Hans Kr. Hoidalen

An All-DC Offshore Wind Farm With Series-Connected Turbines: An Alternative to the Classical Parallel AC Model?

In this paper, the concept of an all-dc wind park with series-connected turbines is investigated as an alternative to the classical ac parallel or radial wind park. This paper presents a literature overview of all-dc wind park concepts with series connection. A three-phase conversion system with permanent magnet machine, ac-ac converter, high-frequency transformer, and diode bridge rectifier is suggested in this paper for the series connection of dc turbines. The dc series park with the suggested conversion system is compared in terms of losses, cost, and reliability to the state-of-the-art park configuration which is the ac radial park with HVDC transmission. It is found that the dc series park becomes comparable with the ac radial design for high ratings of the dc turbines. Furthermore, the comparison shows that emphasis must be put on reducing the losses in the conversion system of the dc turbine and, particularly, the ac-ac converter. Therefore, the efficiency of the ac-ac converter is compared for three different topologies: the direct matrix converter, the indirect matrix converter, and the conventional back-to-back converter. The direct matrix converter is found to be the most efficient, suitable for the suggested conversion system.

Ground effects on induced voltages from nearby lightning

The lightning induced voltage on overhead lines from return strokes and its dependency of a lossy ground is analyzed using a new, analytical vector potential formulation. Nortons (1937) approximation and the surface impedance approach are used to take loss effects into account. The surface impedance method predicts in general induced voltages in good agreement with Nortons approximation, but the accuracy of the method is dependent on the variation of the current along the lightning channel. Nortons method is compared with the exact Sommerfeld solution, showing a deviation <10% even for low conducting grounds and distances from 100-1000 m. The effect of stroke location and line termination is also analyzed, showing that a line terminated by its characteristic impedance and excited by a return stroke at the prolongation of the line is especially sensitive to lossy ground effects. Strokes near the mid-point of an overhead line gives less loss effect than strokes at the end of the line. The surface impedance approximation is derived from Nortons method and the necessary assumptions are outlined.

Comparison of wind farm topologies for offshore applications

Increasing energy demand and environmental factors are driving the need for the green energy sources. The trend, in general, with respect to wind farms is to increase the number and the size of wind farms. The wind farms are also being located offshore with the prospect of more consistent and higher energy capture. The offshore wind farms are likely to move farther off from the shores to reduce visual impacts and increase the size. But, this has implications in terms of design of the collection grid and grid interconnection. Farms of 1 GW size and at distances of about 100 km are envisaged [1]. The design of collection system and turbine interconnection will become very important as the farms move farther offshore. Proper choice of collection system topology is important from the point of view of maximum energy capture while ensuring a high reliability of the design. Different collection system topologies have been proposed by researchers before, with the radial system being most popular. One of the key factors in selection would be the losses in the farm. In order to select the most suitable topology a comparison of different topologies with respect to losses, reliability and costs has to be done. Comparisons of calculations indicate that the DC series and series-parallel wind farm design may be options for future wind farm designs. The DC series and series-parallel design have lower reliability, but can be improved by providing redundancies. The designs have equipment costs almost equal to the AC wind farm costs. The losses in DC series-parallel wind farm are higher by about 12 % when compared to AC wind farms. The DC series design is also very attractive design, but has restrictions with respect to insulation. Also, the required turbine ratings may be significantly and unrealistically high when it comes to designing large wind farms. It can also be concluded that the novel designs require significant amount of work before these can be used in real wind farms.

Topology-Correct Reversible Transformer Model

A topology-correct transformer model, which covers core operation under heavy saturation conditions, is presented. A method of accounting for magnetic fluxes outside the core and windings is proposed. Representation of the magnetization curve at high flux densities is considered. The equivalent air gap in the core is taken into account. The model is capable of reproducing inrush currents accurately regardless of which transformer winding (LV or HV) is energized. The model is illustrated by calculating inrush currents produced by subsequent energizations of a single-phase transformer.

Implementation of Inverse Hysteresis Model Into EMTP—Part II: Dynamic Model

The main features of a dynamic hysteresis model (DHM) and its implementation into the electromagnetic transient program (EMTP-ATP) are described. A method of fitting the DHM to catalog data is proposed. The implementation of the DHM within a transformer model is illustrated by transient calculations. Possible methods of fitting the DHM-based transformer model to a no-load test are outlined.The main features of the history-dependent inverse model of magnetic hysteresis are outlined. Its implementation into the Electromagnetic Transient Program-Alternative Transients Program (EMTP-ATP) is described, and the fitting of the model to catalog data is demonstrated. The abilities of the model are illustrated by its use in a single-phase transformer model.

Investigation of the Overvoltage and Fast Transient Phenomena on Transformer Terminals by Taking Into Account the Grounding Effects

A large number of electromagnetic transient studies have so far reported findings related to the overvoltage behavior of systems within a broad frequency range. However, in most cases, actual grounding effects have been either not taken into account or have been just partially considered. Although an accurate methodology to study grounding effects exists, a detailed analysis of the grounding effects has not yet been fully performed. Moreover, many test applications are performed in laboratories where the grounding is close to ideal. In this paper, the grounding system modeling methodology is described, and its capability is demonstrated for an existing complex grounding system. First, the complex grounding system is modeled in the frequency domain by the TRAGSYS program, which represents the grounding system as an equivalent multiport network. Next, the equivalent network parameters are represented in a broad frequency range enabling time domain computations in ATP-EMTP. Finally, the protection level of the chosen lightning arresters is discussed under most severe conditions.

Reliability assessment of DC wind farms

New offshore wind farms are being proposed to be located at distances of 50-100 km from the shore. These large distances make HVDC connection to grid a feasible option. DC collection grids have been proposed to eliminate offshore platforms in wind farms. Two main topologies previously proposed for DC offshore wind farms with series connection of wind turbines to eliminate offshore platforms have been studied. The aspects of reliability of wind farm collection system are important in order to attract investors and make the designs physically realizable. Wind farms are large economic investments and in order to make good economic decisions a very important factor is the amount of additional income that can be generated over the life time of the wind farm. Thus the reliability calculations have to be included as a part of design process. The collection system reliability for DC Series and Series-Parallel topologies is mainly addressed in this paper. Various reliability indices like the interruption frequency, interruption duration and expected energy not supplied are calculated using an analytical method. The largest contributors to unreliability are identified. Alternative designs with added redundancy are proposed and evaluated. The important factor of possible additional income that can be generated is calculated to select and improve designs to be implemented. The evaluation results indicate that the DC Series-Parallel wind farms offer a great promise in terms of feasibility and reliability. It is observed that the DC Series design can be used as base system for calculations and understanding of the important factors in reliability.

Characterization of Magnetic Losses in the Transformer Tank Steel

The transformer tank is normally built of construction steel and has relatively large thickness of 1.2 mm for distribution up to 12 mm for large power transformers. In such thick materials, skin effect becomes much more significant than thin materials due to both frequency and non-linearity. The flux density in the tank can vary in a wide range of some mT to 1800 mT (in the case of having high zero sequence flux) leading to a large variation in the permeability and, consequently, in skin depth. Since the losses are high, the temperature rise can also be a main problem influencing the conductivity that must be considered in the loss calculations. The main contribution of this paper is to study the magnetic losses in the transformer tank steel through the measurements and finite difference-based numerical calculations. Using the loss separation principle, different loss mechanisms are investigated, and the behavior of each at the different level of inductions and frequencies is studied.

Calculation of off-core inductance in dual-circuit model of transformer

Zero sequence magnetic flux, generated in transformer core in different operation cases, is forced out of the core including the oil gap and the tank. These off-core flux paths can be represented by inductances in a duality transformation based electrical transformer model. These off-core inductances mainly determine the zero sequence impedance that is vital in analyzing a transformer subjected to the GIC event or unbalanced voltages. Estimation of these inductances is one of the challenges in identification of the electric equivalent circuit. The main contribution of this paper is to present an approach for calculation of the mentioned off-core inductances based on 2D-FEM. Since the transformer structure is not symmetric for off-core flux path, 3D-FE analysis is also used to evaluate and improve the presented method. The results calculated for a 3-leg 3-phase transformer have a good agreement with the values obtained from empirical equations typically adopted by manufacturer.

Analysis of Pipe-Type Cable Impedance Formulations at Low Frequencies

This paper analyzes series impedances of screenless cables enclosed by a common conducting pipe. Established analytical formulas for pipe-type cables parameters are analyzed by taking finite pipe thickness and proximity effect into account. Analytical low-frequency approximations are developed and analyzed. The analysis shows that a classical formulation by mixing infinite and finite pipe thickness can result in incorrect inductance and even negative self-resistance at low frequency. This paper analyzes corrections to take a finite pipe thickness into account. Finite-element simulations reveal that the core-to-core proximity effect is substantial for the analyzed cable design with much lower high-frequency inductance than that obtained from analytical formulas. This paper also analyzes given proximity-effect formulations and proposes corrections based on analytical low-frequency approximations. The proposed proximity effect corrections improve resistance and inductance in common and differential mode.