Harmonics in Offshore Wind Power Plants Employing Power Electronic Devices in the Transmission System

xii it will also enable the OWPP developer, such as DONG Energy Wind Power (DEWP), to provide relevant input for the grid code compliance of an OWPP as well as define technical requirements to potential suppliers. Purpose of the Industrial PhD Project The purpose of this Industrial PhD project has been to investigate and address the interaction (from a harmonic perspective) between the OWPP, and the associated control systems in the WTGs and other PEDs in the transmission system with focus on the VSC-HVDC and the STATCOM. The main scope has been on the assessment of the frequency domain evaluation approach, commonly applied in the OWPP design phase. This has been accomplished by comparing results obtained from conventional and linearized frequency domain analysis methods such as the Nyquist stability criterion against the detailed electromagnetic transient (EMT) based model realised in PSCAD/EMTDC. Measurement Campaigns Detailed, yet generic models of the involved PEDs have been developed in order to meet the project requirements of analysing the harmonic stability phenomena. The generic EMT models of the PEDs have been validated based on comparison of measurement data, in order to ensure their trustworthiness. Test and field measurements have therefore been conducted on commercial PEDs such as ±50 MVar Siemens SVC Plus, the 7 MW ABB SVC Light (modified back-to-back configuration) and a commercial multi-megawatt sized type 4 WTG. Measurements constitute a core part in industry-oriented research. Due to this fact, the research project owes its uniqueness and contributes new insight to the academia. A long term harmonic measurement campaign has therefore been prepared and conducted within the PhD project at Clevehill substation, UK, where four STATCOMs are installed. The Cleveill substation serves as grid connection for the 600 MW sized London Array OWPP. The STATCOMs employ the modular multilevels cascaded converter (MMCC) technology, which is considered state-of-the-art within the industry. The measurement data obtained on the STATCOM has allowed the author to acquire detailed information preferred in order to develop a generic, yet detailed, EMT based model of the MMCC STATCOM, including the switching devices and distributed sub-module (SM) dynamics. Additionally, the author has, together with colleagues at DEWP, established a Cooperative Research and Development Agreement (CRADA) between DEWP and the National Renewable Energy Laboratory (NREL) in Boulder, Colorado, USA. NREL has recently commissioned an advanced multimegawatt sized power electronic grid simulator test system. Test results obtained during the author’s stay at NREL are used to develop and evaluate generic, yet detailed EMT models of the grid simulator (referred to as controllable grid interface, CGI) and a commercial multi-megawatt sized type 4 WTG. PED Modelling and Validation The relatively high number of non-linear semi-conductors in the MMCCs possesses some challenges in the EMT programs, such as PSCAD/EMTDC, as the semi-conductors are triggered by relatively highfrequency signals and as the electrical system’s admittance matrix is altered at each switching instant. A significant computational effort is required for re-triangularisation of the electrical network subsystem’s admittance matrix. The computational burden is considerably increased for the large number of semi-conductors, making the simulation of the MMCC in some occasions impractical when using a conventional modelling approach. Previously, a detailed equivalent model of the MMCC VSCHVDC has been devised based on the “Nested Fast and Simultaneous Solution” procedure. However, there are some limitations to the previous model as it is specifically intended for the MMCC employing half-bridge converters in the SMs. The somewhat more complex structure of the full-bridge converter employed in the MMCC STATCOM makes it complex to apply the existing modelling approach for the STATCOM. A universal modelling technique is proposed, which is able to represent both the halfand full-bridge based SMs. The modelling technique is simple as the derivation of the SM’s Norton equivalent is merely based on circuit inspection. The measurement data from the London Array OWPP measurement campaign has successfully been applied in the evaluation of the generic model of the