Wind Turbines and Radar - The Radar Cross Section RCS a Useful Figure for Safeguarding?

Transcription

1 1

2 windturb3.dsf 3/7 Wind Turbines and Radar - The Radar Cross Section RCS a Useful Figure for Safeguarding? Gerhard Greving, Wolf-Dieter Biermann, Rolf Mundt NAVCOM Consult Ziegelstr. 43 D-7167 Marbach navcom.consult@t-online.de Keywords: Radar, RCS, wind turbines, ground, safeguarding distance Abstract Wind turbines WT are often to be located in some distance to ground based navigation, landing and radar systems. Unaccetable distortions have to be avoided. The distance to the system has to be determined in some way often done by the RCS scheme for radar systems. The Radar Cross Section RCS is defined for lane wave excitation. It is a useful arameter for objects of limited size in sace such as the aircraft or other flying objects (high) above ground. Basic fundamental asects are discussed by theoretical and numerical results. Objects on the ground, such as wind turbines, cannot be described er definitionem by the RCS. Concetional and numerical results are resented for the RCS of tyical large wind turbines in different distances and for various heights of the radar above ground. The RCS has a sectrum in general and is satially variant and also timely variant if the blades are rotating. Basic effects of the RCS above ground are shown for sherical reference objects where the standard RCS is constant. It is concluded that the RCS cannot be used strictly seaking for the task of safeguarding. Introduction Wind Turbines and Radar Wind turbines WT are constructed today more and more as art of the renewable energy rogram of governments worldwide. Its otimum location deends on the regional best wind scenario, but many locations deend also on the constraints and conditions of the investors. By that the locations of the WT are often more and more in some distance to existing locations of systems, i.e. navigation, landing, communication and radar systems (Fig. 1). In the context of this aer, the radar cross section RCS and its usefulness in the analysis of the imact of the WT on the radar is discussed. The radar systems can be ATC-radar, air defense radar and weather radar. Fig. 1: Wind turbines in the radiation field of systems Introduction RCS The Radar Cross Section RCS is defined for lane wave excitation [1]. It is a useful arameter for objects of limited size in sace such as the aircraft or other flying objects (high) above ground. It is common to characterize the objects for the radar by the RCS (σ, radar cross section; mono-static, bi-static). The general definition of the RCS (1) [1] assumes an asymtotic infinite distance. That imlies the lane wave excitation or a real far-field aroximation [1],[],[3],[4]. σ q = lim 4πR R Wind turbines in the radiation field of systems Radar-, Navigation-, Landingsystems, Communications ASR/PSR E E s i q (1) The obligatory limit condition R imlies a lane wave excitation as exlained also in the IEEE definition of terms [1]. The lane wave is characterized by constant amlitude and by a linearly rogressing hase across the object. The natural consequence is that all tools for the RCS require the inherent lane wave source and in the measurements a lane wave has to be aroximated. A ground lane is rejected thereof in the RCS calculation. SSR/MSSR IFF DME PAR ADR HADR RAT-31S TACAN GPS ILS MLS VOR/DVOR ADS-B weather radar S/C NDB VHF/UHF com WAM radio relay ADF/UDF effects / distortions / conditions? 539

3 An aroximate RCS is ublished for erendicular incidence and A >>λ of a square metal late. σ = () A misleading intuition and interretation of equation () seems to suggest that a metal late of double area will have a double RCS just by adding the individual RCSs of objects such as arts of a WT or of WT in windarks. Assume a square late of size A 1 and a second late of A =A 1. Alying the aroximate formula () σ 1 = πa1 / 4 λ = πa / λ = 4π (A1 ) / λ 4 σ 4 = σ (3) it can be readily seen that a metallic late of double size yields a four times larger RCS. The resulting RCS of the larger doubled area is not the sum of the individual RCS. The reason for that is that the basic equation contains the ower ratio by the squared fieldstrengths. The following modeling and rigorous numerical calculations shall visualize that fact. A square metal late of m by m is treated in 3 different ways for the S-band radar frequency of 3GHz: 1. RCS function of the full total late within ±1. RCS of half late each and adding the RCS 3. RCS of a quarter late each and adding the 4 RCS irs7_rcs1.dsf 7/7 RCS [db m ] πA / λ RCS Radar Cross Section of a square metal late total late sum of half lates half late each sum of 4 quarter lates quarter late each 6dB 1dB frequency: 3GHz vertical olarisation Ez late in yz lane full late: m x m half late: 1m x m quarter late: 1m x 1m method: IPO Angle ρ[ ] (ϑ=9 ) Fig. : RCS calculation of a square metal late and subdivisions into or 4 arts. It can be easily seen from Fig. that the added RCS is smaller by 3dB or 6dB resectively. A next rincial RCS-calculation (Fig. 3) shows the RCS for a rotated square late of the same size and a second modified version where two quarters are 3dB 1 RCS 1 i symmetrically back setted by a quarter wave length. It can be nicely seen that at broadside the modified late has a numerical minimum of more than 5dB where the normal late has the maximum. On one hand this last result is not surrising because the RCS is the result of the normalized scattering rocess and is by that to be calculated by the real scattering attern which includes the interference effects. On the other hand both examles show drastically that a simle addition of the individual RCS is not ossible. The electrical fieldstrengths in () have to be added u in a comlex vectorial way and have to be rocessed. irs7_rcs.dsf 7/7 RCS [db m ] Angle ϕ [ ] (ϑ=9 ) Fig. 3: RCS of a rotated square late (45 ) flat and modified by back setting quarters by λ/4 In [] a rincial method is roosed to suerose the individual RCS and to determine the so-called coherent RCS j e Φ σ = Σ σ (4) RCS of a rotated square late (flat, modified) z + - E z x quarter lates in different lanes 1m y 1m λ/4 full late 4 quarter lates; different lanes single quarter late frequency: 3GHz vertical olarisation Ez lane wave all lates in yz lane and rotated by 45 full late: m x m quarter lates: 1m x 1m 4 quarter lates in lanes distance between lanes: λ/4 method: IPO The roblem here is the hase term Φ - where to take the reference oint in the general case of extended electrically large objects and tyical higher radar frequencies. By that it does not seem to be ractical. Arbitrary results can be achieved. The noncoherent RCS [] neglects the hase term and the root and square function and is, by that, highly questionable and not alicable in any case for relatively small numbers of objects. Conclusions for the basic RCS: 1. The evaluation of the RCS requires the excitation by a lane or an aroximately lane wave.. The RCS is based on the scattered field. 3. RCS figures cannot be simly added. 4. The rojected area as seen from the radar is not a measure for the RCS in general. 5.6dB 1dB 54

4 H N H max RCS and Wind Turbines WT in General As outlined above, the fundamental assumtion in the definition of the RCS is the excitation by a uniform lane wave. The wind turbines WT are naturally installed on the ground and, by that, the ground interactions have to be taken into account if the ground is significantly illuminated. In case of encil beam antennas, such as 3D air defence radar or weather radar the lowest beam ositions are relevant for the WT in most cases only. FieldStrength[dBV/m] dist. = 3m dist. = 5m dist. = 1m dist. = m dist. = 3m Field Strength Above Ground source height 3m, elevation angle frequency: 5.64 GHz (C-band, λ=.53m) source attern: encil beam: az.1.,el.1. (3dB-width);SL-8dB olarization: horizontal 1V/m (dbv/m) at 5m for the main beam in free sace Calculation Method: MoM ground: ε r =1, κ=.1 S/m Height above Ground [m] Cbandrefscatatdist1e8 blades nacelle windturbine in the exciting radar field H max >> H D >> H, H max RCS? radground.dsf 7/7 Fig. 5: Exciting field at the location of the WT; C-band source 3m above ground; nacelle height 19m. shaft direct direct rescattered field of the windturbine radar -8-9 shere: diameter.8m ka=16.5, σ/(πr )=1. height: 1m (fixed) Scattering of a Shere above Ground distance 5m; var. source heights (h) scattered field strength (co-olar) at the hase center of the source antenna ground reflected ground D ground reflected Fig. 4: Schematic setu of the radar and the WT, direct and ground reflected rays The effects of the ground are twofold, for the excitation caused by the radar at the WT (Fig. 4) and for the echo resonse scattered by the WT (Fig. 4) at the location of the radar. The illuminating field across the object does not have a constant amlitude and does not have a constantly rogressing hase. The radar has some height above ground and the WT effectively also (Fig. 4). Fig. 4 shows the non-uniform illumination schematically. Fig. 5 shows the numerically calculated variable exciting field at the location of the WT for different distances of a C-band weather radar from 3km to 3km. In the close distance of 3km, the encil beam antenna does not even illuminate the WT comletely. By all that, two sources of fundamental errors occur if the RCS-scheme is alied to the WT-case, namely the non-existence of the lane waves for the validity of the RCS and by that an erroneous illumination is assumed (Fig. 5) the distorted lobing field amlitude in the backscattered resonse at the location of the radar (Fig. 6 and Fig. 7). The amount of error of the RCS-scheme comared to the reality is unredictable. The WT-radar-scenario has to be analyzed by aroriate tools and adequate numerical methods in a deterministic sense case by case. H not scaled Field Strength [dbv/m] Y [m] X [m] ground; el= Z[m] frequency: 5.64 GHz (C-band, λ=.53m) source attern: encil beam: az. 1., el. 1. (3dB-width); SL -8dB olarization: horizontal 1V/m at the object distance for the main beam in free sace Calculation Method: MoM ground: ε r =1, κ=.1 S/m Source Height above Ground [m] Cbandshere_scatatd5rh1-4b Fig. 6: Back scattered signal at the location of the radar having a variable height by a metallic shere at 1m height above ground; distance of the shere 5m FieldStrength[dBV/m] Scattered Field of a Wind Turbine above Ground distance 5m; var. source heights, elevation wind turbine: ENERCON E7 rotor diameter 7m nacelle height 65m shaft bottom 6m shaft to m rotor blade itch yaw 1 Z[m] source - Y[m] co-olar electric field at the source osition frequency: 3 GHz (S-band, λ=.1m) source attern (encil beam): az. 1., el. 1. (3dB-width); SL -8dB olarization: horizontal elevation: 1V/m at the object distance for the main beam in free sace calculation Method: IPO ground: ε r =1, κ=.1 S/m Source Height above Ground [m] Sband_E7_Scat_5kmry Fig. 7: Back scattered signal at the location of the radar of variable height by comlete WT above ground, nacelle height 86m, distance of the WT 5m One could argue that the RCS-treatment of objects above ground would constitute the worst case. First, this is not roven and, second if so, this concet would enalize the siting of the WT and would yield unjustified large safeguarding distances. 541

5 If searately and indeendently calculated or measured RCS are used, the discussed effects are not taken into account. The scattering behavior of an object above ground is fundamentally different from the inherent free sace condition for the standard RCS. Conclusions for the alication of RCS: 1. The standard numerically calculated or measured RCS in free sace does not describe the hysical effects for the scattering at a WT above ground.. The system errors are unredictable. 3. The worst case concet is not justified. The RCS of a Wind Turbine Although the RCS is not a justified scheme for the evaluation of the effects of wind turbines on radar in real scenarios, some detailed results for the standard RCS shall be resented and discussed in the following chater. A lane wave excitation is assumed. single RCS is a characteristic figure for the WT assuming that the RCS would be alicable at all. Fig. 8 shows the 3D model of a large WT to be analyzed by modern imroved numerical methods [3,4,5-8]. The 3D-model is discretized and comosed of a large number of metallic triangles. This is justified desite the dielectric material of glass fibre for the blades. Under strong rain conditions the water layer is almost erfectly reflecting for tyical radar frequencies (L-, S-, C-band) RCS of a Wind Turbine (E7) 3 33 φ Enercon E7 blade diameter: 7m blade itch: tower diam. b/t: 6.1m/m blade rotation: 1 yaw angle: Windgenerator Enercon E66 1.8/MW 3D-model ca m 148 Patches / 11MHz ILS/VOR 1888 Patches / 13MHz SSR Sband_E7_RCSco1 Fig. 9: RCS of a large WT in the azimuthal lane time variant scattering attern of blades generator house blades ca m α If rotorlane faces DVOR almost no Doler-shift Dolershift- Frequency of the scattered fields The Fig. 9 and Fig. 1 show that the mono-static RCS of a WT in free sace is extremely structured and lobed due to the electrically large size of the WT. The dynamic is at least 5dB, ranging from >7dBm in the elevation lane when the view angle is orthogonal to the metallic surface of the mast down to dbm and smaller in minima. The RCS is also very sensitive for the satial direction. By that again, strictly seaking the RCS is different for the direct signal and the ground reflected signal also if one would use the image theory for two lane waves. Large amlitudes in narrow eaks ( flashes ) and interference are suerosing an avca. 4m f D = f Dmin f D = shaft ca. m f Dmax f D = quasistatic slow, wind direction <ca. /min v max = ca. 3km/h v r f T c windgenere66a.dsf 6/4 Fig. 8: 3D-model of a WT for numerical evaluation The RCS as a single figure is sometimes requested as a basis for the safeguarding distance of the radar station. The RCS can be measured by scaled modeling or can be modeled and simulated by adequate tools. But the question arises which r fd Fig. 1: RCS of a large WT in the elevational lane RCS of a Wind Turbine (E7) Elevation (±5deg) θ Sband_E7_RCScoel Enercon E7 blade diameter: 7m blade itch: tower diam. b/t: 6.1m/m blade rotation: ("A") yaw angle: 54

6 φ Sband_E7_RCSco 3 33 θ Sband_E7_RCScoel Enercon E7 blade diameter: 7m blade itch: tower diam. b/t: 6.1m/m blade rotation: ("A") yaw angle: Elevation (±5deg) Enercon E7 blade diameter: 7m blade itch: tower diam. b/t: 6.1m/m blade rotation: ("A") yaw angle: erage RCS generated by the conical frequently metallic shaft. The RCS of lattice tye shafts is tyically much lower. Fig. 11 shows a statistical evaluation of the RCS in the azimuth lane (Fig. 9) lus the maximum value taken from the elevation lane. The eak is some average generated by the mast as seen from that angle. frequency [%] Fig. 11: Statistical frequency distribution of the RCS of Fig. 9 lus the maximum of Fig. 1 in the elevation lane Electric Field [dbv/m] RCS Frequency Distribution of a Wind Turbine "Enercon E7" horizontal lane, elevation max Scattering Pattern (Coolar) of a Wind Turbine Above Ground Varying Rotor Position; Source height 3m, Turbine Distance 5m horizontal vertical olarization shaft+nacelle h shaft+nacelle v frequency: 3 GHz (S-band, λ=.1m) source attern: encil beam: az.1.,el.1. (3dB-width);SL-8dB 1V/m at the object distance for themainbeaminfreesace calculation Method: IPO ground: ε r =1, κ=.1 S/m wind turbine: ENERCON E7 rotordiameter7m nacelle height 65m shaft bottom 6m shaft to m blade itch yaw Rotor Angle [deg] Time [sec] ( RPM;.73sec=36 ) Sband_hv_E7_SCAT5q3rotr Fig. 1: Scattering resonse of a large WT from some asect angle for a rotation cycle of 1 of the trile blades; distance 5km Fig. 1 shows the scattering resonse of a large WT above ground for rotation cycle of 1 of the trile blades. The scattering resonse is almost indeendent of the vertical/horizontal olarization. Again it is remarkable that the dynamic is more than 5dB. Fig. 13 shows a comarison of the scattering resonse for a fixed radar geometry (height 3m) and different distances, 5km and km. Large differences can be observed. In the RCS-scheme the cases would be identical. A further interesting technical toic of the WT with regard to the ulse-doler-radar and the RCS is that of the rotating blades. The blades may rotate u RCS of a Wind Turbine (E7) RCS of a Wind Turbine (E7) maximum of elevation Sband_E7_RCS_hdb18_histo1 to /min (Fig. 8). By that their radial velocity may be u to 3km/h at the tis and a high Doler-shifted frequency sectrum will be created by the rotation (Fig. 14 ±1.5kHz; u to ca. 3kHz for a C-band radar). Electric Field [dbv/m] Fig. 13: Scattering resonse of a large WT from some asect angle for a rotation cycle of 1 of the trile blades; distances 5km and km The amlitudes of the Doler shifted sectral signals deend on many factors such as the orientation of the WT and the back scattering roerties of the blades. However, in any case the Doler-shifted back-scattered signal reresents a continuous sectrum and contains ositive and negative Doler shifts (Fig. 14). rel. Electric Field [db] WTE8_dol.dsf 1/ Scattering Pattern (Coolar at Source) of a Wind Turbine Above Ground Varying Rotor Position; Source height 3m, WT Distance: 5km,km Rotor Angle [deg] Time [sec] ( RPM;.73sec=36 ) Z[m] Distance 5km km Doler shift sectrum of back scattered signal WT E8 Sectrum at radar antenna osition (monostatic) Wind Generator Blades and Nacelle Seed: rm Y[m] Frequency 3GHz Pencil beam 1 Radar height 3m Distance 5km rm from Wind Generator without Blades frequency: 3 GHz (S-band, λ=.1m) source attern: encil beam: az. 1., el. 1. (3dB-width); SL -8dB vertical olarization 1V/m at the object distance (5km) for themainbeaminfreesace calculation Method: IPO ground: ε r =1, κ=.1 S/m Fig. 14: Scattering Doler-sectrum resonse of a large WT from some asect angle for a rotation cycle of 1 of the trile blades in 4 ositions; distance 5km By that again, the simle stationary monostatic RCS of the blades is not reresentative for the rotating blades. In fact, the RCS is distributed and much reduced by the Doler-sectrum sread. One can define a RCS-frequency-function in dbsm/hz. One can understand that easily since only a small subart of the blades creates the related Doler frequency and not the total blade (Fig. 8). Only in case of the non-rotating stationary blades the total blades X[m] Wind Generator Blades and Nacelle Doler Shift [Hz] Y[m] - Band Width 1Hz wind turbine: ENERCON E7 rotor diameter 7m nacelle height 65m mast bottom 6m mast to m blade itch yaw Sband_hv_E7_SCATq3rotr Antenna Weather Radar Position (x/y/z): m/m/3m Pencil Beam (-3dB Az/El=1 /1 ) Horizontal Polarisation Elevation =.5 Position: x/y/z = m/m/3m Frequency= 3GHz Wind Generator Enercon E8 Position (x/y/z): 5km/m/m Nacelle Height: 18m Rotor Diameter: 8m Yaw Angle: 45 ( =Blades in YZ-lane) Rotor Angle: /3 /6 /9 Rotation: rm Ground Parameter: ε r =1, σ=.1 S/m W_do_sh_E8_5km_1Hz_u 543

7 contribute to the Hz-signals which are suressed by the MTI/MTD mechanism by 4dB minimum in case of modern radar. Conclusions for the RCS and WT 1. The RCS of a WT has a dynamic of more than 5dB in the azimuthal lane. A total dynamic in sace of more than 7dB has been observed.. The shar maximum RCS is created by the metallically assumed blades in the azimuthal lane and low elevation angeles. 3. The largest RCS aear for erendicular incidence to the metallic mast. 4. The RCS has a broad Doler sectrum in the rotating case. The sectral lines are lower for faster rotations. 5. In the non-rotating case the RCS shows a broad frequency distribution. The statistical maximum is defined by the mast. 6. A single RCS-figure is imossible to define for a WT from a reasonable engineering oint of view. [5] GREVING G. Modern threats to recision aroach and landing - The A38 and windgenerators and its adequate numerical analysis, ISPA 4; October 4, Munich/Germany [6] GREVING G. Analysis of the scattering fields for ATC radar by objects - Consequences of the alication of different methods, IRS 5, Berlin/Germany [7] GREVING G. Numerical Analysis of the effects by scattering from objects on ATC-radar and various methods for its reduction - Theory, results, IRS 6, Krakow/Poland [8] GREVING G. Numerical Simulations of Environmental Distortions by Scattering of Objects for the Radar - SSR and flat roofs, RCS and windturbines, EURAD 6, Setember 6 Manchester/UK [9] GREVING G., Malkomes M. On the Concet of the Radar Cross Section RCS of Distorting Objects like Wind Turbines for the Weather Radar, 4 th ERAD 6, Setember 6, Barcelona/Sain Final Conclusion, Summary The concet of the RCS is not alicable for objects on the ground if the ground is significantly illuminated by the radar. Fundamental theoretical and hysical reasons rohibit the alicability of the RCS in the case of objects on the ground. By that it is not a useful figure and cannot be used also as a ragmatic aroximation and alication for safeguarding the radar with resect to the WT. Since the RCS measured or calculated in free sace for the required lane wave condition is not reresentative for the WT, these results cannot be used for the analysis and treatment of the safeguarding task. The error made in this scheme is not defined. Generally as a tendency, the safeguarding zones will be much too large when using the standard RCS deending on the single RCS-figure chosen. The general scattering 3D-case has to be modeled and analyzed taking into account the given 3D-geometry, the antenna radiation atterns and the ground. References [1] IEEE STD , Definition of terms [] SKOLNIK Radar Handbook, McGraw Hill, Boston 199 [3] LO, LEE Antenna Handbook I, Chaman&Hall, NewYork 1993 [4] KNOTT et.al. Radar Cross Section, Artech House, Boston