searching for quick clay, using of electromagnetic waves from helicopters.
When the waves hit the ground, an electric current is induced in conductive materials located in the subsurface. The currents move downwards and create a secondary magnet field. This is then registered by the equipment hanging below the helicopter.
The force of the secondary magnetic field is dependent on what the subsurface consists of – the resistivity (the specific electrical resistance) varies, which is what will be key when interpreting the data.
EMerald Geomodelling, which emerged from the Norwegian Geotechnical Institute (NGI), is now processing extensive amounts of data from a 35 square kilometer area in Nannestad through its computers. They do this in order to provide a better picture of where quick clay is located and what should be done in response. The helicopter has flown back and forth and collected data from more than 400 line kilometers.
“We want to develop better models for mapping quick clay areas, and this has potential,” says section chief Lars Harald Blikra in the Norwegian Water Resources and Energy Directorate (NVE).
The surveys are part of a larger mapping project NVE is currently carrying out in the municipalities Gjerdrum, Nannestad, and Ullensaker. Bilkra says the reason why EMerald is only being used in Nannestad is because the Norwegian Geological Survey (NGU) has already mapped large areas.
“And this measuring method cannot be used too close to buildings, creating some limitations,” says Blikra.
The measuring method is well-known and widely used internationally by the mining industry. More than ten years ago, NGI started researching whether this technique could be utilized for different purposes. It worked excellently, and after a while EMerald was founded as a result of the research. In the spring of 2019, they ran the first mapping for Nye Veier at a stretch in Trøndelag. The objective at the time, was to map loose material and quick clay occurrences, and weak zones in bedrock.
Using four different algorithms
The algorithms are now being developed further from StartupLab’s offices in Forskningsparken, Oslo, with quick clay as the key focus. The measuring equipment is rented from the Danish SkyTEM, while EMerald’s core capabilities lie in the processing of the data collected, self-developed algorithms and machine learning.
CEO Andreas Pfaffhuber says the process occurs in several steps. First, loose materials are separated from bedrock. Subsequently, the loose materials must be separated from each other.
EMerald has developed four different algorithms that, together with machine learning, are continuously making the analyses more accurate.
The main benefit with this type of mapping is not to spot quick clay so accurately that traditional methods are made redundant. It is rather about reducing insecurities and identifying three levels: there is, with certainty, a presence of quick clay – there is, with certainty, not a presence of quick clay – and lastly, there is uncertainty, and more thorough surveys are necessary to find the answer.
Pfaffhuber explains that the data gives them an indirect indication of salt contents, which is an indicator of potential quick clay.
“NGI first used this technique in 2013, for the State Highway Authority. We found some connections that indicated the presence of quick clay, but they were difficult to interpret. This was repeated for Bane Nor (the Norwegian railway) in 2016. This time the findings were better, but still difficult.”
Gathering data from national database
Since then, the algorithms have been refined and improved. The results are best if Emerald’s models also use supporting data from other surveys. Solutions have been developed to gather this data directly from the national database Nadag, operated by the NGU.
Many companies are reluctant to deliver their survey results to Nadag, which is a problem. Several companies justify this by saying their drillings are done in specific situations and thus cannot be used for different purposes.
Pfaffhuber rejects this, saying any skilled geo-technician knows how this kind of data can and should be used.
“Additionally, our systems are capable of handling this uncertainty. That is one of the advantages of machine learning,” says Pfaffhuber.
Also Blikra in NVE confirms that lacking submissions to Nasdag is a problem. He refers to the fact that NVE is now preparing a proposition for mandatory submissions, which will be assessed by the government before it is put in place.
“But this will only apply for new data,” says Blikra.
Preparing to use drones and smaller systems
EMerald Geomodelling has received international attention for its results, with an assignment completed in India, and will soon be embarking on another in Brazil.
EMerald collaborates with Danish SkyTEM, who delivers the technical measuring system. The goal is reducing the size of the equipment used. Today, the hexagon-shaped structure hanging beneath the helicopter covers more than 300 square meters and weighs almost 700 kilos.
“The aim is to make it so small that we can use drones – below 100 kilos, possibly all the way down towards 30,” says project lead Guro Skurdal.
Even though there are several companies out there that deliver similar services, Skurdal believes EMerald is unique.
“We combine geophysics and geotechnics with data skills. We have these algorithms that allow us to combine different data in a way no one else can,” she says.