A Computational Intelligence Approach to Voltage and Power Control in HV-MTDC Grids

An ever increasing interest in renewable energy sources (RES) in order to reduce global carbon emissions and alleviate problems posed by climate changes has lead to dramatic increase particularly in offshore wind energy projects. Several different projects are currently being executed and so many are being planned for the future. The location of wind power plants is increasingly going farther away from shore with more and more challenges and opportunities. With the increased distance to shore, HVAC (High Voltage Alternating Current) transmission is consequently becoming impossible to be used as a result of increase in cable charging currents with distance. As such, HVDC (High Voltage Direct Current) transmission is becoming the only alternative to transmit power from far offshore wind power plants to shore. Two ways of doing this; point-to-point connection and multi-terminal connection. Multi-terminal connection offers dramatic improvement in flexibility, security, reliability of power supply, and with current advancement in power electronic control, offers so much. Thus, a multi-terminal grid connection is the subject of this thesis. The most important issue for multi-terminal grids has been it's controllability. Two most important controllable parameters in multi-terminal grids is voltage and power with voltage being more important as IGBT (insulated gate bipolar transistors) switches are still very sensitive devices. Beside, controlling voltage entails controlling power as they are both dependent on each other. This Thesis proposes a new computational-based control philosophy for the direct voltage and active power control of a VSC (Voltage Source Converter) based multi-terminal offshore DC grid. The limitations of the classical control strategies to HV-MTDC (High Voltage Multi-terminal Direct Current) grids were studied and in particular direct voltage droop control strategy. The main drawbacks of classical voltage droop control is its difficulty to reach power reference set-points and does not ensure a minimum loss profile in the event of contingencies. The proposed strategy is capable of addressing these weaknesses by combining the advantages of the droop controller such as robustness and exceptional ability to compensate for imbalance during contingencies, and the advantages of the constant active power controller which has the ability to reach easily power set points. Thus, the direct voltage droop control strategy and constant active power control strategy were combined, simultaneously solving the drawbacks of each. The advantages of the new Fuzzy controller is the reduced computational effort, the high degree of flexibility, limited influence of topology or the size of grid, and the near zero percentage error. The control strategy is demonstrated by means of time domain simulations for a three terminal VSC-based offshore HVDC grid system used for the grid connection of large offshore wind power plants. Furthermore, a high level controller in the form of an optimal dispatcher was implemented using a genetic algorithm (GA) optimizer to form a complete hierarchical control system - The fuzzy based strategy is used at the local layer and an optimizer/scheduler at the upper layer. The Optimizer optimizes for losses and provides optimal reference set points to the fuzzy-based controllers. The optimizer also checks regularly the available wind power and when it changes, it defines new set points. It uses the new information on wind generated to recalculate set points and return all nodal power fixed terminals to pre-disturbance levels. However results of the GA optimizer in comparison with traditional Newton-Raphson method do not show considerable improvement in reducing losses as expected and this was confirmed by similar works reported in literature. Finally, simulation results are presented in order to demonstrates the capabilities of the proposed control strategy in meeting all the design objectives. No deviation in power or voltage from the references, no influence of topology, configuration, or size. Hence, there will be no need for a secondary corrective action to alleviate the deviations.

Subject

HVDC
VSC
MTDC
knowledge-based control
fuzzy control
genetic algorithm
transmission system
offshore grids
HV-MTDC
hierarchical control
optimization

To reference this document use:

http://resolver.tudelft.nl/uuid:446a43d3-f729-4392-b72d-8f68737e5a64

Embargo date

2020-07-11

Part of collection

Student theses

Document type

master thesis

Rights

Files

PDF

Adedotun_Agbemuko_Master_Thesis.pdf

9.86 MB

Close viewer