About this Journal Submit a Manuscript Table of Contents
Advances in Mechanical Engineering
Volume 2013 (2013), Article ID 978140, 8 pages
Research Article

Efficiency of a Directly Driven Generator for Hydrokinetic Energy Conversion

Division of Electricity, The Ångström Laboratory, The Swedish Centre for Renewable Electric Energy Conversion, Uppsala University, P.O. Box 534, 751 21 Uppsala, Sweden

Received 23 May 2013; Revised 6 September 2013; Accepted 6 September 2013

Academic Editor: Fabrizio Marignetti

Copyright © 2013 Mårten Grabbe et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.


An experimental setup for hydrokinetic energy conversion comprising a vertical axis turbine, a directly driven permanent magnet generator, and a control system has been designed and constructed for deployment in the river Dalälven in Sweden. This paper is devoted to discussing the mechanical and electrical design of the generator used in the experimental setup. The generator housing is designed to be water tight, and it also acts as a support structure for the turbine shaft. The generator efficiency has been measured in the range of 5–16.7 rpm, showing that operation in the low velocity range up to 1.5 m/s is possible with a directly driven generator.

1. Introduction

Research activities in the area of energy conversion from freely flowing water have increased over the last couple of years, and much has been published regarding tidal currents [14] as a promising renewable resource. Furthermore, a handful of rather large tidal turbines have been going through initial tests recently at test sites such as EMEC (http://www.emec.org.uk/about-us/emec-history/, 2013-01-07) or at individual sites such as the 1.2 MW Seagen in Strangford Lough (http://www.marineturbines.com/, 2013-01-07). Less has been published regarding energy conversion from the freely flowing water in rivers, however, but it should be noted that Khan et al. have reviewed the technology [5, 6] and that Toniolo et al. recently have performed interesting work regarding the river resource in Alaska [711]. Here, the term hydrokinetic energy conversion is used to emphasize that any kind of water current can be of interest, be it tidal currents, rivers, or other ocean currents.

Turbines for tidal currents and turbines for rivers may in many instances have a lot in common. At the same time, however, other technical solutions might be preferable in a river setting, for instance, regarding foundation design and deployment procedure, as well as the fact that smaller turbines might be required due to the limited water depth. Some examples of turbines for hydrokinetic energy conversion in rivers include Verdant Power’s turbines in East River, New York (http://verdantpower.com/, 2013-01-07) and a floating Darrieus turbine by New Energy Corporation (http://newenergycorp.ca/, 2013-01-07) deployed in Winnipeg River at Pointe du Bois, Manitoba [12].

The concept presented in this paper comprises a fixed pitch, vertical axis turbine connected to a directly driven permanent magnet generator. The generator and control system is designed to electrically start, control, and brake the turbine to exclude the use of a blade pitch mechanism and a mechanical brake. An experimental station has recently been finalized, and the concept is set to be tested in a river setting in river Dalälven at Söderfors, Sweden. This paper presents experimental results pertaining to the generator performance in the full range of operation expected at the river site. To the best of the author’s knowledge, this is the first time a direct drive generator will be used in the low velocity range presented by a river. The results are important for evaluating the turbine performance in the prototype energy converter, and the mechanical design and the generator performance may be of interest for other research projects looking to utilize the hydrokinetic resource.

2. The Söderfors Project

A brief overview of the Söderfors project is included to give the context in which the generator has been designed and under what conditions it will be operated. As mentioned in [13], the experimental station at Söderfors is primarily intended for proof of concept, validation of simulation tools, and experience.

The site was chosen as it was deemed suitable from a practicable viewpoint: it is less than an hour’s drive from the university, the unit can be deployed from a bridge, and close cooperation with the owner of the upstream hydropower plant will allow for control of the flow to a certain extent during experiments. The water velocity at the site is mostly in the range of up to 1.5 m/s [14], well suited for studying system performance at the comparatively low velocities presented by rivers.

As with many early prototypes, several design choices have been made with simplicity in mind. For instance, the turbine is a straight bladed Darrieus turbine with non-tapered blades, and a tripod gravity foundation (see Figure 1) is used to allow for deployment with no underwater construction work. A sea cable connects the generator to a measuring station on shore, housing the control system and electrical load as the experimental station is not grid connected initially.

Figure 1: Overview of the whole unit with turbine and generator mounted on a tripod foundation. As a reference of scale, the turbine blades are 3.5 m long and the turbine radius is 3 m. Design by Anders Nilsson.

3. Theory

The amount of kinetic energy extracted by a turbine from a flowing fluid can be expressed in terms of its power coefficient as where is the water velocity and is the turbine cross sectional area. The value of depends on the relative velocity between the turbine blades and the water, usually called tip speed ratio and defined as where is the angular velocity of the turbine and is the turbine radius. For a certain tip speed ratio, , the turbine will give the highest power coefficient, . It is preferable to operate the turbine near its optimal tip speed ratio from the cut in velocity up to the nominal velocity. In order to maintain maximum power capture while the water velocity changes, the generator and turbine rotational speed should be controlled so that

The mechanical power on the generator shaft will thus be proportional to the rotational speed cubed, or , during fixed tip speed operation according to (1). The mechanical system can be described as where is the electrical power output of the generator, is the frictional loss in bearings and sealing, is the total electrical losses in the generator, and is the total angular inertia of turbine, shaft, and rotor. At the same time, the terminal voltage of the generator is roughly proportional to the rotational speed . It is thus of interest to choose the nominal current carefully, in order to allow for electrical control and braking of the turbine at higher water velocities. Knowledge of the velocity distribution is therefore an important piece in the generator design procedure.

The generator has been designed with the aid of an in-house developed design tool where the combined set of field and circuit equations are solved in the finite element environment ACE [15]. The magnetic field inside the core of the generator is assumed to be axisymmetrical and modelled in two dimensions. The displacement field is neglected and the permanent magnets are modelled using the current sheet approach [16]. Furthermore, coil end impedances are introduced in the circuit equations, the laminated stator core is modelled using a single valued magnetization curve, and a correction factor of 1.5 is used for all iron losses.

The complete generator model is described by a combined set of field and circuit equations. The magnetic vector potential inside the generator is described by where is the conductivity, is the permeability, is the axial component of the magnetic vector potential, and is the applied potential.

The circuit equations are described by where , , and are the conductor currents in the three phases , , and , respectively. and are the terminal line voltages, and , , and are the phase voltages obtained from solving the field equation. is the winding resistance, and describes the coil end inductance. Furthermore, it should be noted that frictional losses in the bearings and windage losses are neglected in the simulation model while evaluating the efficiency due to the low rotational speed and high torque.

4. Experimental Setup

The main goal with the experimental setup is to test the system under realistic conditions in a river setting, and therefore, several design choices have been made with simplicity in mind rather than optimizing for energy production. The depth at the site is around 7 m, and in order to leave ample room for the foundation and generator, as well as avoiding floating debris and ice near the surface, it was decided to use turbine blades of 3.5 m in height. The simple NACA0021 chord was used for the blades as there are measurements available, and five blades were used to limit the torque oscillations. Simulations suggest that the 21 m2 turbine should have a of 0.35 at a tip speed ratio of 3.5 (see [14] for more information), resulting in the nominal electrical power of 7.5 kW at a velocity of 1.4 m/s. The system was designed to operate the turbine at a lower tip speed ratio in the range 1.4–1.7 m/s so as to limit the power captured by the turbine. At higher velocities, the generator will be connected to a dump load to brake and stop the turbine. The highest expected velocity at the site is 2 m/s.

4.1. Generator

The electrical design of the generator is based on a laboratory prototype [17], adjusted to the turbine characteristics and the flow velocities at the site. It is essentially a cable wound permanent magnet generator rated at 7.5 kW, 128 V at unity power factor, and a nominal speed of 15 rpm. It has 112 poles and six cables per slot and is wound with a 5/4 fractional wave winding. No slot wedges are used, as the cables are put in axially in each slot rather than radially from the air gap. The nominal design values are given in Table 1, and the FE model of the generator is shown in Figure 2.

Table 1: Generator design data and measured values.
Figure 2: The simulated field distribution in a symmetrical section of the generator under nominal load conditions.

Most of the work concerned integration of the mechanical design with the turbine and the foundation, as well as sealing of the generator housing and inclusion of equipment for control and monitoring. All of the parts were designed at the university, machined at a local work shop, and assembled by hand at the university before being transported to the test site.

The housing supports both the turbine and the generator and is designed to be water tight. The stator frame is basically a thin metal ring, or steel tube, reinforced with beams and flanges (see Figure 3, item 2). Firstly, the laser cut stator sheets (M800-100A) were stacked and secured, and the stator was wound with a standard 16 mm2 cable (RK 16 450/750 V); see Figures 4 and 5.

Figure 3: Overview of stator support structure and the generator housing. 1: the flat bottom flange with holes used for mounting magnets, 2: stator support, 3: rotor and shaft with bearings, and 4: lid and turbine support with sealing and bearing housing.
Figure 4: The stator before winding.
Figure 5: The stator winding with white, grey, and black cables used for the three phases.

After winding, the stator section was fastened to the flat flanged bottom with standard O-rings for sealing. The un-magnetized rotor was then lowered into the stator, resting on a spherical roller bearing (SKF 23940 CC/W33); see Figure 6. The conically shaped lid was mounted onto the upper stator flange, and the rotor was thus secured with the toroidal roller bearing (SKF C 3056 K) in the top of the cone. The forces on the turbine bearing are thus transmitted through the conical lid and stator frame down to the foundation. The conical lid allows for just one bearing and one seal for turbine support and generator housing. The seal (Trelleborg Turcon Roto VL Seal) was, however, not used in the laboratory experiments as it is supposed to operate in water.

Figure 6: The un-magnetized rotor in position resting on the thrust bearing.

In order to facilitate the laboratory experiments, the generator was tilted on its side and fastened in a provisional support structure. The rotor could then be magnetized with three NdFeB magnets per pole inserted through the holes in the flat flanged bottom and into the milled grooves in the rotor. In order to drive the generator in a laboratory setting, it was connected to a 22 kW induction motor with a gearbox (gear ratio 89.89) and a 30 kW frequency inverter, allowing operation up to 16.7 rpm and 12 kNm. After testing was completed, the sealing was mounted and filled with biodegradable grease (SKF LGGB 2) before a short test run. The holes in the flat flanged bottom were then permanently sealed, and the generator was painted before being transported to Söderfors to be mounted on the foundation.

4.2. Measurement System

The control and measurement system of the Söderfors station is based on LabVIEW and a CompactRIO (http://www.ni.com/,  2013-01-07). This interface is used to control the electrical system of the Söderfors station, as described in [18], so as to operate the turbine at a designated tip speed ratio and to handle start and brake procedures. The control system also includes current and voltage measurements which are used during these laboratory experiments with the generator. Aside from the current and voltage measurements that are incorporated in the Söderfors station, the laboratory experiments have allowed the measurement of frictional losses, torque, and B-fields as described below.

Currents are measured for each generator phase and to and from the capacitor bank using HAL 100 and 200 current transducers from LEM (http://www.lem.com/,  2013-01-07). They have an accuracy of 1% and give a voltage signal proportional to the current within a specified range. Voltages from the generator and over the load have been measured after voltage division. Three line-to-line voltages and one phase-to-neutral voltage are measured.

Measurement signals are acquired using a C-series module NI9205 in the CompactRIO, and a sampling rate of 2 kHz was used. The voltage signals were calibrated using an APPA 207 True RMS meter, which gave a scale factor for each voltage division. The offset of the current transducers was set to be within ±0.018 A.

In the laboratory setup, torque on the generator shaft was measured with a DF-30 torque sensor from Lorenz Messtechnik, and the signal was transmitted to the NI 9205 module via a 2.4 GHz SG-link from MicroStrain.

Before magnetizing the rotor, retardation tests were performed. The rotational speed of the rotor was then measured using an IR-sensor (Photologic OPB715) which detected the milled grooves in the rotor. After magnetization, a handheld gaussmeter (Lakeshore Model 410) with an accuracy of ±2% of reading and ±0.1% of full scale (±2 T) was used to measure the B-field in the air gap.

5. Experiments

Frictional losses prior to magnetization were measured by accelerating the rotor manually and recording its deceleration. The retardation from over 20 rpm to standstill was recorded, capturing the expected range of operation.

The B-field in the air gap was measured both during standstill and at nominal speed.

Two sets of experiments were performed using the motor-drive system described in Section 5. In the first, the generator was operated at variable speed (2–16.7 rpm) against a purely resistive, Y-connected balanced three-phase load. Different loads were used to cover the expected range of power for fixed tip speed ratio operation of the vertical axis turbine. In the second set of experiments, the generator was operated against an uncontrolled diode rectifier bridge connected to a resistive load in parallel with a capacitor bank of 26.4 mF to more closely resemble the intended operation in Söderfors. In both cases, the resistive load was a set of 1 Ω resistors (Vishay RPS 500 series) mounted on heat sinks.

6. Results and Discussion

Three retardation tests were analysed, suggesting that the frictional torque is close to 32 Nm in the operating range. However, once in operation, the bearings will operate under much higher radial loads, likely increasing the frictional losses somewhat.

The B-field was measured at the surface of several stator teeth positioned in front of a pole. The field varied slightly, 0.66–0.68 T, measured with an accuracy of ±0.02, perhaps slightly lower than the simulated field; see Table 1. Part of the variation may be explained due to differences between individual magnets, specified to have a typical remanence of 1.22 T and a minimum remanence of 1.17 T.

During magnetization, it was noticed that a part of the rotor was not machined properly, resulting in unexpected leakage flux. At that point in time, it was decided not to delay the project by dismantling the machine.

The machine worked well during operation, although at a lower voltage level due to the mentioned leakage flux, in turn resulting in higher copper losses and lower efficiency at the intended design point (7.5 kW at 15 rpm; see Table 1). Simulations predict the same no-load voltage as the experiments (138 V) with an effective length of 177 mm, 20 mm shorter than the stator stack length. This can to a large extent be explained by the leakage flux at the top of the rotor; see Figure 7 for a graphical illustration of the situation. Furthermore, the simulations do not account for leakage due to each pole being made of three separate magnets or leakage due to equal rotor and stator stack length as discussed in [19].

Figure 7: Artistic impression of the flux components created by the permanent magnets in a radial cross section of the machine. (a) The simulated case with no leakage flux. (b) The actual machine with significant leakage at the top of the rotor. The effective length is thus shorter than the stator stack length.

The measured efficiency in the range 5–16.7 rpm with a resistive load and with a rectifier is shown in Figures 8 and 9, respectively. The generator is expected to operate in the presented range during fixed tip speed ratio operation of the turbine. It will operate at higher efficiency at low speed and low load, while to handle the quickly increasing power from the turbine at higher velocities, the efficiency will drop at higher rotational speed during fixed tip speed operation as discussed in [20]. To illustrate this, the measured voltage drop at constant rotational speed (15 rpm) and increasing power is shown in Figure 10. As expected, the voltage drops quicker as the generator output is rectified.

Figure 8: Measured efficiency during variable speed operation with different resistive loads.
Figure 9: Measured efficiency during variable speed operation with rectifier and different resistive loads.
Figure 10: Comparison of the measured voltage drop and efficiency at nominal speed as a function of power output for the two cases, that is, operation with increasing resistive load and with rectifier and increasing resistive load .

7. Conclusions

A cable wound PM generator for hydrokinetic energy conversion has been finalized. Laboratory experiments show that the generator efficiency is above 80% with a resistive load in the expected range of water velocities up to the rated velocity of 1.4 m/s. In operation with a diode rectifier and a resistive load, the efficiency remains above 75% in the same range of operation. The efficiency is lower than expected, mainly due to axial leakage flux. In conclusion, the results show that the generator is suitable for operation with a vertical axis turbine in the range of velocities expected at the site at hand.

Conflict of Interests

The authors have no direct relation, financial or otherwise, with the commercial identities mentioned in the paper that might lead to a conflict of interests for any of the authors.


The work reported was financially supported by the Swedish Centre for Renewable Electric Energy Conversion, STandUP for Energy, Ångpanneföreningen’s Foundation for Research and Development (ÅForsk), the J. Gust. Richert Foundation, Vattenfall, AB, and the Swedish Research Council (Grant no. 621-2009-4946). The mechanical design of the generator performed by Anders Nilsson is deeply appreciated. Anders Karlsson, Johanna Lundqvist, Daniel Käller, and Erika Schweitz are acknowledged for assembly work.


  1. L. S. Blunden and A. S. Bahaj, “Tidal energy resource assessment for tidal stream generators,” Journal of Power and Energy, vol. 221, no. 2, pp. 137–146, 2007. View at Publisher · View at Google Scholar · View at Scopus
  2. M. Grabbe, E. Lalander, S. Lundin, and M. Leijon, “A review of the tidal current energy resource in Norway,” Renewable and Sustainable Energy Reviews, vol. 13, no. 8, pp. 1898–1909, 2009. View at Publisher · View at Google Scholar · View at Scopus
  3. A. S. Bahaj, “Generating electricity from the oceans,” Renewable and Sustainable Energy Reviews, vol. 15, no. 7, pp. 3399–3416, 2011. View at Publisher · View at Google Scholar · View at Scopus
  4. F. O'Rourke, F. Boyle, and A. Reynolds, “Tidal current energy resource assessment in Ireland: current status and future update,” Renewable and Sustainable Energy Reviews, vol. 14, no. 9, pp. 3206–3212, 2010. View at Publisher · View at Google Scholar · View at Scopus
  5. M. J. Khan, M. T. Iqbal, and J. E. Quaicoe, “River current energy conversion systems: progress, prospects and challenges,” Renewable and Sustainable Energy Reviews, vol. 12, no. 8, pp. 2177–2193, 2008. View at Publisher · View at Google Scholar · View at Scopus
  6. M. J. Khan, G. Bhuyan, M. T. Iqbal, and J. E. Quaicoe, “Hydrokinetic energy conversion systems and assessment of horizontal and vertical axis turbines for river and tidal applications: a technology status review,” Applied Energy, vol. 86, no. 10, pp. 1823–1835, 2009. View at Publisher · View at Google Scholar · View at Scopus
  7. H. Toniolo, P. Duvoy, S. V. Anlesberg, and J. Johnson, “Modelling and field measurements in support of the hydrokinetic resource assessment for the Tanana river at Nenana, Alaska,” Journal of Power and Energy, vol. 224, no. 8, pp. 1127–1139, 2010. View at Publisher · View at Google Scholar · View at Scopus
  8. C. Walsh, J. Fochesatto, and H. Toniolo, “The importance of flow and turbulence characteristics for hydrokinetic energy development on the Tanana River at Nenana, Alaska,” Journal of Power and Energy, vol. 226, no. 2, pp. 283–299, 2012. View at Publisher · View at Google Scholar · View at Scopus
  9. H. Toniolo, “Hydrokinetic assessment of the Kvichak River near Igiugig, Alaska, using a two-dimensional hydrodynamic model,” Energy and Power Engineering, vol. 4, no. 6, pp. 422–431, 2012.
  10. P. Duvoy and H. Toniolo, “HYDROKAL: a module for in-stream hydrokinetic resource assessment,” Computers and Geosciences, vol. 39, pp. 171–181, 2012. View at Publisher · View at Google Scholar · View at Scopus
  11. H. Toniolo, “Bed forms and sediment characteristics along the thalweg on the Tanana River near Nenana, Alaska, USA,” Natural Resources, vol. 4, no. 1, pp. 20–30, 2013.
  12. E. Bibeau, S. Kassam, J. Woods, T. Molinski, and C. Bear, “Operating a 5-kW grid-connected hydrokinetic turbine in a river in cold climates,” Journal of Ocean Technology, vol. 4, no. 4, pp. 71–82, 2009.
  13. K. Yuen, S. Lundin, M. Grabbe, E. Lalander, A. Goude, and M. Leijon, “The Söderfors project: construction of an experimental hydrokinetic power station,” in Proceedings of the 9th European Wave and Tidal Energy Conference (EWTEC '11), pp. 1–5, Southampton, UK, September 2011.
  14. M. Grabbe, K. Yuen, A. Goude, E. Lalander, and M. Leijon, “Design of an experimental setup for hydro-kinetic energy conversion,” International Journal on Hydropower and Dams, vol. 16, no. 5, pp. 112–116, 2009. View at Scopus
  15. ACE User Manual, Modified Version 3.1, ABB Common Platform for Field Analysis and Simulations, ABB Corporate Research Centre, Västerås, Sweden, 2001.
  16. R. Gupta, T. Yoshino, and Y. Saito, “Finite element solution of permanent magnetic field,” IEEE Transactions on Magnetics, vol. 26, no. 2, pp. 383–386, 1990. View at Publisher · View at Google Scholar · View at Scopus
  17. K. Thomas, M. Grabbe, K. Yuen, and M. Leijon, “A low-speed generator for energy conversion from marine currents—experimental validation of simulations,” Journal of Power and Energy, vol. 222, no. 4, pp. 381–388, 2008. View at Publisher · View at Google Scholar · View at Scopus
  18. K. Yuen, System perspectives on hydro-kinetic energy conversion [Ph.D. thesis], Uppsala University, Uppsala, Sweden, 2012.
  19. J. Pyrhönen, V. Ruuskanen, J. Nerg, J. Puranen, and H. Jussila, “Permanent-magnet length effects in AC machines,” IEEE Transactions on Magnetics, vol. 46, no. 10, pp. 3783–3789, 2010. View at Publisher · View at Google Scholar · View at Scopus
  20. K. Yuen, K. Thomas, M. Grabbe et al., “Matching a permanent magnet synchronous generator to a fixed pitch vertical axis turbine for marine current energy conversion,” IEEE Journal of Oceanic Engineering, vol. 34, no. 1, pp. 24–31, 2009. View at Publisher · View at Google Scholar · View at Scopus