Проф. др Јовица Милановић

The international power industry, electrical power utilities in particular, are required to improve efficiency of planning and operation of integrated electrical system (IES), i.e., distribution and transmission networks (D&TN), to achieve environmental targets at affordable cost. This is coupled with significant workforce shortages and delays in procuring relevant equipment, hardware and software tools to achieve required levels of observability and controllability of both individual and integrated networks and help with improved evidence for their regulatory submissions, strengthened whole-system capability and increased innovation capacity. The current approach to interconnected D&TN simulations is based on assumption that DN is passive (does not generate or export power to transmission network) and that the power follow is from TN to DN. Therefore, a DN is typically modelled using equivalent static exponential or polynomial load model even for dynamic studies. The assumption of exclusively unidirectional power flow already does not hold true with modern and will not hold true with future Active DN (ADN) with embedded generation and storage, as already seen in many cases of operation of existing power systems. In addition to bi-directional power flow in future interconnected power system there will be dynamic interactions between D&TN. Even though there have been various equivalent models of ADN proposed, from conventional equivalent models, to grey and black box models based on data analytics and machine learning approaches, they have not been used much, if at all, in integrated system studies. At present both networks are modelled, studied and operated separately using different software, different data and the information between the two only recently started to be exchanged at greater scale, or rather there is a strong drive by electricity network regulators worldwide to standardize data formats and enable data exchange between D&TN. Consequently, many working groups involving experts from D&TN have been formed to develop guidelines for data sharing and standardization. The recently promoted and promising approach to deal with cost effective, efficient, secure and stable operation of such complex integrated system is development of its digital twin. The advantages of the approach include improved data integration and interoperability, enhanced real-time monitoring and simulation, integration of machine learning, co-simulation frameworks, integration of distributed energy resources (DER), cyber-physical security and resilience, scalable cloud-based architectures and potential use in grid modernization and planning. Considering existing practice in system planning and operation, however, there are still significant challenges associated with development of, and transition to, digital twin to deliver stated potential benefits. Some of those include high initial and ongoing costs (sensors, ICT, modelling tools, skilled workforce) and complexity; data integration (issues between software platforms and data formats), quality and availability; difficulty to synchronize models of transmission and distribution systems, which are currently managed by different system D&TN operators; data privacy and cybersecurity risks; model accuracy, validation, scalability and performance as integrated T&D systems require significant computing power and may suffer from latency or scalability issues; lack of standardization and proprietary issues; and equally importantly, workforce and skill gaps, e.g., lack of personnel with expertise in data science, artificial intelligence (AI), and complex systems modelling, which are essential to develop and maintain digital twins. The approach discussed in this presentation, once fully developed, will potentially have profound impact on planning and operation of integrated systems (not only electrical) and facilitate faster, reliable and secure transition to efficient, resilient and affordable net-zero power networks.It will ensure secure, stable and economic operation of the IES while avoiding unnecessary data exchange (hindered by data volumes, latency, security and standardisation) and excessive computational, modelling, infrastructure, human resource and security costs associated with purely data driven approaches, such as digital twin of an IES.

Кључне речи: uncertainties, low carbon technologies, planning, operation, integrated electrical systems