The lever and the wheel were the technological revolutions of their time. Now, in the 2020s, robotics and artificial intelligence will revolutionize the way we work. Software robotics can be used, for example, in various dimensioning and routing tasks in electricity network planning. The resulting network designs will be comparable and consistent in quality and take into account both technical and economic considerations. These designs provide a solid basis for prioritizing investment projects and thus support the utility's decision-making.
Requirements on the carbon neutrality of production and security of supply are tightening. More distributed and weather-dependent generation will be connected to the network, and electric cars are rapidly gaining ground. Utilities are responding to these challenges by investing heavily: improving the weatherproofing of distribution networks and the ability to respond to the energy transition, and – as urbanization intensifies – placing overhead lines underground.
How can utilities respond to all the challenges in an increasingly complex environment? Taking into account different variables, assessing multiple scenarios, and finding solutions that are both cost-effective and take into account electrotechnical safety is a time-consuming and demanding task for the planner. Old and new cross-sections of conductors, distribution cabinets and fuse sizes; in an electricity network, all these factors are intertwined and can easily lead to a very laborious process of trial and error. What if there was a robot that would take care of these routines as a tireless assistant to the planner?
To meet this need, Trimble developed Network Optimizer in collaboration with Caruna and Elenia, the two largest Investor Owned Utilities in Finland.
The robot's power lies in its ability to perform a given task over and over again, quickly and tirelessly. Network Optimizer finds the best possible network structure at the planner's command and ensures that the network is correctly dimensioned. If the input data changes, the calculation can be rerun at the touch of a button. The design of the network itself is based on minimizing the life cycle costs over a given period – within the technical constraints. By taking over certain laborious and time-consuming routines, Optimizer frees up time for experts to focus on more relevant issues.
– The job of a network analyst involves a lot of understanding of the operating environments of different network areas and analyzing future trends. Thanks to Network Optimizer, more time is available for analyzing such large-scale issues, as time-consuming manual steps can be done faster than before," says Janne Sorsanen, Network Analyst at Caruna.
Also Vesa Hälvä, Development Manager at Elenia, says that the main benefit and goal with Network Optimizer is to improve and speed up the planning process. The robot virtually eliminates one manual step.
– Network Optimizer enables a smoother design process involving less repetitive work. In addition to the actual technical-economic optimization of the network, we have also worked with Trimble to develop post-processing of the design so that routine tasks that were previously manual are now automated," says Hälvä.
Detailed information earlier in the project
At Caruna, Network Optimizer has transformed the network development process that covers the steps from long-term planning all the way to network construction.
– Thanks to Optimizer, we are able to do detailed plans early in the process, while we are still in the preliminary planning phase, so we have better visibility of the end result also in financial terms," says Elina Lehtomäki, Senior Vice President of Electricity Network Management and Operation at Caruna.
– We get a good idea of the likely cost level and cost-effectiveness of the renovation. And on the other hand, we can avoid cost surprises later on. Optimizer plays a big part in the management of the investment portfolio in particular," Lehtomäki continues.
Also in Vesa Hälvä’s view, the improvement in forecasting accuracy is an essential benefit.
– We get a more accurate cost and volume estimate of the project earlier in the process, he says.
Consistent, comparable, strategy-based network designs
According to Hälvä, Network Optimizer is also a useful tool for putting design principles into practice.
– Above all, Optimizer produces network designs of consistent quality, enabling Elenia to model the way we construct networks. Similarly, any strategic changes to construction principles can be fed directly into Optimizer, allowing the change to be implemented quickly and efficiently across the entire project portfolio," says Hälvä.
– At this stage of using robotically-assisted planning, the biggest benefits can be seen in the efficiency of the planning process and the consistent quality of the network designs, which strengthens network asset management, he continues.
According to Lehtomäki the planning robot can also be used to ensure supply chain transparency, i.e. that the network complies with Caruna's planning principles.
– In case of any deviations, there is no need to figure out the reasons later, as Optimizer helps you go back to the previously built network and see the principles by which the network was designed and built.
The division of labour between the expert and the planning robot is clear: the expert provides input data, which Network Optimizer then uses to perform hundreds to thousands of calculation rounds. Within minutes, Optimizer returns the best network design in terms of minimized life cycle costs. This network is also correctly dimensioned from an electro-technical point of view and follows utility-specific planning principles from network structures to component types and distribution reliability requirements.
So utilizing robotics speeds up the planning and cost estimation processes, but what other benefits does it bring to the user?
– It significantly speeds up the user's thinking process because, in my experience, it is easier to evaluate an existing network solution than to create a completely new one. I also see a significant learning aspect in using the solution in this way to allow the user to broaden their view on network development," says Janne Sorsanen.
As a potential challenge Sorsanen identifies the quality of network modeling.
– The actual quality of solutions is largely dependent on the quality of the input data. This underlines the importance of a good documentation quality," he says.
Hälvä highlights the division of labour between the planning robot and its user.
– It is important to know the principles according to which the tool works in order to anticipate its suitability for different situations. It is not even desirable to try to build a tool that can handle all special cases, but it makes sense to leave these to an expert.
Therefore, humans will still be needed in network planning.
Network Optimizer has undoubtedly already revolutionized network planning, but at the same time it is easy to see many possible ways in which the robot could further expand process support. What are the utilities' thoughts on future developments?
Elenia’s Hälvä says that it would be great if in the future, on the way to more holistic optimization, Network Optimizer could learn to interact more with its user and use network asset and condition data to enrich decision-making.
At Caruna, the thoughts are moving forward in the same direction. Lehtomäki plays with the idea of Network Optimizer taking into account the overall condition of the network and choosing the most suitable solution for each purpose from the toolbox. Today, the choice is still made by an expert. But what if network modelling evolved from using history, energy and power data to predicting what is most likely to happen in each part of the network in the future?
– If we would be able to at least model where there is growth and where there is migration loss, says Lehtomäki.
– What makes the operation of utilities challenging is the long-term perspective, as investments are made for a very long time. As dependence on electricity grows and phenomena change at the same time, it is precisely the predictability of all this that must be brought into system management, she continues.
Perhaps future machine intelligence will make it possible to detect and understand subtle, underlying signals, so that when it comes to the electricity network in a particular region, we would be able to meet the requirements that were still looming around the corner. Whatever the future role of the planning robot, one thing is certain: the professional skills of an expert will become increasingly important.
