EP1886166A1

EP1886166A1 - Method for measurement of air speed by doppler radar

Info

Publication number: EP1886166A1
Application number: EP06755534A
Authority: EP
Inventors: Pierre Tabary; Laurent Perier
Original assignee: METEO-FRANCE; METEO FRANCE
Current assignee: METEO-FRANCE; METEO FRANCE
Priority date: 2005-06-02
Filing date: 2006-05-30
Publication date: 2008-02-13
Also published as: FR2886737A1; US20080211714A1; FR2886737B1; WO2006129006A1; US7579977B2

Abstract

The invention relates to a method for measurement of air speed in an atmospheric region by means of the Doppler effect using radar, comprising the following steps: transmission of bursts of three pulses at different rates F1, F2, F3, determination of a speed V1, V2, V3 for the air from the reflected pulses of each burst and calculation of the air speed V from the air speeds V1, V2, V3 determined by the reflected pulses for each burst.

Description

DOPPLER radar air velocity measurement method.

The present invention relates to ^"a method of measuring the air velocity in a region of the atmosphere.

BACKGROUND OF THE INVENTION In meteorology, it is known to measure the air velocity by means of a radar using the Doppler effect: the radar transmits sinusoidal wave trains or pulses which are returned to the radar by particles suspended in the air; if the air has a movement causing the particles to move away from or towards the radar, the pulses returned to the radar have a phase shift with respect to the pulses emitted, an offset which makes it possible to calculate the radial velocity of the particles relative to the radar and therefore the speed of the air that carries them. The speed can thus be determined without ambiguity if the real speed of the particles is within a so-called Nyquist interval depending on the rate of emission of the pulses (also called emission frequency). If the actual velocity of the particles is outside this range, the calculated velocity is equal to the actual velocity modulo the width of the Nyquist interval. We then speak of a folding in the Nyquist interval of the calculated speed. To increase the width of this interval, it is known to increase the pulse transmission rate. However, there are a number of drawbacks, in particular a strong solicitation of the transmitter, a significant consumption of the latter and a reduction in the range of the radar.

It is also possible to use a radar of longer wavelength. Such a radar is however expensive.

It is furthermore known to emit pulses of pulses according to a first and a second rate pulse emission, the substitution of one rate to the other speaker after each burst (process "dual pulses repetition frequency - PRF" or double pulse repetition frequency). By combining the velocities calculated from the pulses received from the pulses transmitted in successive bursts, the particle velocity can be determined unambiguously in an expanded Nyquist interval. However, as the radar antenna rotates continuously, the atmosphere zone to which a burst is emitted at the first transmission rate is slightly different from that to which the next burst is transmitted at the second transmission rate. This results in an inaccuracy of the determination of speed all the more important that the radar is located in an area where the air velocity has large local variations and the speed of rotation of the radar is high.

OBJECT OF THE INVENTION It would therefore be advantageous to have a means for precisely determining the air velocity with a relatively large Nyquist interval.

BRIEF DESCRIPTION OF THE INVENTION

For this purpose, according to the invention, there is provided a method for measuring the air velocity in a Doppler zone of the atmosphere by means of a radar, comprising the steps of:

- emit gusts of three pulses at different rates F _x , F ₂ , F ₃ ,

determining a speed V ₁ , V ₂ , V ₃ of the air from pulses received in return from the pulses of each burst,

calculating the air velocity V from the velocities V ₁ , V ₂ , V ₃ determined for the pulses received in return from each burst. Thus, each pulse is emitted at a rate different from that of subsequent pulses. The Nyquist interval is obtained by combining the three emission rates so that the Nyquist interval is relatively wide. The pulses at the three emission rates are emitted successively to the same zone of the atmosphere, which limits the imprecision of the process.

Preferably, the calculation of the velocity V of the air comprises the phases of:

calculate the Nyquist velocities V _nl , V _n2 , V _n3 corresponding to each rate Fi, F ₂ , F ₃ and the speed of

Nyquist equivalent V _{neg /}

for each value of an integer k varying in the interval [-V _neq / 2V _nl +%; V _neq / 2Vm +%],

. calculating a speed V = Vi + t _your 2kV _n i,. fold the speed V _test in the intervals

[-V ₁₁₂ , V ₁₁₂ ] and [-Vn ₃ , V _n3 ] to obtain the speeds V ₂ 'and V ₃ ',

. calculate the deviations ΔV ₂ = V ₂ '-V ₂ and ΔV ₃ = V ₃ ' - V ₃ and the mean squared difference + ΔV ₃ ² ) / 2), - remember as speed V the _test speed V corresponding to the smallest mean square deviation.

This calculation method proves to be relatively reliable and easy to implement by computer with relatively low computing resources. Advantageously, the rates F ₁ , F ₂ , F ₃ are relatively close.

This causes only a relatively small load on the radar transmitter and limits its wear.

According to a particular mode of implementation, the method comprises a step of determining the rates

Fi ^ F ₂ , F ₃ com, taking the phases of:

-. determine pairs of parameters p / q and r / s such that p and q and r and s are prime between them, q and s are greater than p and r respectively, p is greater than q / 2 and r is greater than s / 2, select a frequency Fi and calculate a corresponding Nyquist velocity V _n i,

- choose a speed V corresponding to the maximum speed of the air in the measuring zone and fold the speed V in the interval [-V _nl , V _nl ] to obtain the speed V ₁ ',

- for each pair p / q and r / s:

.calculate rates F ₂ = p / q * Fi and F ₃ = r / s * Fi and Nyquist velocities V _{n2 /} V _n3 , .calculate the equivalent Nyquist velocity

V _neq = ppcm (p, r) * V _n i,

for each value of an integer k varying in the interval [-V _neq / 2Vm +%; V _neq / 2Vm +%], * calculate a speed V _test = Vi '+ 2kV _n i,

* fold the speed V _test in the intervals [-V _{n2 /} Vn ₂ ] and [-V _n3 , V _n3 ] to obtain the speeds V ₂ 'and V ₃ ',

* calculate the deviations ΔV ₂ = V ₂ '-V ₂ and ΔV ₃ = V ₃ ' -V ₃ and the mean squared difference E = V ((ΔV ₂ ²

+ ΔV ₃ ² ) / 2),

.Hold _your speed V t corresponding to the lowest standard deviation.

- Compare _your speed V t obtained for all couples and speed V to select the best couple.

This mode of determination is relatively simple, reliable and fast.

Advantageously then, the comparison comprises the phases of calculating for all the couples the difference Δ '= V _test - V. and to check if Δ 'is less than half the speed of Nyquist V _nX .

This mode of comparison combines simplicity and efficiency. More preferably, the method comprises the step to assign to the speeds V ₁ ', V ₂ ' and V ₃ ', before their use in the calculations, a noise corresponding to the noise specific to the radar and to the usual atmospheric conditions in the measuring zone and the step of folding the speeds V ₁ ', V ₂ ' and V ₃ 'and noisy in the intervals [-Vm, Vm] / [-V _n2 , V ₁₁₂ ] and [-V _n3 , V _n3 ] to obtain the speeds V ₁ ', V ₂ 'and V ₃ ' used later in the calculations.

It is thus possible to determine transmission rates that are optimized for the atmosphere zone where the speed measurements are made.

Other characteristics and advantages of the invention will emerge on reading the following description of a particular, nonlimiting embodiment of the invention.

BRIEF DESCRIPTION OF THE DRAWINGS

Reference will be made to the appended drawings, among which:

FIG. 1 is a block diagram illustrating the general course of the process of the invention;

FIG. 2 is a block diagram illustrating the progress of the frequency determination step according to the invention,

- Figure 3 is a diagram illustrating the emission of pulses over time.

DETAILED DESCRIPTION OF THE INVENTION With reference to the figures, the method according to the invention is implemented by means of a Doppler radar capable of emitting pulses of pulses by changing the transmission rate between each pulse. .

For the measurement of the air velocity, the method of the invention starts with the step of emitting gusts of three pulses 1, 2, 3 at different rates F ₁ , F ₂ , F ₃ (the duration t1 separating the pulses 1 and 2 is different from the duration t2 separating the pulses 2 and 3 and the duration t3 separating the pulse 3 from a burst of the pulse 1 of the following burst is also different from the times t1 and t2, see Figure 3).

When an impulse encounters particles suspended in the air, these particles return a pulse towards the radar.

The method is thus continued by a step 40 for determining the velocities V ₁ , V ₂ , V ₃ of the air from the pulses received in return from the pulses 1, 2, 3 of each burst. The calculation of the velocities Vi, V ₂ , V ₃ is known in itself and rests on the following formula: V = Fd * λ / 2 where Fd is the frequency offset of the received pulse with respect to the transmitted pulse (also called Doppler frequency). The speed V of the air is then calculated from the speeds V ₁ , V ₂ , V ₃ determined for the pulses received in return for each burst (step 50).

Calculation of the speed V of the air requires having computed the Nyquist velocities V _nl , V _{n2 /} V _n3 corresponding to each rate Fi, F ₂ , F ₃ and the speed of

Nyquist equivalent V _neg (step 20). It is recalled that the Nyquist speed is equal to the product of the wavelength of the pulse and the transmission rate divided by 4. For example, V _n i = λ * F _x / 4. calculating the equivalent Nyquist velocity V _ne q is produced from the rate ratios relative to the rate F ₁ and the Nyquist velocity V _n1, as explained further in the description of the step of determining the rates. Then, an integer k is varied in the range [-V _neq / 2V _nl +%; V _neq / 2V _nl + ^] and for each value of k in this interval:

.a _your speed V t = V ₁ + 2kV _nl is calculated, .the _your speed V t is folded in the intervals [-V _n2, V _n2] and [-V _n3, V _n3] for speeds V ₂ 'V and _3', V ₂ '= V _test modulo (2V _n2) and V ₃ = V _test modulo (2 V ₁₁₃₎

.the deviations ΔV ₂ = V ₂ '-V ₂ and ΔV ₃ = V ₃ ' -V ₃ and the mean squared difference (ΔV ₂ ² + ΔV ₃ ² ) / 2) are cal- culated.

Then the average quadratic differences obtained for all the values of k are compared with each other and the speed V _te st corresponding to the smallest mean square deviation is selected as the speed V of the air (step 60).

To implement the method of the invention, it is necessary to have previously determined the transmission rates of the pulses F ₁ , F ₂ , P ₃ . The step 10 for determining the rates Fi, F ₂ , F ₃ is detailed in FIG. 2 and begins with a phase 11 for determining pairs of parameter ratios p / q and r / s which will be used to define the rates F ₂ , F ₃ according to Fi. F ₂ = p / q * Fi and F ₃ = r / s * F _x .

Preferably, in order to optimize the search for these pairs, constraints are imposed in the choice of the parameters p, q, r, s:

- p and q as well as r and s are first between them,

q and s are greater than p and r respectively, p / q is greater than r / s,

p is greater than q / 2,

r is greater than s / 2.

The parameters p and r are advantageously greater than q / 2 and s / 2 respectively, to prevent the rates F ₂ and F ₃ are much lower than the rate F ₁ , which may cause wear magnetron radar. ^•

In practice, to limit the number of possibilities, it is possible to limit the value of p to 11. In addition, it was found that for efficiency maximum of the method the parameter q is preferably equal to p + 1.

The step of determining the pulse emission rates continues with a phase 12 during which the selection of a first rate F ₁ depends on the technical characteristics of the radar and the calculation of a Nyquist velocity V _nl. corresponding. A velocity V corresponding to the maximum velocity of the air in the measuring zone is folded in the Nyquist interval [-Vm, V _nl ] to obtain the velocity V ₁ ', that is to say V ₁ ' = V modulo (2 * V _nl ). During phase 13, a noise corresponding to the noise specific to the radar and the usual atmospheric conditions in the measurement zone is assigned to the speed V ₁ 'and the velocity V ₁ ' thus noisy is folded as before in the range [- Vm _/ V _nl ] to obtain the speed V ₁ 'used later in the calculations. The noise added to the speed V ₁ ¹ is a Gaussian distribution noise of zero average and standard deviation parameterizable so that this noise corresponds to that encountered under conditions of use.

The following operations are then performed for each pair p / q and r / s (phase 14):

. calculating the rates F ₂ = p / q * F ₁ and F ₃ = r / s * F ₁ and the corresponding Nyquist velocities V _n2 , V _n3 .

. calculating the equivalent Nyquist velocity V _neq by multiplying the Nyquist velocity V _nl by the smallest common multiple of the parameters p and r (V _neq = ppcm (p, r) * V ₁₁₁ ), for each value of k integer varying in _; the interval [-V _neq / 2V _nl + U; V _neq / 2V _nl +%],

* Calculating a speed V _thy t = V ₁ '+ 2kV _{nl /}

* fold the speed V _test in the intervals [-V _n2 , V _n2 ] and [-V ₁₁₃ , V _n3 ] to obtain the vi- tesses V ₂ 'and V ₃ ',

* as for the speed Vi ¹ , assign to the velocities V ₂ 'and V ₃ ' a noise corresponding to the noise specific to the radar and to the usual atmospheric conditions in the measuring zone and fold the velocities V ₂ 'and V ₃ ' as well noised in the intervals [-V _n2 , V ₁₁₂ ] and [-V _n3 , V _n3 ] to obtain the velocities V ₂ 'and V ₃ ' used in the following calculations, * calculate the deviations ΔV ₂ = V ₂ '-V ₂ and ΔV ₃ = V ₃ ' -

V ₃ and the mean squared difference (ΔV ₂ ² + ΔV ₃ ² ) / 2), hold the _test speed V corresponding to the smallest mean squared difference. The step of determining the pulse transmission rates is completed by comparing the _test speeds V obtained for all the couples and the speed V to select the best torque. This comparison comprises the phases of calculating for all the couples the difference Δ '= V _tes t - V', of comparing the deviations Δ 'and of checking if Δ' is less than half the speed of Nyquist V _nl . To refine the selection of the best pair, it is also possible to compare the average quadratic differences obtained by looking for the pair with the difference Δ 'and the mean squared difference E the lowest or the couple offering the best compromise between these two deviations.

For example, given the atmospheric conditions in France and more particularly the air turbulence in this country, the following parameters give satisfactory results: p = 6, q = 7, r = 4, s = 5 where p = 7, q = 8, r = 2 and s = 3.

Thus, by choosing a rate F ₁ at 375 Hz, one obtains, with the first set of parameters, F ₂ = 321 Hz and F ₃ = 300 Hz. These parameters are of course usable for all areas with conditions similar to those encountered in France.

Of course, the invention is not limited to the embodiment described and variants may be provided without departing from the scope of the invention as defined by the claims.

The frequency determination method can be carried out for several noise levels in order to evaluate the relevance of the selected pairs with respect to the noise levels encountered.

The numerical values are given for information only and other values are of course usable.

Claims

A method for measuring air velocity in a Doppler-effect zone of the atmosphere using a radar, comprising the steps of:

- emit gusts of three pulses at different rates F ₁ , F ₂ , F ₃ ,

calculating the air velocity V from the velocities V ₁ , V ₂ , V ₃ determined for the pulses received in return from each burst.

2. Method according to claim 1, characterized in that the calculation of the speed V of the air comprises the phases of:

- calculate the Nyquist speed V _D, V _n2, V _n3 corresponding to each rate Fi, F _2, F ₃ and the speed V _neg Nyquist equivalent, - for each value of an integer k varying in the interval [ -V _neg / 2V _nl + y,; V _neq / 2V _nl +%], calculate a speed V _test = V ₁ + 2kV _nl , fold the speed V _test in the intervals [-V _n2 , V _n2 ] and [-V _n3 , V _n3 ] to obtain the speeds V ₂ 'and V ₃ ', calculate the deviations ΔV ₂ = V ₂ '-V ₂ and ΔV ₃ = V ₃ ' -V ₃ and the mean squared difference (ΔV ₂ ² + ΔV ₃ ² ) / 2),

- remember as velocity V the velocity V _te s _t corresponding to the smallest mean square deviation.

3. Method according to claim 1, characterized in that the rates F ₁ , F ₂ , F ₃ are relatively close.

4. Method according to claim 3, characterized in that the rates F ₂ and F ₃ are respectively 6/7 * F ₁ and 4/5 * F _x .

5. Method according to claim 3, characterized in that the rates F ₂ and F ₃ are respectively 7/8

* F _x and 2/3 * F ₁ .

6. The method of claim 1, characterized in that it comprises a step of determining the rates F _1, F _2, F ₃ comprising ^the phases are:

- determine pairs p / q and r / s such that p and q and r and s are prime between them, q and s are greater than p and r respectively, p is greater than q / 2 and r is greater than s / 2,

- Select a rate Fi and calculate a corresponding Nyquist V _nl speed,

- choose a speed V corresponding to the maximum speed of the air in the measurement zone and fold the speed V in the interval [-V _nl , V _nl ] to obtain the speed V ₁ ',

- for each pair p / q and r / s:

. calculate rates F ₂ = p / q * F ₁ and F ₃ = r / s * F ₁ and the velocities of Nyquist V _n2 , V _n3 ,

. calculate the equivalent Nyquist velocity

Vneq = ppcm (p, r) * V _nl ,

. for each value of an integer k varying in the interval [-V _neq / 2V _nl +%; V _neg / 2V _nl + H],

* Calculating a speed V _thy t = V ₁ '+ 2kV _nl,

* fold the speed V _test in the intervals [-V _n2 , V _n2 ] and [-V _n3 , V ₁₁₃ ] to obtain the velocities V ₂ 'and V ₃ ', * calculate the deviations ΔV ₂ = V ₂ '-V ₂ and ΔV ₃

= V ₃ '-V ₃ and the mean squared difference remember the speed V _test corresponding to the smallest mean squared difference. - compare the V _tegt speeds obtained for all couples and speed V to select the best torque.

7. Method according to claim 6, characterized in that the comparison comprises the phases of calculating for all couples the difference Δ '= V _test - V' and to check if Δ 'is less than half the speed of Ny - quist Vm.

8. Method according to claim 6, characterized in that the comparison comprises a comparison phase of the average quadratic differences obtained for all couples.

9. The method of claim 6, characterized in that it comprises the phase of assigning to the velocities V ₁ ', V ₂ ' and V ₃ ', before their use in the calculations, a noise corresponding to the specific noise radar and to normal atmospheric conditions in the measuring area and the step of folding the speeds V ₁ ', V _2' V and ₃ 'and noisy in the ranges [i -V _n, V _n i], [-V _n2, V _n2 ] and [-V _n3 , V _n3 ] to obtain the velocities V ₁ ', V ₂ ' and V ₃ 'used later in the calculations.

10. Method according to claim 6, characterized in that the parameter q is equal to p + 1.