Portal de Periódicos da CAPES

Self‐powered autonomous underwater vehicles: results from a gyroscopic energy scavenging prototype

2016; Institution of Engineering and Technology; Volume: 10; Issue: 8 Linguagem: Inglês

10.1049/iet-rpg.2015.0210

1752-1424

Nicholas Townsend,

Fluid Dynamics Simulations and Interactions

IET Renewable Power GenerationVolume 10, Issue 8 p. 1078-1086 Research ArticleFree Access Self-powered autonomous underwater vehicles: results from a gyroscopic energy scavenging prototype Nicholas C Townsend, Corresponding Author Nicholas C Townsend nick@soton.ac.uk Faculty of Engineering and the Environment, University of Southampton, Southampton, SO16 7QF UKSearch for more papers by this author Nicholas C Townsend, Corresponding Author Nicholas C Townsend nick@soton.ac.uk Faculty of Engineering and the Environment, University of Southampton, Southampton, SO16 7QF UKSearch for more papers by this author First published: 08 June 2016 https://doi.org/10.1049/iet-rpg.2015.0210Citations: 20AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinkedInRedditWechat Abstract This study describes and presents preliminary experimental results from a novel prototype energy scavenging system installed in a model 2 m cylindrical autonomous underwater vehicle (AUV). The system, which is based on control moment gyroscope principles, utilises the gyroscopic response of a gimballed flywheel mounted within an AUV body to generate energy from the wave induced rotational motions of the vehicle. This method, using the reaction of a spinning wheel under an input torque to provide an output torque of greater magnitude, orthogonal to the input torque axis and the spin axis provides a means to harvest energy in-situ, without external appendages and additional hydrodynamic drag. The system promises to extend AUV mission durations indefinitely and reduce support vessel time currently required for periodical recharging and redeployment. A description of the system operation, design and experimental results from a series of regular wave tests conducted at zero speed in a towing tank are presented in this study. The results show that the system can harvest energy, with the greatest power generation around resonance, tailing off as the frequency increases and typically non-linear in nature. The system could potentially be applied to any rotationally excited platform e.g. autonomous surface vessels, buoys or boats 1 Introduction Autonomous underwater vehicles (AUVs) have limited endurance capabilities [1]. Nearly all AUVs depend on stored energy for their operation [2]. Currently, the majority of AUVs use batteries as an energy supply for their operation [1, 3]. However, batteries have a finite life (stored energy), require periodical recharging (and redeployment) from a dedicated host platform or support vessel and represent a significant proportion of the total vehicle mass, typically around 20% [2]. A summary of various AUV sensors and AUV power requirements is given in Tables 1 and 2. With current AUV endurance, measured in hours or days [30], alternative power systems or in-situ charging strategies are required to extend missions. Table 1. Example AUV sensor power requirements Sensor Power, W pressure sensor 0.1 (typ) [4] digital compass 0.132 (typ), 0.014 (sleep) [5] sound velocity sensor 0.25 (typ) [6] echo sounder 0.25 (max) [7] fluorometer 0.3 (typ) [8] precision timing reference 0.3 (max) [9] hydrophone 0.12–0.3 (typ) [10] MEMS AHRS and GPS/INS 0.675–0.95 (typ) [11] turbulence sensor 1 (typ) [12] 2D imaging sonar 3 (typ) [13] CTD sensor 3.42 (incl. pump) [14] digital camera 5 (typ) [15] sidescan sonar 5 (typ exclude CPU) [16] LBL acoustic positioning system 2.5–5.5 (transmit), 1.3 (max receive), 0.005–0.285 (listen mode), 0.0025 (standby) [17] nitrate sensor 7.5 (max) [18] Doppler velocity log 12 (max transmit), 2 (average transmit), 1.1 (typ) [19] 3D imaging sonar 15 (typ) [20] underwater RF 16 (transmit), 5 (receive), 0.005 (sleep) [21] current profiler 20–0.3 (transmit), 0.2–1.4 (typ) [22] side scan sonar and sub bottom profiler 30 (max) [23] navigation and control system 50 (max), 2 (active listening), 0.7 (sleep) [24] multibeam swath bathymetry and sidescan 50 (max), 20 (standby) [25] transponder 50 (max), 2 (active), 0.7 (sleep) [26] underwater laser scanner 144 (typ) [27] acoustic communications 300 (transmit), 1.8 (receive/standby), 0.08 (standby) [28] Table 2. Example power consumption of various torpedo style battery powered AUVs [29] Vehicle Mass, kg Length, m Diameter, cm Stored energy, kWh Endurance, h Power, W Iver2-580-S 20.5 1.52 14.7 0.6 14 43 Folaga AUV 31.0 2.04 16.0 0.5 6 90 REMUS 100-S 38.5 1.60 19.0 1.0 10 100 REMUS 600 240.0 3.25 32.4 5.2 70 74 Bluefin 12D 260.0 4.32 32.0 7.5 30 250 Dorado class AUV 680.0 5.30 53.0 11.0 17.5 629 Bluefin 21 750.0 4.93 53.0 13.5 25 540 Abyss AUV 880.0 4.00 66.0 11.2 24 467 REMUS 6000 884.0 3.84 71.0 11.8 22 536 Eagle Ray AUV 940.0 4.60 69.0 29.4 30 979 Explorer-class MARUM SEAL AUV 1300.0 5.50 74.0 14.0 19 737 Echo Surveyor IV 1450.0 5.40 100.0 45.0 50 900 With most of the oceans containing an average wind power density greater than for more than 80% of the year (with a maximum of ) [31], an average solar energy power density at the surface of ∼168 W/m2 (∼1270 W/m2 at the limits of the atmosphere) [32] and the global oceans containing a wave energy density greater than 2 kW/m for 90% of the year (and wave energy densities greater than 20 kW/m in mid-latitudes) [33], the available ambient power is comparable to that required by an AUV. That is, energy scavenging systems, in particularly wave energy scavengers, promise to; enable AUV systems to be remotely and renewably recharged at sea, extend mission durations and capabilities, negate the necessity to carry sufficient energy reserves (size and weight) for entire mission(s), reduce costs by freeing support vessel time (a major cost component in AUV deployment) and provide flexibility in system deployment and recovery (time and/or location). Solar powered AUVs and several solar assisted commercial ASVs systems, e.g. SAUV-II [34], Autonaut [35], WaveGlider [36] and C-Enduro [37], have been developed. While solar potentially offers unlimited mission durations, as found by the SAUV II it is limited to night-time missions and daylight recharging strategies and is susceptible to bio fouling [34]. A prototype thermal energy harvesting underwater vehicle, the SOLO-TREC, has also been developed, which uses a phase-change material (a waxy fluid) that melts and expands in warm water (at the surface) and solidifies in cooler (deeper) water to drive a hydraulic generator and provide power [38]. The SOLO-TREC is reported to provide over 7000 J per dive, sufficient to power GPS, iridium and conductivity–temperature–depth (CTD) sensors [38]. In the past internal combustion engines have been used to power AUVs. However, these systems were found to be limited as additional power is needed to expel the exhaust gases at depths greater than 200 m [1]. Fuel cells have also been trialled, e.g. on the AUV URASHIMA [39] and IDEF Ifremer [40], however, these systems are expensive and complex [40]. The Royal Swedish Navy has used Stirling engines [1] and (Slocum) gliders have been developed using ocean temperature gradients and battery power to provide propulsion [41]. Various wind based concepts, e.g. C-enduro and the Submaran [42], are being developed. The C-enduro uses a deck mounted wind turbine to generate power and the Submaran uses a fixed wing (sail) for propulsion. Various sail based systems have also been developed for the Microtransat Challenge [43] – a fully autonomous sailing boat transatlantic race. Interestingly, no successful transit of the Atlantic within this competition has been made to date. Ideas of recharging AUV power supplies using wave-energy absorbers and sea current generators have also been proposed [44, 45], however, no practical demonstrations have been made. While wave propelled devices, e.g. the Autonaut and Wave Glider have been developed, these systems do not currently generate power. To date, no wave energy scavenging systems have been demonstrated. This paper reports the results from the first demonstration of a gyroscopic wave-based energy scavenging system for an AUV. Currently gyroscopic wave energy converters (WECs) are gaining interest with several systems being developed in Japan [46-48], Italy [49-51] and Spain [52]. Unlike the systems being developed in Italy and Spain, with large, slow flywheels with limited precession and horizontal precession axis and the large (kW scale) WEC systems being developed in Japan, this study demonstrates: A small-scale system specifically for energy scavenging applications. A system with a vertical precession axis (whereby the rolling and pitching motions contribute to the energy harvested). A system with high flywheel rpm (thousands not hundreds). A system with unrestricted precession motion (where the gyroscopic precession can take any angle and the gyroscopic equations cannot necessarily be regarded as linear). A detailed description of the principal of operation, system advantages and the prototype design is given in Section 2. The experimental setup and investigations are described in Section 3 and the results and discussion of results are presented in Section 4. 2 System description 2.1 Principal of operation A schematic of the gyroscopic system is illustrated in Fig. 1. The system, which is based on control moment gyroscope (CMG) principles, utilises the gyroscopic response of a gimballed flywheel mounted within an AUV body to generate energy from the wave induced rotational motions of the vehicle. Fig. 1Open in figure viewerPowerPoint System schematic and coordinate frame definitions [(Xe, Ye, Ze) represents an Earth fixed inertial axis system, (Xh, Yh, Zh) represents the hydrodynamic or equilibrium axis system that moves with the average motion of the AUV, but is not fixed to the AUV, (Xb, Yb, Zb) represents the body (AUV)-fixed axis system, and (Xf, Yf, Zf) represents the flywheel axis system. This axis precesses, but does not spin with the flywheel] The system equations of motion can be described by the Newton–Euler equations. Defining and as the flywheel angular velocities and mass moment of inertias of the flywheel, with respect to (Xf, Yf, Zf), the angular momentum of the flywheel can be expressed as (1) In the (AUV) body-fixed axis (Xb, Yb, Zb) this can be expressed as (2) where represents the rotation matrix describing the transformation of the momentum component from (Xf, Yf, Zf) to (Xb, Yb, Zb). Including the AUV motion, the moments acting around each axis in the AUV body-fixed coordinate frame (Xb, Yb, Zb) can then be expressed as (3) where represents the moments acting around each axis in the (AUV) body-fixed coordinate frame, represents the skew-symmetric form (equivalent to the cross product operation) of the (AUV) body motions experienced by the flywheel. represents angular momentum of the flywheel in the AUV body-fixed axis system (Xb, Yb, Zb). Assuming, as illustrated in Fig. 1, the flywheel is restricted about the x-axis (Xf), precesses about the z-axis (Zf) and has an angular velocity, , about the y-axis (Yf) and the flywheel and AUV centres of mass lie at the origin of the body-frames of reference and the body-frames of reference coincide with the principal axes of inertia of the bodies. As shown in [53], (3) can be expanded yielding the torque about the precession axis (Zf = Zb) (4) Here Ixx, Iyy, Izz represent the mass moment of inertia about the flywheel fixed axis, spin axis and precession axis, respectively. represents the flywheel spin rate and represent the roll and pitch angular velocities of the AUV body, respectively. τp, β represent the torque about the precession axis and the precession angle, respectively. Equation (3) shows that with AUV roll and pitch motions () a gyroscopic precession () and a precession torque (τp) is experienced. That is, with precession motion and precession torque, the precession axis can be used to drive a generator and produce power. 2.2 System advantages (and disadvantages) The system utilises the reaction of a spinning wheel under an input torque (AUV motion) to provide an output torque of greater magnitude, orthogonal to the input torque axis and the spin axis to drive a generator. This provides a means to harvest energy in-situ with no external appendages, avoiding any additional hydrodynamic drag. Furthermore, with no direct exposure to the marine environment the system is not susceptible to environmental performance degradations, i.e. bio fouling and would not be limited to daylight recharging and night time missions. Potentially the system could be applied to any rotational excited platform(s), e.g. ASVs, boats, buoys and WECs. In practice as the effect of surface waves and swell diminishes with depth [54], similar to solar-based AUVs (e.g. SAUV II), the AUV system would need to surface to recharge, as depicted in Fig. 1 – admittedly exposing the AUV to the potentially hazardous wave environment. However, as waves are a concentrated form of solar energy (formed by winds passing over bodies of water created by the differential heating of air masses by the sun on the earth's atmosphere) with a reported spatial concentration of time-averaged power flow of typically 0.1–0.3 kW/m2 (solar) to 0.5 kW/m2 (wind) to 2–3 kW/m2 (wave) [55], greater energy capture is anticipated compared to solar and wind strategies. 2.3 Prototype design The prototype design is illustrated in Figs. 2a–c. Based on the majority of AUV systems, a torpedo style AUV, with a cylindrical body design was selected. The body diameter (≃ 30 cm) was approximately based on the REMUS 600 AUV, see Table 2. To minimise the swept volume and mass, the flywheel was designed to precess about the central axis of the flywheel, with the flywheel thickness comparable to the flywheel diameter, with an I-shaped cross section. The gyroscopic system was designed to allow full, unrestricted 360° precession motion. This was achieved by using a slip ring mounted on the precession axis to provide power to the spin motor. To convert the mechanical power to electrical power a DC generator (motor) was employed. A discussion of the generator efficiency is given in Section 4.5. A summary of the AUV, gyroscopic unit and generator particulars are presented in Table 3. Table 3. System particulars AUV system particulars Value overall length (Lb), m 2 diameter, m 0.2929 assembled AUV prototype mass, kg 110.3 trim angle, deg. 11.8 Gyroscopic system particulars Value gyroscopic unit mass (including end plate), kg 11.2 gyroscopic unit mass (excluding end plate), kg 6.3 flywheel mass, kg 3.6 flywheel diameter, m 0.1 flywheel mass moment of inertia (Iyy), kgm2 0.00482 flywheel mass moment of inertia (Ixx = Izz, kgm2 0.00363 Generator system particulars Value speed constant, rpm/V 1040 torque constant, mNm/A 9.18 terminal resistance, Ω 1.01 speed/torque gradient, rpm/mNm 114 gearing (gyroscopic precession to generator shaft) 1:73.3 resistor (across terminals) (Vr), Ω 15 3 Experimental setup and investigations A series of regular wave tests over a range of frequencies (0.5–1.4 Hz) and wave amplitudes (0.1 and 0.05 m) were conducted in a towing tank with the gyroscopic system operating (within the AUV) at several spin rates (1000, 5000 and 9000 rpm) and with the system disabled. The prototype was tested following a conventional sea keeping methodology, with the AUV attached to the tow post of the towing tank carriage, constraining the AUV in yaw, roll, surge and sway. For comparison, a series of slack moored tests were also conducted. The experimental setups are illustrated in Fig. 2. The generated data was logged at 100 Hz (using a National Instruments compactRIO). The generated power was calculated by measuring the voltage (Vr) across a resistor (Rr) connected to the terminals of the generator, i.e. . The generator particulars are summarised in Table 3. Fig. 2Open in figure viewerPowerPoint Prototype system and experimental setups (a) CAD assembly and manufactured gyroscopic unit, (b) Experimental tow-post arrangement, (c) Experimental slack moored arrangement 4 Results and discussion The following results are presented in this section: 3: The equivalent harvested energy (Wh) over a range of frequencies and spin rates, per hour. 4: The generated power (peak and rms) over a range of frequencies, spin rates, wave heights and constraints. 5: The non-dimensional pitch response (ϕa/kζa) of the AUV over a range of frequencies. That is, the AUV motion response in waves without the gyroscopic system operating. 6: Example gyroscopic precession angle variation over time. Fig. 5Open in figure viewerPowerPoint Non-dimensional pitch response (ϕa/kζa) of the AUV in regular waves (system off) Fig. 4Open in figure viewerPowerPoint Peak and rms generated power (a) Over a range of frequencies and spin rates (ζa ≃ 10 cm, 5000 rpm and 9000 rpm), (b) Over a range of frequencies and wave heights (ζa ≃ 5 cm and ≃ 10 cm, 5000 rpm), (c) With tow-post and slack moored constraints (ζa ≃ 10 cm, 9000 rpm Fig. 3Open in figure viewerPowerPoint Equivalent harvested energy (Wh per hour) over a range of frequencies, spin rates (ζa ≃ 10 cm) Fig. 6Open in figure viewerPowerPoint Change in precession angle over time ζa ≃ 10 cm, 5000 rpm, 10 wave encounters (a) 0.6 Hz, (b) 0.9 Hz, (c) 1.2 Hz These results are based on ten wave encounters. Example time and frequency domain responses are presented in Figs. 7a–c and 8a–c and flywheel acceleration and deceleration profiles (on start-up and shutdown) are also presented in Fig. 9. Fig. 7Open in figure viewerPowerPoint Example time histories and fast Fourier transforms (FFT) of generated power ζa ≃ 10 cm, 0.6 Hz (a) 1000 rpm time history, (b) 1000 rpm FFT, (c) 5000 rpm time history, (d) 5000 rpm FFT, (e) 9000 rpm time history, (f) 9000 rpm FFT Fig. 8Open in figure viewerPowerPoint Example time histories and FFT of generated power ζa ≃ 10 cm, 1.1 Hz (a) 1000 rpm time history, (b) 1000 rpm FFT, (c) 5000 rpm time history, (d) 5000 rpm FFT, (e) 9000 rpm time history, (f) 9000 rpm FFT Fig. 9Open in figure viewerPowerPoint Flywheel acceleration and deceleration profiles (a) Flywheel spun up to 5000 rpm, (b) Flywheel spun down from 5000 rpm Fig. 3 shows the harvested energy per hour, calculated by integrating the power profiles over ten wave encounters and extrapolating to a one-hour period. The results show that the harvested energy is greatest around resonance, tailing off as the frequency increases. To provide a representative comparison, as the wave amplitudes were found to vary slightly between tests, the generated power (over the ten wave encounters) were normalised with respect to the (maximum) wave amplitude, see Fig. 4. The results show that the power generated is greatest around resonance (≃ 0.7 Hz) and rapidly reduces at higher frequencies (e.g. above 1Hz). Furthermore, the results show that the maximum instantaneous power compared to the rms power is large (up to six times), indicating that the generated power is not sinusoidal. As shown in Figs. 7 and 8, the generated power ‘pulses’, as the flywheel precesses and drives the generator. Interestingly, the pulses can vary in frequency and magnitude over time and are rarely proportional to the (regular) wave frequency. That is, the generated power is typically non-linear. Although this finding is not unexpected (see [56]), the non-linear responses were found to be readily achievable, particularly the case of continuous precession rotation, see Fig. 6a. As a continuous precession motion can be achieved, a continuous power output is possible with the system. With a 1 degree of freedom excitation it is known that when the flywheels orientate themselves in line with the forcing excitation, no gyroscopic precession can occur and no energy can be harvested [56]. In practise, it is anticipated that any slight additional axial excitation will ‘kick’ the system out of these dead zones and this may account for the irregular power output. Furthermore, these results indicate that an optimised restoring term and/or control of the gyroscopic precession (position and rate) could further increase the energy harvested and widen the operating range. 4.1 Spin rate With low spin rates, e.g. 1000 rpm, the gyroscopic torque is small and no significant power is generated. Generally increasing the spin rates, greater powers (peak and rms) were observed. However, the difference between 5000 and 9000 rpm does not appear to represent a significant step change in the generated power. At these spin rates, the system response can become non-linear and it is expected that the flywheel orientation is not always optimally aligned with the wave induced excitation, reducing the energy capture (see Fig. 6b), and that there is an optimal spin rate for a given condition and system setup. 4.2 AUV motion Fig. 5 shows the non-dimensional pitch response, a measure of the pitch amplitude relative to the wave slope, of the AUV prototype. This was calculated from the pitch angular velocity amplitude, i.e. , k = ω2/g. The pitch angular velocity, , was measured with a xSENS MTi-100 inertial measurement unit. The response shows a characteristic resonance response, around 0.7 Hz, with the motion amplitude decreasing with increasing frequency. For the investigated conditions, the AUV motion responses are relatively small, with the largest investigated case (ζa ≃ 10 cm) equating to a peak pitch amplitude response of approximately 7.5°. That is, the test conditions represent a relatively benign sea state and AUV response. As shown in Fig. 4b, greater power can be generated in larger waves. Therefore in a real seaway, where larger waves and responses are expected, it is anticipated that greater power can be generated from the system than presented. 4.3 Experimental setup The prototype was tested following a conventional sea keeping methodology, with the AUV attached to the tow post of the towing tank carriage, constraining the AUV in yaw, roll, surge and sway. Varying the experimental constraint, in Fig. 4c the results were found to be similar in magnitude for a slack moored and tow post constrained setup. The results provide some confidence that the constraint does not significantly influence the system. However, the constraint is believed to influence the system response and a more detailed study is needed to identify the effect(s). 4.4 Energy balance Fig. 9 shows an example, acceleration and deceleration profiles for the prototype. As the system requires power to initially accelerate and then maintain the flywheel spin rate, overcoming friction, an estimate of the input power to accelerate the flywheel and the power to overcome friction and maintain the flywheel rpm was made. Assuming a constant energy loss, the power loss due to friction (to maintain the flywheel rpm) was estimated as (5) where represents the stored kinetic energy in the flywheel. As the prototype system takes ∼470 s for the flywheel to slow from 5000 rpm to rest, see Fig. 9b, the power loss due to friction was estimated as That is, a continuous power input of 1.4 W (estimate) was required to maintain the flywheel at its operating spin rate. Similarly, the input power to accelerate the flywheel was calculated, see Fig. 9a, yielding approximately a 4 W power demand for a 160 s start up time, assuming no losses. Considering the supplied current was limited to 4 A and 24 V (≃ 100 W), or considering the motor rating (40 W), the efficiency of start-up of the prototype system is estimated at between 4 and 10%. 4.5 Generator efficiency Expressing the generator overall efficiency as the ratio of electrical power output, Pe, to the mechanical power input, Pm: (6) and estimating the electrical power and mechanical power delivered to the generator as (7) and (8) respectively, where τp/g and represents the torque and angular velocity at the generator shaft, respectively, and g represents the gearing between the gyroscopic precession axis and the generator shaft. In addition, estimating the magnitude of the gyroscopic torque τp, as (9) Then, the estimated mechanical power delivered to the generator shaft (using the rms and maximum measured values for V, and ), as shown in Figs. 10a and b (for a spin rate of 5000 rpm), is between 0.05–4 W rms and 0.4–18 W peak. In comparison the electrical power ranged from 0.02–0.8 W rms to 0.14–3.2 W peak. That is, the estimated efficiency of the prototype generator is estimated to be ≃ 30% (ranging between 14 and 50%), as shown in Fig. 10. Fig. 10Open in figure viewerPowerPoint Generator efficiency (a) Mechanical power (Pm) and electrical power (Pe), (b) Percentage efficiency In the tested conditions, the prototype system generated power up to 0.8 rms (3.2 peak) W with and up to 1.24 rms (8 peak) W with (see Fig. 10). This shows that the system can harvest energy, however, given the estimated power loss due to friction in the tested conditions a net gain may not have been achieved. Given that the investigated conditions and responses were relatively small and that the estimated mechanical energy delivered to the generator shaft is typically greater than the friction loss (see Fig. 10). Then a net gain appears readily feasible with operation in larger waves and/or improved generator efficiency [e.g. if ηpto = 100% (instead of ≃ 30%) then the system could be expected to generate up to 2.67 rms (10.67 peak) W with and up to 4.12 rms (26.67 peak) W with ]. This could be achieved by controlling the torque and rpm delivered to the generator, for example through the use of a controlled, variable gearing between the precession axis and generator shaft. Compared to other energy scavenging systems, Table 4, and AUV powering requirements, Tables 1 and 2, the results are very encouraging. Table 4. Comparison to other energy scavenging systems Prototype gyroscopic system Power Notes power (actual) up to 0.8–1.2 W (rms) (note: ζa < 10 cm) (ignoring frictional losses) 3.2–8 W (peak) power per wave amplitude up to 20 W/m (rms) , ζa ≃ 10 cm (ignoring frictional losses) 80–115 W/m (peak) power (maximum – if ηpto = 100%) up to 2.6–4.1 W (rms) (note: ζa < 10 cm) (ignoring frictional losses) 10.7–26.7 W (peak) Other systems Power, W Notes SOLO-TREC (thermal) [38] 1.6 6100 J per dive, assumed one-hour dive SAUVII (solar) [57] 7.5–22.5 300–900 Wh per day, scaled to AUV footprint (0.6 m2) waveglider (solar) [36] ≃ 8.6 solar panel peak power, assumed 20% efficiency and scaled to AUV footprint (0.6 m2) Forgen 500NT (small wind turbine) [58] 10 nominal output, 15 knots, 7 kg, 200 mm diameter Autonaut (solar) [35] ≃ 10 solar panel peak power, assumed 20% efficiency and scaled to AUV footprint (0.6 m2) C-enduro (solar) [37] ≃ 14.2 solar panel peak power, assumed 20% efficiency and scaled to AUV footprint (0.6 m2) leading edge LE-v50 (small wind turbine) [58] 14 nominal output, 15 knots, 9 kg, 290 mm diameter Forgen 1000NT (small wind turbine) [58] 15 nominal output, 15 knots, 15 kg, 300 mm diameter 5 Conclusions This paper describes a novel prototype energy scavenging system for a torpedo style AUV. The system, based on CMG principles, provides a means to harvest (wave) energy in-situ, internally, without external appendages. A description of the system operation, design and experimental results from a series of regular wave tests conducted at zero speed in a towing tank are presented. The results show that the system can harvest energy, although in the tested conditions a net energy gain was not achieved. In the tested conditions (ζa ≃ 10 cm), equating to a maximum (AUV) pitch response of 7.5°, the prototype system harvested the equivalent of between 0.05 and 0.6 Wh per hour, with a maximum recorded peak power of 8 W. The generated power was found to be the greatest around resonance, tailing off as the wave frequency increased. Typically, greater spin rates and wave amplitudes yielded greater power with the response becoming increasingly non-linear. The experimental results show that the system has the potential to provide additional hotel load or power specific systems on an AUV or similarly rotationally excited platform, e.g. ASVs, buoys or boats. 6 Acknowledgment This research was funded by the Defence Science and Technology Laboratory (DSTL) through the Centre for Defence Enterprise (CDE). 7 References 1Reader, G., Potter, J., Hawley, J.: ‘ The evolution of AUV power systems’. OCEANS ’02 MTS/IEEE, 2002, vol. 1, pp. 191– 198 2Griffiths, G., Jamieson, J., Mitchell, S., et al: ‘ Energy storage for long endurance AUVs’. Proc. of Advances in Technology for Underwater Vehicles, London, UK, 2004, pp. 8– 16 3Griffiths, G., Reece, D., Blackmore, P., et al: ‘ Modeling hybrid energy systems for use in AUVs’. Proc. 14th Unmanned Untethered Submersible Technology (UUST), 2005 4Paroscientific series 9000 pressure instrument. Available at: http://www.auvac.org/sensors/view/23?fromsearch=1 5Dc-4e digital compass. Available at: http://www.spartonnavex.com/product/dc-4e/ 6Vale