On 13th February 2020 had place at University of Southampton a Distinguished Lecturer Program (DLP) seminar called: “DC Transmission Grids: Components, Modelling, Control and Protection Challenges”. This seminar was given by Professor Dragan Jovcic from Aberdeen HVDC Research Centre, University of Aberdeen.
High Voltage DC Transmission has seen rapid technology advances in the last 20 years driven by the implementation of VSC (Voltage Source Converters) at GW powers and in particular introduction of MMC (Modular Multilevel Converters). The development of interconnected DC transmission grids will require significant further advance from the existing point-to-point HVDC links, but it is widely believed that complex DC power grids can be built with comparable performance, reliability, flexibility and losses as traditional AC grids. The primary motivation for DC grid development is the need to interconnect multiple DC lines located in close proximity, and to enable power trading between many DC terminals, as an example in the proposed North Sea DC grid, or EU-wide overlay DC grid. AC transmission is not feasible with long subsea cables, and it is inferior to DC systems in some other conditions. This presentation will address some key technical challenges in DC Grid development and discuss the current technology status.
MMC concept, including half-bridge and full-bridge modules, will underpin building blocks in most DC grid converters and further improvements are expected in terms of efficiency and fault handling. DC/DC converters for transmission grids are not yet commercially available but there is lot of research world-wide and some prototypes have been demonstrated. DC/DC converters may take multiple functions including: DC voltage stepping (transformer role), DC fault interruption (DC CB role) and power flow control. DC hubs can be viewed as electronic DC substations, capable of interconnecting multiple DC lines.
Very fast DC CB circuit breakers (2ms) have become commercially available recently, but the cost is still considerably higher than AC CBs. Slightly slower mechanical DC CBs (8ms) are becoming commercially available at high voltages, while new technical solutions are emerging worldwide for achieving faster operation and lower size/weight/costs.
DC grid modelling will face the challenge of numerous converters dynamically coupled through low-impedance DC cables/lines. A compromise between simulation speed and accuracy is required, leading to average-value modelling, commonly in rotating DQ frame and with variable structure to represent blocked state under fault conditions.
The control of DC grids will require new solutions, since there is no system-wide common frequency to indicate power unbalance. DC voltage responds to both: global power balance and local power flow scenario. DC grid dynamics will be 2 orders of magnitude faster than traditional AC systems and most components will be controllable implying numerous, fast control loop interactions. Because of lack of inertia, and minimal overload capability for semiconductors, DC grid primary and secondary control should be fast, autonomous and robust.
The protection of DC grids is a significant technical challenge, both in terms of components and protection system development. A reasonably accurate and reliable DC grid protection can be developed using local traveling wave measurements. Nevertheless, there is a substantial challenge to achieve security margins, to manage self-protection on various components, back-up grid-wide protection, and in general to achieve reliability levels comparable with AC grids. .
The topology with full bridge MMC converters and slow DC CB avoids any MMC blocking for DC faults and represents a possible low-risk solution.