written by Mileta Žarković
The power sector is experiencing rapid digitalization, and the amount of data actively monitored and stored within electric power systems is growing exponentially each day. With data influenced by meteorological, thermal, mechanical, and electromagnetic parameters, the need for innovative tools to manage and interpret this wealth of information is becoming increasingly critical. As we look to the future, power system experts and engineers will require advanced tools powered by artificial intelligence (AI) to make more informed decisions and gain clearer insights into complex datasets. This is where the Sunrise Project comes into play, pushing the boundaries of digital transformation in the power sector, particularly through the application of AI to optimize power systems.
The newly established Smart Grid Laboratory within the Sunrise Project enables the real-time simulation of decentralized power systems with various assets. Data obtained from these simulations is used to develop Artificial Neural Networks (ANNs) as an alternative to standard power flow models. This AI tool can significantly assist in identifying technical losses in transmission networks and non-technical losses in the distribution grid. Through various datasets, power quality can be analyzed, along with the impact of energy storage systems and distributed renewable generation.
Developed Machine Learning (ML) algorithms can be tested in real-time simulations within the Smart Grid Laboratory to evaluate their effectiveness across Energy Hubs of different sizes. These algorithms can achieve multiple objectives, including: Enhancing energy efficiency in energy hubs, Improving sustainability, Reducing greenhouse gas (GHG) emissions, and Maximizing profits in the electricity market.
Through real-time simulation, various types of faults in power plants and high-voltage substations can be replicated to generate valuable datasets. These datasets are otherwise difficult to obtain, as replicating diverse failures in a laboratory setting is nearly impossible. However, simulations enable a comprehensive analysis of such failures. Using these datasets, machine learning algorithms can be trained to develop auxiliary tools for early failure detection and predictive maintenance, estimating the probability of power equipment failure. At the core of both existing and future research is the application of supervised and unsupervised ML algorithms, which are used to: Cluster monitoring data, Detect anomalies, and Predict failures or accelerated aging of power equipment.
This approach supports the development of optimal maintenance strategies that minimize downtime and reduce operational costs. Additionally, various anomalies and cyberattacks can be simulated, allowing AI models—such as autoencoders and other advanced techniques—to be trained for enhanced system protection. This is crucial for: early fault detection, improved power system diagnostics, and overall power system resilience.
What sets this research apart is AI’s ability to learn and understand the dependencies between various parameters within an energy system. The Sunrise Project has demonstrated that existing power system databases are not static; rather, they are dynamic, living knowledge bases from which timely conclusions and decisions can be drawn. By utilizing these datasets with AI-driven tools, we can not only predict but also proactively address potential issues in the power system.
The Sunrise Project brings cutting-edge research and technology to real-time and simulated data monitoring and decision-making in the power sector. Among its groundbreaking initiatives, the project showcases how data from the Smart Grid Laboratory can be used by AI-based systems to enhance energy efficiency and reliability in decentralized power systems. The key takeaway from this research is the transformative impact of AI in enhancing the reliability, safety, and efficiency of energy supply systems. The use of AI in power systems extends beyond automation; it enables smarter, more proactive decision-making and improves the overall stability of power grids.
As the energy landscape continues to evolve, the Sunrise Project is shaping the future of power systems by providing innovative AI-driven solutions that: enhance power system performance, optimize maintenance schedules, and improve overall energy efficiency.