Blog

Powering up: The future potential of drone inspections for power & utility network management

03 December 2019

Written by National Drones

WA_Pole_Gif

 

Drones or Remotely Piloted Aircraft Systems (RPAS) have been around for quite some time in one form or another, however, recent advances in technology have propelled them into every-day use. With studies into drone technology and research into the economic impact of RPAS being announced regularly you could be led to believe that everyone is using them. Our belief is that it's just the beginning, particularly for the power and utility sector where new opportunities for drone surveys and drone mapping software are being discovered and explored every day.

We’re already seeing cases around the world where this is  happening – where drones are allowing data to be captured more efficiently, saving time and resources on asset inventory calculations, asset management and condition reporting, pre-construction planning, and vegetation management. Yet, there is still so much potential to be uncovered.

It is estimated that power and utilities (P&U) sector losses related to network outages cost $169 billion globally. Drones and the data they collect have the potential to improve the reliability of supply through better condition monitoring and improved fault finding response on areas where supply faults may already be identified.  Leveraging additional technologies such as deep learning and artificial intelligence will enable further savings, cost reductions and efficiencies to be realised, enabling better outcomes for both customers, and P&U companies themselves.

Here are just a few of the opportunities we feel are ripe for the picking.

 

The opportunities for drone Surveys &
drone Mapping Software

Screen Shot 2019-11-22 at 10.38.36 am

Thermographic inspection of HV Tower in SmartData

Using Drone Thermography for network inspection programs

In the past, network mapping and inspection have been completed via expensive helicopter and light aircraft flights to aggregate Light Detection and Ranging (LiDAR) and imagery data for further analysis. As drone and sensor technology improves, there is now an opportunity to use integrated hardware sensors on the drone itself to capture high-resolution data such as RGB imagery, thermal imagery and LiDAR data in the one inspection. Furthermore, data capture from multiple angles of an asset allows for the generation of 3D models that are incredibly useful for maintenance planning, situational awareness, and fault finding.

 
WA_Pole_Gif

High Resolution 3D Model of Power asset in Western Australia

 

Where power quality is suspected to be an issue, an inspection of all relevant assets (poles and towers, cross-arms, hardware and fittings) on a pole can be conducted using the one piece of hardware. When this data is combined with thermal imagery and other data sources, defects can be identified, and the asset condition can be assessed.

Furthermore, thermal imagery will allow companies to determine whether electrical components are operating within effective limits. Thermography surveys of line cables will assist in identifying potential hotspots, as well as potentially show possible earths in SWER lines. Thermal technology can also be used on substations to assist in predictive maintenance.

 

Using deep learning & artificial intelligence (AI) analytics for asset inventories

image008

National Drones Deep Learning Engine can be used to extract features or defects

 

Deep learning is a subset of artificial intelligence. It's one type of technology that can deliver real cost savings and efficiencies in power networks through its capacity by

a) learning how to identify and report on individual elements and

b) recording asset inventory based on large amounts of raw data.

Large numbers of assets aren't easy to review manually as they require significant investments of labour and time. Deep learning can be used to identify initial areas of interest quite quickly, which can then be reviewed by people for accuracy. By coupling the automatic recognition of assets and inventory with the identification of defects on structures and potential vegetation encroachment, many of the challenges associated with manual inspections can be solved.

The overall benefit is that the speed and quality of data capture is enhanced providing more detailed and useful reports and analytics. High-resolution inspection data carried out by a drone would enable planners to make accurate decisions on high priority areas for maintenance, thereby reducing the potential downtime of network assets.

Once the data has been captured, and the AI technology has been trained to learn what is required, condition-based risk monitoring scores and outputs can be created. Potential risks and hazards to the network can be identified allowing maintenance planners to better create maintenance schedules.

 

Enhancing asset monitoring using Drone LiDAR & photogrammetry

LiDAR and photogrammetry both have their strengths and weaknesses - depending on the outcomes to be achieved, and both can be deployed on an RPA.

LiDAR Drone Technology

Pacific Hwy Lidar
                                                             Lidar Survey in NSW of Power Infrastructure for Asset Identification

LiDAR uses light pulses emitted from a rotating sensor and then measures the time taken for a return to reach the sensor. As the speed of light is a known figure,  the distance to an object can be calculated using the time measured. By integrating this data with other sensor information (accurate GPS and IMU), an XYZ co-ordinate can be applied to that point. Some of the LiDAR sensors on a drone are capable of capturing nearly a million points per second. From this, a detailed point cloud can be generated.

LiDAR can assist Power & Utility companies to:

• inspect heavily vegetated areas and identifying vegetation encroachment without human or manned aircraft.

• conduct powerline sag surveys safely, particularly during hot weather and peak loads.

• inspect complex structures such as substations.

• conduct night-time operations as the LiDAR uses a laser scanner as opposed to a camera so point clouds can be generated during night flights.

• conduct operations which require faster processing and turnaround times than standard imagery. (LiDAR data can generally be turned around generally much faster than imagery related products. For vegetation encroachment calculations and measurements, a point cloud can often be generated immediately after the flight has finished).

• reduce flying time and associated costs - LiDAR generally requires less flying time on an RPA than a camera payload.

• identify thin wires and structures, such as powerlines.

• collect vital data for improved network planning and asset identification.

Photogrammetry

Pipeline_Survey-1
                                                            Corridor Survey of Power Assets in South Australia captured by National Drones 

Photogrammetry is progressing at an extremely rapid rate, allowing the creation of complex datasets in the form of 3D Models, point clouds, orthomosaics, DTM’s and DSM’s to be created from standard imagery, provided enough overlap is captured by the drone. Photogrammetry does require large amounts of overlap between subsequent images, and longer processing times.

Photogrammetry offers several advantages for P&U companies:

• Colourised point clouds can be generated with the pixel data captured by the camera. This is an advantage over the height returns or intensity returns shown by a LiDAR alone.

• Raw imagery can also be used for condition reporting.

• Colourised products such as orthomosaics and 3D Meshes can be generated.

• Photogrammetry can also be used for the measurement of volumes and other data either using the point cloud, DTM, or DSM.

• It’s easier to identify objects in the point cloud due to the colourisation

 

Long-range drone inspections

IMG_9625 copy
                                                             National Drones Long Range Aircraft on Survey in South Australia

Another opportunity on the horizon is utilising long-range RPAS to conduct network inspections and monitoring programs over larger distances (say 100km corridors). Currently, this is an extremely time-intensive and expensive exercise for P&U companies.

The technology is advancing rapidly enough for these inspections to become business as usual in the not-too-distant future, and the regulations will start to catch up over the next two to five years. Power companies who anticipate these changes early will be in a good position to adopt the technology faster than those who don’t.

The Civil Aviation Safety Authority (CASA) is adopting a SORA (Specific Operations Risk Assessment) process for reviewing beyond the visual line of sight (BVLOS) applications. This is less cumbersome than how CASA previously reviewed these projects and takes a quantitative approach to risk assessment.

In time, the current network monitoring that takes place using helicopters or light aircraft could be completed using long-range, endurance drones. This will mean less risk for aircrew operating at a low level, often in less than favourable conditions, and the data deliverables are likely to be of a higher quality. There are large potential cost reductions to be realised as well. 

A sensor suite comprised of RGB cameras, LiDAR, thermal imaging cameras, and potentially multi-spectral or hyper-spectral cameras could be carried to capture an array of data for analysis. Capturing all of this data and then merging it would allow P&U companies to generate valuable insights into the condition of their network. 

This will be an area where a large cost benefit saving can be realised. Manned aircraft are expensive to operate from a fuel, insurance, crew and mobilisation perspective. RPAs operate at approximately one-tenth of the fuel cost with reduced maintenance and other operational costs.

This will be a natural progression from helicopters and fixed-wing aircraft to RPAs, especially as the regulations to conduct these types of operations become more flexible. P&U companies that are investigating, and trialling these technologies now will be well placed into the future to realise the benefits of automated inspections from RPAs.

Long-range RPAS operations would be particularly suited to remote area inspections and data collection in locations such as in Western Australia and North West Queensland where the network may not be inspected as regularly as metropolitan areas.

 

The Challenges of using drone inspections &
Drone Mapping Software

"While the opportunities regarding advances in hardware and software technology are promising for the P&U sector, certain challenges need to be overcome before the adoption can be widespread"

Australia's Regulatory Environment

20190516_190240 (1)

Completing an RePL Course in Queensland

 

One of the current challenges preventing drones from being used more regularly in power applications is the requirement in Australia to keep the RPA within line of sight of the person conducting the operations. Typically, this limits the radius of operations to 200-300 metres, depending on the person’s eyesight and size of the RPA itself.

Future iterations of the regulations, namely the Manual of Standards, will allow drones to be operated under what's called extended visual line of sight (EVLOS).

The updated extended visual line of sight regulations will allow for the RPA to be operated up to 1500 metres away from the relevant observer providing certain conditions are met. There will be two classes of EVLOS operations provided for under the Manual of Standards.

Current regulations are particularly limiting for power line corridor surveys where drones could offer huge time and cost savings when it comes to inspections. Traditionally, these are completed by manned fixed-wing aircraft and helicopters, combined with satellite imagery in some cases. These aircraft carry expensive LiDAR sensors and cameras to map and create a point cloud/mesh of the condition of the network, as well as identify vegetation encroachment areas which could cause network outages.

There are industry estimates that flying drones beyond visual line of site (BVLOS) will cost between $200-$300 per mile, compared to helicopter flights which can average $1200-$1600 per mile, offering P&U companies a significant cost saving.

Drones can already capture a higher quality dataset generally with more information per image/scan than traditional methods. Valuable analytics can then be extracted from each asset, whether that be an individual pole, transformer, substation or otherwise.

To this point, technology has far exceeded the regulations in keeping up with the capabilities of the RPAS. New regulatory processes mean that the potential for drone technology for longer projects such as corridor mapping is now a real possibility in the power sector.

The regularity of RPAS deployment

"RPAS technology can help with fault finding and troubleshooting of the current network when outages were notified, but the drone needs to be deployed regularly for sufficient data to be gathered and organised"

If organisations aren't deploying drones on a regular basis, or are using a mix of data collected by drones and traditional means, there is a risk of insufficient and inconsistent data being captured for decision making and analysis.

Drones, when used effectively, not only reduce risk but can save on labour costs. Whilst there is a capex or opex consideration this is often offset by the reduction in labour costs and, depending on the asset, the increased speed of surveys and improved quality of data compared to traditional methods.

For a drone program to be successful in the long terms, companies should ensure that drones are deployed regularly and used consistently for projects.

 

Collecting sufficient data for deep learning

Deep learning has enormous potential to enhance the analysis and use of data captured by drones. But it’s important to understand that for deep learning to be effective, the more data that is captured and built into the models, the more accurate the data becomes over time. Whilst the model is being trained initially, it is important to take a long-term view of data capture so that the model can reach a high confidence interval.

Deep learning models also perform well when the data used to train the model is of a high standard and has been labelled correctly in the first instance. The first step to correctly deploying deep learning and data analytics is to ensure the data captured is of a sufficiently high quality, to allow useful analysis. A poorly trained deep learning model will create additional challenges with false positives and incorrectly identified aspects and will require re-training.

There are multiple elements to the deep learning/AI puzzle, and the technology is still in its infancy. While concepts such as feature extraction and data identification are now easily achievable with well-trained datasets, the next steps is to create accurate, predictive models based on the data which is being captured now.

A Final Word on drone inspections

Implementing a drone survey program within any business, including P&U companies, comes with challenges. Traditionally, drone programs fail due to poor data management and an inability to property utilise valuable data analytics. Companies often purchase a drone or multiple drones without fully considering what they want to achieve for the business. We hear from program managers that they set out to capture video and photos only for them to be stored on an employe's computer never to be shared or analysed. 

A successful drone program is outcome-led and has a whole-of-business focus. We recommend considering the following:

            • Who are the key internal and external stakeholders that can benefit from the data? Often multiple business units can benefit, so think broadly.
            • What data needs to be captured so that useful analytics can be derived?
            • How will the data be analysed and presented? Do you have access to drone mapping software and analytics software? You may need to purchase a subscription
            • How can the data captured improve current business processes?
            • What are the business objectives? For example, is the business looking to save money, improve process, increase quality and efficiency of data capture?
            • Is there capex and opex considerations which would dictate the suitability of an in-house program versus outsourcing the capture and visualisation of drone data?

As we’ve explored above, there are numerous ways that drone technology can save P&U companies time, money and maintenance both now and in the future. We believe that it is a matter of staying abreast of technological changes and planning for them now so that they can be taken advantage of as soon as they are ready. 

Interested in exploring drone technology for your organisation? Please get in touch with us.

Contact Us