Machine Learning in the energy sector

from Michaela Tiedemann | 22 February 2018 | Basics

Machine Learning in the energy industry - The power grid is one of the most largest machines ever built. The construction, operation and maintenance of all the individual elements are technological masterpieces. The complexity of the power grid is enormous - the simple fact that electricity always comes out of the socket cannot be taken for granted.

Especially since the increasing conversion to renewable energies and also due to the electrification of more and more areas, the challenges to operate the electricity grid in a stable manner are also increasing. Machine Learning-methods can not only help to overcome these challenges, but they can transform the energy industry in numerous areas.

The power grid from an engineer's perspective: everything you need to know about the biggest machine in the world.

Inhaltsverzeichnis

Predictive maintenance for wind turbines

Predictive maintenance has become one of the new standards in Industry 4.0 in recent years. Many European energy suppliers are also pursuing the goal of preventing the possible failure of wind turbines, for example, through preventive measures. In doing so, the strategies usually pursue a dual purpose. On the one hand, the Reduce repair costs and on the other hand the Security of supply to be able to guarantee this without restriction.

The specific solutions can look very different. For one of our customers from the energy industry, we have therefore developed a Hackathon carried out. This is a very good way of identifying the concrete Use Case and to develop initial approaches to solutions or prototypes.

As a result, the already existing infrastructure and operational data were used to compare these with Machine learning algorithms or to train them to predict transmission failures. This also required training the experts on site in statistical modelling and machine learning.

By training the experts in data analysis methods, they were able to implement faster and improved analyses within the company. The client also received a comprehensive list of technical and organisational recommendations to increase the likelihood of success of the use of Predictive maintenance for wind turbines to improve.

Link Tip: Learn more about Predictive Maintenance in our free whitepaper.

Power plant maintenance

However, individual use cases like this are usually only the beginning of a comprehensive data strategy. Power plants are not only very complex structures, but also involve an extreme amount of investment. That is why they depend on the longest possible service life in order to be profitable. Energy supply companies that want to reduce the maintenance costs of their plants are therefore increasingly relying on data-driven and condition-based Maintenance planning. Machine learning plays an important role in such maintenance plans.

The Maintenance measures are based on the current condition of the facilities and aim to ensure their availability and increase their efficiency. Due to intensive use, signs of wear are the order of the day, so permanent monitoring is necessary.

Decisions on when to maintain, modernise or repair which component must always be made depending on the economic efficiency of the entire plant and the security of supply. The use of machine learning algorithms leads to these Decisions on a reliable, data-based foundation can be made. At the same time, this makes it possible to better understand individual processes and thus identify optimisation potential.

Visualisation of electricity consumption

But not only the generation of electricity, but also the best possible supply to customers is a central concern in the energy market. The better they understand the consumption behaviour of their customers, the better energy service providers can do their job. For a renowned energy service provider, we should therefore bring transparency to the consumption data of its customers in order to achieve a Benchmark for sales and marketing to obtain.

In this context, the electricity consumption of certain types of use, such as hotels, supermarkets or schools, was to be analysed and characterised. Furthermore, he wanted to know which of his customers might be migrating to other providers in order to be able to actively counteract this.

The solution initially involved determining the electricity consumption of the individual customer segments on an annual basis. In the course of this, the development of the electricity consumption of individual customer groups over the years could also be shown in connection with customer churn.

The total electricity consumption was additionally calculated with Building information such as year of construction, floor area, number of storeys and type of use. Machine learning algorithms were also used to analyse the existing usage data in order to identify consumption patterns. At the end of the solution process was the conception and realisation of a Visualisation tools for sales and marketing.

The potential of machine learning in the energy industry

The spectrum of possible applications of machine learning in the energy industry is wide. From industrial contexts to the prediction of future consumption to the design of the customer journey. The use of machine learning is therefore particularly suitable in such a complex and versatile environment as the energy industry. Three core areas can be identified in which machine learning can bring enormous improvements to the energy industry:

Increase the Reliability from Mechanical components
Mastery the rising Complexity
Reduction the Life cycle costs large plants

The challenges in the entire energy sector are increasing in the coming years. In this context, machine learning methods can ensure the permanent use of all components of the power grid. Machine learning in the energy sector thus not only helps to prevent penalties as far as possible, but also to ensure that security of supply can be permanently guaranteed in the future.

Author

Michaela Tiedemann

Michaela Tiedemann has been part of the Alexander Thamm GmbH team since the early start-up days. She has actively shaped the development from a fast-moving, spontaneous start-up to a successful company. With the founding of her own family, a whole new chapter began for Michaela Tiedemann at the same time. Hanging up her job, however, was out of the question for the new mother. Instead, she developed a strategy to reconcile her job as Chief Marketing Officer with her role as a mother.