Smart grid monitoring: new opportunities and challenges with digitalization and 5G integration

Published at #SINTEFblogg

Advanced communication technologies in smart grid monitoring. Studying smart grids is turning into dealing with a cross-disciplinary subject. It encloses power system and ICT system, which are becoming closely interdependent. This interdependence has a huge impact on performance and reliability of the smart grid. Therefore new methodologies for studying the future distribution grid are necessary.

In this blog post, and the article it is based on, a novel methodology for analyzing the performance and dependability of advanced communication technologies in smart grid monitoring applications is proposed. In particular, this methodology (and the correspondent software developed) has been employed to analyze the impact of communication failures on distribution grid state estimation.

The exploitation of ICT in power system operation promises to achieve a significant enhancement in the management of the power network, in terms of better performances, reliability and quality of service.

Along with these advantages, ICT inclusion in power system management introduces new challenges, due to the multiple factors that may affect the correct exchange of information within the Smart Grid Wide Area Monitoring, Control and Protection system (SG-WAMPAC). These factors can be classified in two main categories:

In the Centre for Intelligent Electricity Distribution (CINELDI) we develop new risk assessment methodologies in smart grids that take into account, among other things, the growing  uncertainty in future distribution systems.

The main issue in studying the future distribution grid is to allow the two main components of the smart grid (power system and ICT) to be analyzed in a comprehensive way, despite being two systems with their own particular characteristics. Today’s methodologies do not allow this.

Our attention should be focused on the design of new simulation tools. They must be able to merge these two systems and describe the interdependencies among the elements that compose the complex system. Different demands coexist: on one hand, the necessity of reproducing the operation of the system in an accurate way, on the other hand the need of simplifying the system according to the detail of interest.

This scenario opens some challenges in the research area. First of all the cross disciplinary nature of the system, where different experts have to merge knowledge, but also their background, reference models, research approaches, in order to target a clearly defined and shared goal. Moreover, another big challenge is to create a balanced model, where the level of detail in modeling both ICT and power systems has to be balanced, according to the target of the research.

The methodology that we developed is based on the Stochastic Activity Network (SAN) formalism, a stochastic generalization of Petri Nets purposely designed for dealing with complex system modeling. Specific C++ libraries have also been developed and deeply integrated within the SAN framework, that allow performing power flow calculations and state estimation calculations within the simulation studies.

The methodology developed has been adopted for quantifying the impact of ICT failures on the performances of a traditional state estimation algorithm. State estimation consists in the calculation of the electrical quantities that define the state of the network from data collected by measurement devices distributed on the electrical network. Typical measured quantities are power flows on the branches, and voltages and power injections on the buses.

A standard IEEE radial distribution network is used in our simulations, where voltage and power sensors are distributed along the network and interconnected to a central Supervisory Control And Data Acquisition system (SCADA), that gathers the measurement data from the sensors and performs the state estimation calculation. When failures occur in the monitoring system (failures on sensors, communication link, communication core network, etc.) the observability of the power network is reduced, as well as the accuracy of the state estimation of the grid. Quantifying the sensitivity of the state estimation accuracy upon the failure rate of a communication technology allows achieving a deeper insight on the most proper communication technology for this specific smart grid application.

In our analysis, the communication technology considered is the 5^th generation cellular communication (5G). 5G technologies represent a promising candidate for implementing the communication infrastructure that allows data traffic to be transmitted from measurement devices to control centers in Wide Area Monitoring, Control and Protection (WAMPAC) systems. In fact, 5G is expected to meet the requirements for a smart grid implementation, with highly reliable communication, low latencies, high safety against malicious intruders and high scalability.

Two different versions of 5G Radio Access Technologies (RAT), namely 5G with LTE and 5G with URLLC (Ultra Reliable and Low Latency Communication) are compared with an ideal failure-free communication technology, in order to highlight the impact of communication failures on the state estimation accuracy. The study showed significant improvements in the performances of the state estimation with a monitoring system supported by URLLC-based 5G technologies compared with LTE-based communication infrastructures. For example, in terms of voltage mean estimation error, the URLLC-based 5G technology presented a reduction of 21%, when compared with LTE-based communication technologies. A URLLC-based 5G solution also demonstrated a stronger resilience towards simultaneous failures on wide areas of the power network.