Invisible faults in power lines can be revealed with a hyperspectral camera

Senop HSC-2 hyperspectral camera attached to a drone

Can a fault be repaired already before it occurs? That is the aim of a research project initiated by Rejlers, in which the possibilities of enhancing fault location and condition management of electricity distribution networks are examined with the help of hyperspectral cameras. The camera makes the invisible visible.

The image produced by a hyperspectral camera looks like a washed out black and white image, but it reveals more of its object than is visible to the human eye.

Rejlers’ expert Joona Ehrnrooth says that the camera is able to identify the materials and characteristics of various surfaces in the image.

“The camera images wavelength ranges which are invisible to the human eye. This way, specific ‘spectral fingerprints’ can be created of the materials imaged. Any deviations in the materials are also revealed.”

“In this project we are looking for answers, could the images be used for purposes such as analysing the rottenness of pillars, the condition of insulators, the rustiness of transformer machines, and oil leakages. The trees standing along the power line could perhaps also be analysed to determine their species or condition,” Ehrnrooth outlines.

“Hyperspectral technology enables a more preventive mode of operation,” says Tomi Öster, development engineer at Järvi-Suomen Energia, a company involved in the project.

“Our goal is to become able to find faults by identifying them in the materials before something has broken. Ideally this could result in a more reliable electricity supply as well as cost and energy savings.”

Rejlers believes in new technology and in its air force

The condition of electricity distribution networks is currently monitored and faults are located from the ground, on foot, and by helicopter, with drones also now seen as a promising option. Imaging is performed with conventional colour cameras, thermal imaging and infrared cameras and with laser scanners producing three-dimensional images.

Joona Ehrnrooth says that Rejlers firmly believes in the future use of drones. In addition to this, microtechnology has reduced the size of hyperspectral cameras to the extent that they can be attached to drones.

“Modern drones are capable of almost completely autonomous flight along pre-defined routes, and their operating times and performance have increased. We believe that the use hyperspectral cameras in addition to other cameras improves the reliability of analyses and observations,” Ehrnrooth reports.

“The research project expressly aims to determine the system package and the settings needed to carry out the inspection tasks and to find out how the materials collected during inspection can be used for automatically detecting what we are looking for.”

Hyperspectral imaging is a versatile tool

Hyperspectral imaging is not a new innovation. It has already been investigated and applied, for example, in the analysis of the health of vegetation, and on the production lines of factories for quality control and identification. Research Professor Eija Honkavaara from the National Land Survey of Finland, who is responsible for the project’s hyperspectral measurements, says that hyperspectral and drone technology and operation are developing fast.

“We have investigated the use of hyperspectral technology in applications related to the environment, especially in precision farming, forest inventories, measuring forests’ health and nutritional status, and analysing the quality of waterbodies. Hyperspectral analysis can be used on a broad scale for other areas of application as well, such as skin cancer research, examination of food composition and waste sorting, for example.”

At the same time, the legislation governing unmanned aviation is developing. Autonomous drone flights will soon become possible in this regard, too.

“There are many interesting questions related to the utilisation of the technology in the condition survey of electricity networks. Among other things, we are interested in the correlations between the wear of components and the hyperspectral images, in whether the technology can be used for achieving a better level of automation compared to traditional cameras, and in finding the most efficient artificial intelligence-based methods,” Honkavaara explains.

A technology still largely unused

The cameras to be used in the project come from the Finnish security and defence company Senop Oy, which offers night vision equipment, intelligent sensor technology-based solutions and system integration services to the authorities. Small hyperspectral cameras represent the newest technology developments of the company.

“They are ‘made in Lievestuore’,” as sales director Matti Rautiainen proudly says.

“The potential of the cameras is far from exhausted, and the entire industry has yet to be built up. The camera’s price is still comparable to that of a car, but this is already low enough for universities and research institutes to increasingly use the cameras in their research work.”

“The service provider layer for this technology is still completely non-existent, but as awareness of the hyperspectral cameras and their possibilities grows, their applications will also increase.

The ongoing research project is a very good example of the discovery of new possibilities.”

“If the project succeeds, it can create new companies and jobs for this niche industry of experts.”

New understanding for the benefit of the industry

Matti Rautiainen says that the data collected by a hyperspectral camera is not measured in tonnes but in truckloads. Joona Ehrnrooth is not worried about this.

“Through Eija’s (Honkavaara’s) team, we have access to top-level expertise in the processing and analysis of the data produced by the camera.”

“In the final report of the project, we will present specifications for data processing and for the software platform. We will also compare the performance of the hyperspectral camera to colour camera technology and other common methods. In addition to this, we will report our research findings regarding the automatic identification of component condition observations and faulty states,” Ehrnrooth says.

As a result, operators in the energy sector will receive access to the models created in the study on the spectrum of the key components used in electricity networks both in their fault-free condition and in faulty states, and these spectrums will have been tested under field conditions.

Eija Honkavaara states that the strategical goal of the Finnish Geospatial Research Institute at the National Land Survey of Finland is exactly this: to provide research results for the benefit of industry and society.

“Hyperspectral technology allows for a better resolution than conventional cameras. This might be of critical significance in terms of the level of automation.”

“It is wonderful that companies such as Rejlers and the energy providers have the courage to examine the utilisation of this technology in their operations,” she sums up.

Tomi Öster of Järvi-Suomen Energia also sees a lot of potential in the technology.

“Development work can take years, as it is not enough to just get our hands on new information through the technology – we must learn to utilise it as well. This, however, is a truly interesting and significant research project for the future.”


More information:

Rejlers offers comprehensive expertise in the area of electricity distribution networks for the entire lifecycle of the network, from strategic consultancy to design, construction and condition management. The company is actively searching for and testing innovations that can produce added value for its customers. In addition to Rejlers, the research project related to the use of hyperspectral cameras also involves the Finnish Geospatial Research Institute (FGI) at the National Land Survey of Finland, hyperspectral camera manufacturer Senop Oy, hybrid drone manufacturer Avartek Drones Oy Ab, software company Terrasolid Oy and power distribution network company Järvi-Suomen Energia.

The project is financed by the Electricity Research Pool (Sähkötutkimuspooli), Suur-Savo Energy Foundation, Järvi-Suomen Energia Oy, the Finnish Geospatial Research Institute at the National Land Survey of Finland and Rejlers Finland Oy. The project started in September 2020 and it will run until the end of October 2021.

Contact details

Rejlers Finland Oy, Joona Ehrnrooth, +358 50 463 5855, joona.ehrnrooth@rejlers.fi

National Land Survey of Finland, Eija Honkavaara, +358 40 192 0835, eija.honkavaara@nls.fi

Senop Oy, Matti Rautiainen, +358 40 763 6830, matti.rautiainen@senop.fi

Järvi-Suomen Energia, Tomi Öster, +358 50 363 3646, tomi.oster@sssoy.fi