CN104239614A - Method for simulating aerodynamic instability signal of compressor - Google Patents

Abstract

The invention provides a method for simulating aerodynamic instability signals of a compressor and belongs to the technical field of impeller mechanical surge prevention. The method for simulating the aerodynamic instability signals of the compressor comprises the following steps: 1) constructing a basic signal set {f<i>(t)}, which consists of four subsets, including (a) a subset{s<i>(t)} of a small amount of aerodynamic instability signals obtained through testing, CFD (Computational Fluid Dynamics) numerical simulation or analytical numerical simulation, (b) a subset {g<i>(t)}of damped sine signals, (c) a subset{h<i>(t)} of half-sine pulses, and (d) a subset{n<i>(t)} of noises; 2) selecting basic signals; 3) conducting waveform transformation and superposition by adopting operations such as multiple transformation, scaling transformation, translation transformation and time domain superposition. The method can be used for generating a great number of aerodynamic instability signals covering characteristics such as different frequencies, amplitudes and stall inception modes at an extremely small production cost, and can be used for assessment and evaluation of aerodynamic instability detection/control devices of compressors.

Description

The analogy method of pneumatic plant aerodynamic unstability signal

Technical field

The present invention relates to turbomachine preventing surge technology, be specially a kind of analogy method of pneumatic plant aerodynamic unstability signal.

Background technology

Pneumatic plant aerodynamic unstability mainly comprises stall and surge two kinds of phenomenons.Aerodynamic unstability not only reduces the performance of turbomachine, and brings strong vibration, may cause structural damage and the generation caused a serious accident.In the examination and assessment of aerodynamic unstability detection/control device, aerodynamic unstability signal has a very important role.Always expect to obtain sample as much as possible, in order to examine accuracy and the reliability of detection/control device.

Generally speaking, the approach obtaining pneumatic plant aerodynamic unstability signal has three kinds: (1) Compressor test is measured and obtained; (2) numerical simulation based on Fluid Mechanics Computation (CFD) obtains; (3) numerical evaluation based on pneumatic plant analytic model obtains.

Test be obtain pneumatic plant aerodynamic unstability characteristic the most directly, one of the most real means.During test, pneumatic plant is stabilized in a certain rotating speed, adopts and regulate blower outlet air throttle or increase the modes such as compressor inlet flow distortion, force pneumatic plant to enter aerodynamic unstability state.By being arranged on the high frequency sound sensor on pneumatic plant, the signal in pickup aerodynamic unstability process, is recorded by parallel high-speed data acquisition system and is stored.A kind of high energy consumption, excessive risk and macrocyclic pilot project are not only in the test of pneumatic plant aerodynamic unstability, and test parameters adjustment difficulty, the aerodynamic unstability signal characteristic of acquisition is comparatively single.

Based on the method for numerical simulation of CFD, first pneumatic plant gas circuit is carried out stress and strain model, then carry out nonsteady aerodynamics calculating by professional software on computers.The signal in aerodynamic unstability process can be obtained by virtual digital sensor.CFD is labor Field Characteristics, one of important means of research aerodynamic unstability mechanism.Its computational complexity is relevant to stress and strain model, model complexity etc.Usual calculated amount is huge, for a state of pneumatic plant, when needing the computing machine of a few hours to several weeks.By parameter such as adjustment compressor model, status condition etc., the aerodynamic unstability signal with different characteristic can be obtained.

Pneumatic plant aerodynamic unstability based on analytic model is analyzed, and has carried out a large amount of simplification to the flowing of gas and compression process, is expressed as comparatively simple non-linear partial difference equation.Conventional analytic model has Moore-Greitzer model etc.By solving analytic equation group, the Dynamic Signal of aerodynamic unstability can be obtained.By the change of Model Parameter, the aerodynamic unstability signal with different characteristic can be obtained.

Above-mentioned three kinds of methods are the important means obtaining true pneumatic plant aerodynamic unstability characteristic and signal.Because aerodynamic unstability feature under same model pneumatic plant different operating state is different; The pneumatic plant aerodynamic unstability feature of different model is different.In the development and examination of pneumatic plant aerodynamic unstability device, a large amount of aerodynamic unstability signals is usually needed to reflect the aerodynamic unstability feature of the pneumatic plant of different model under different conditions.Adopt said method to obtain the aerodynamic unstability signal of large sample, need the manpower of at substantial, financial resources and time, be not only and there is no need, and feature coverage rate is narrow, is difficult to the requirement meeting comprehensive assessment aerodynamic unstability pick-up unit.

Summary of the invention

The object of this invention is to provide a kind of analogy method of pneumatic plant aerodynamic unstability signal.The method can be avoided or reduce pneumatic plant and force and breathe heavily test or the numerical simulation calculation based on CFD or analytic method, the pneumatic plant aerodynamic unstability signal of simulation can cover the feature such as frequency, amplitude, stall precursor pattern widely, for the examination of pneumatic plant aerodynamic unstability detections/control device and assessment provide the sample of abundance.

Technical scheme of the present invention is:

The analogy method of described a kind of pneumatic plant aerodynamic unstability signal, is characterized in that: comprise the following steps:

Step 1: build baseband signal collection { f _i(t) }: baseband signal collection is made up of following 4 signal subsets: (1) is by test, CFD numerical simulation or a small amount of aerodynamic unstability signal { s resolving numerical simulation acquisition _i(t) }; (2) attenuated sinusoidal signal { g _i(t) }; (3) half-sine pulse { h _i(t) }; (4) noise { n _i(t) };

Attenuated sinusoidal signal { g _i(t) } be expressed as:

ζ in formula _ifor damping ratio, f _ifor frequency, for initial phase;

Half-sine pulse { h _i(t) } be expressed as:

T in formula _ifor the width of half-sine pulse;

Step 2: get the aerodynamic unstability signal of f (t) for generating, and set the characteristic parameter of f (t); The baseband signal needed for f (t) is generated according to following process choosing:

If 1 f (t) waveform and { s _i(t) } in a certain signal s ₁t () similar, only frequency or amplitude different time, then from { s _i(t) } in choose this signal as baseband signal;

If 2 { s _i(t) } in the waveform of no signal similar to f (t) time, choosing following signal is baseband signal:

(1) attenuated sinusoidal signal g ₁(t), wherein ζ ₁=0, f ₁the rotor frequency of=signal f (t),

(2) attenuated sinusoidal signal g ₂(t), wherein ζ ₂=0, f ₂the blade passing frequency of=signal f (t),

(3) attenuated sinusoidal signal g ₃(t), wherein ζ ₃span [-1 0), f ₃the stall frequencies of=signal f (t),

(4) attenuated sinusoidal signal g ₄(t), wherein ζ ₄span [-1 0), f ₄the surge frequency of=signal f (t),

(5) if the stall precursor section of signal f (t) is prominent sharp wave, then half-sine pulse { h is selected _i(t) }, wherein T _ispan [0.1 20] ms; If the stall precursor section of signal f (t) is modal waves, then select attenuated sinusoidal signal g ₅(t), wherein ζ ₅=0, f ₅0.1 ~ 0.9 times of the rotor frequency of=signal f (t),

(6) noise signal n ₁(t), amplitude is 1;

Step 3: to be converted by baseband signal waveform and superposition obtains signal f (t):

If baseband signal is s in 1 step 2 ₁t (), then adopt multiple transformation to change signal amplitude; Companding conversion changes frequency, obtains signal f (t):

f(t)＝As ₁(t/b)

Wherein A is number multiplying factor, and b is companding coefficient;

If baseband signal is not s in 2 steps 2 ₁(t), then carry out following waveform conversion:

If the stall precursor section of signal f (t) is prominent sharp wave, then

$\begin{array}{c}f\left(t\right)=[A_1{g}_1\left(t\right)+A_1{g}_2\left(t\right)][U\left(t\right)-U(t-T_{T3})]\\ +\underset{i=0}{\overset{N}{\mathrm{\&Sigma;}}}\left[A_h{h}_i(t-T_{T1}-{\mathrm{iT}}_h)\right]\&CenterDot;[U(t-T_{T1})-U(t-T_{T2})]\\ +A_3{g}_3(t-T_{T2})[U(t-T_{T2})-U(t-T_{T3})]\\ +A_4[g_4(t-T_{T3})-1]U(t-T_{T3})\\ +A_n{n}_1\left(t\right)\end{array}$

If the stall precursor section of signal f (t) is modal waves, then

$\begin{array}{c}f\left(t\right)=[A_1{g}_1\left(t\right)+A_1{g}_2\left(t\right)][U\left(t\right)-U(t-T_{T3})]\\ +A_5{g}_5(t-T_{T1})[U(t-T_{T1})-U(t-T_{T2})]\\ +A_3{g}_3(t-T_{T2})[U(t-T_{T2})-U(t-T_{T3})]\\ +A_4[g_4(t-T_{T3})-1]U(t-T_{T3})\\ +A_n{n}_1\left(t\right)\end{array}$

Wherein T _hfor the cycle that prominent sharp wave occurs, N is the number occurring prominent sharp wave, T _t1for stall precursor section start time, T _t2for stall section start time, T _t3for surge section start time, A ₄for the signal amplitude of surge section, A ₃for the signal amplitude of stall section, A ₁for 1/2, A of the signal amplitude of stable section _hfor the signal amplitude of prominent sharp wave, A _nfor the amplitude of noise signal, A ₅for the signal amplitude of modal waves; U (t) is step function, and function expression is:

$U\left(t\right)=\left\{\begin{array}{cc}0& t<0\\ 1& t\&GreaterEqual;0\end{array}\right.;$

Step 4: repeat the aerodynamic unstability simulating signal that step 2 ~ step 3 generates further feature parameter.

Beneficial effect

The analogy method of pneumatic plant aerodynamic unstability signal of the present invention, solve and how to pass through to cover the problem of the aerodynamic unstability signal of broad characteristic range with approach acquisition easily fast, the examination and the assessment that can be pneumatic plant aerodynamic unstability detection/control device provide abundant test samples.There is following characteristics:

1. method described in adopts test, CFD calculates or the pneumatic plant aerodynamic unstability signal of analytical Calculation acquisition, and attenuated sinusoidal signal, half-sine pulse and noise form baseband signal set.

2. method described in chooses baseband signal in baseband signal set, adopts companding, number takes advantage of and the fundamental operation such as superposition or conversion generate aerodynamic unstability simulating signal.

3. the method described in comparatively pneumatic plant aerodynamic testing, CFD calculates and analytical Calculation, have method simple, easy to implement, great amount of samples can be produced and cover the wide feature of aerodynamic unstability feature.

Accompanying drawing explanation

Fig. 1. compressor stall signal (original signal);

Fig. 2. compressor stall signal (time domain 2 times stretching, extension);

Fig. 3. the pneumatic plant aerodynamic unstability signal of simulation.

Embodiment

Now in conjunction with two embodiments and accompanying drawing, the present invention is described further.

Aerodynamic unstability signal is divided into stable section, stall precursor section, stall section and surge section in time history, and its characteristic parameter mainly comprises rotor frequency (20 ~ 850Hz), blade passing frequency (100 ~ 26000Hz), stall frequencies f _s(20 ~ 300Hz), surge frequency f _g(0.1 ~ 50Hz), stall precursor stage mode (prominent sharp wave or modal waves), the amplitude (0.05 ~ 0.5 of stable section amplitude) of prominent sharp wave and time span (0.1 ~ 20ms), the amplitude (0.05 ~ 0.2 of stable section amplitude) of modal waves and frequency (0.1 ~ 0.9 of rotor frequency), stable section signal amplitude (0.01 ~ 0.9 of surge section amplitude), stall segment signal amplitude (0.01 ~ 0.9 of surge section amplitude), surge segment signal amplitude (being normalized to 1) etc.

Embodiment 1:

In the present embodiment, the analogy method of pneumatic plant aerodynamic unstability signal comprises the following steps:

Step 1: build baseband signal collection { f _i(t) }: baseband signal collection is made up of following 4 signal subsets: (1) is by test, CFD numerical simulation or a small amount of aerodynamic unstability signal { s resolving numerical simulation acquisition _i(t) }; (2) attenuated sinusoidal signal { g _i(t) }; (3) half-sine pulse { h _i(t) }; (4) noise { n _i(t) }; Wherein, { s _i(t) } in have test obtain compressor stall signal as Fig. 1.

Attenuated sinusoidal signal { g _i(t) } be expressed as:

ζ in formula _ifor damping ratio, typical value scope-1 ~ 1, f _ifor frequency, typical value scope 0 ~ 300Hz, for initial phase, span-π ~ π.

Half-sine pulse { h _i(t) } be expressed as:

T in formula _ifor the width of half-sine pulse.

Step 2: get the aerodynamic unstability signal of f (t) for generating, and set the characteristic parameter of f (t), as stall frequencies, surge frequency, surge segment signal amplitude etc.; The baseband signal needed for f (t) is generated according to following process choosing:

If 1 f (t) waveform and { s _i(t) } in a certain signal s ₁t () similar, only frequency (with stall frequencies or the definition of surge frequency) or amplitude is different time, then from { s _i(t) } in choose this signal as baseband signal;

If 2 { s _i(t) } in the waveform of no signal similar to f (t) time, choosing following signal is baseband signal:

(1) attenuated sinusoidal signal g ₁(t), wherein ζ ₁=0, f ₁the rotor frequency of=signal f (t),

(2) attenuated sinusoidal signal g ₂(t), wherein ζ ₂=0, f ₂the blade passing frequency of=signal f (t),

(3) attenuated sinusoidal signal g ₃(t), wherein ζ ₃span [-1 0), f ₃the stall frequencies of=signal f (t),

(4) attenuated sinusoidal signal g ₄(t), wherein ζ ₄span [-1 0), f ₄the surge frequency of=signal f (t),

(5) if the stall precursor section of signal f (t) is prominent sharp wave, then half-sine pulse { h is selected _i(t) }, wherein T _ispan [0.1 20] ms; If the stall precursor section of signal f (t) is modal waves, then select attenuated sinusoidal signal g ₅(t), wherein ζ ₅=0, f ₅0.1 ~ 0.9 times of the rotor frequency of=signal f (t),

(6) noise signal n ₁(t), amplitude is 1;

In the present embodiment, the default feature of f (t) is: stall frequencies is 1/2 of signal shown in Fig. 1, so select Fig. 1 signal to be s ₁(t).

Step 3: to be converted by baseband signal waveform and superposition obtains signal f (t):

If baseband signal is s in 1 step 2 ₁t (), then adopt multiple transformation to change signal amplitude; Companding conversion changes frequency, obtains signal f (t):

f(t)＝As ₁(t/b)

Wherein A is number multiplying factor, and b is companding coefficient;

If baseband signal is not s in 2 steps 2 ₁(t), then carry out following waveform conversion:

If the stall precursor section of signal f (t) is prominent sharp wave, then

$\begin{array}{c}f\left(t\right)=[A_1{g}_1\left(t\right)+A_1{g}_2\left(t\right)][U\left(t\right)-U(t-T_{T3})]\\ +\underset{i=0}{\overset{N}{\mathrm{\&Sigma;}}}\left[A_h{h}_i(t-T_{T1}-{\mathrm{iT}}_h)\right]\&CenterDot;[U(t-T_{T1})-U(t-T_{T2})]\\ +A_3{g}_3(t-T_{T2})[U(t-T_{T2})-U(t-T_{T3})]\\ +A_4[g_4(t-T_{T3})-1]U(t-T_{T3})\\ +A_n{n}_1\left(t\right)\end{array}$

If the stall precursor section of signal f (t) is modal waves, then

$\begin{array}{c}f\left(t\right)=[A_1{g}_1\left(t\right)+A_1{g}_2\left(t\right)][U\left(t\right)-U(t-T_{T3})]\\ +A_5{g}_5(t-T_{T1})[U(t-T_{T1})-U(t-T_{T2})]\\ +A_3{g}_3(t-T_{T2})[U(t-T_{T2})-U(t-T_{T3})]\\ +A_4[g_4(t-T_{T3})-1]U(t-T_{T3})\\ +A_n{n}_1\left(t\right)\end{array}$

Wherein T _hfor the cycle that prominent sharp wave occurs, 1.1 ~ 10, N of span rotor cycle (inverse of rotor frequency) is the number occurring prominent sharp wave, span 1 ~ 20, T _t1for stall precursor section start time, span 3 ~ 30s, T _t2for stall section start time, span T _t1+ NT _h~ T _t1+ (N+1) T _h, T _t3for surge section start time, span T _t2+ M/f ₃, M=1 ~ 300, A ₄for the signal amplitude of surge section, A ₃for the signal amplitude of stall section, span A ₄0.01 ~ 0.9, A ₁for 1/2, span A of the signal amplitude of stable section ₄0.01 ~ 0.9 and be not more than A ₃/ 2, A _hfor the signal amplitude of prominent sharp wave, A ₁0.025 ~ 0.25, A _nfor the amplitude of noise signal, A ₁0.005 ~ 0.1, A ₅for the signal amplitude of modal waves, A ₁0.025 ~ 0.1; U (t) is step function, and function expression is:

$U\left(t\right)=\left\{\begin{array}{cc}0& t<0\\ 1& t\&GreaterEqual;0\end{array}\right.,$ So, $U(t-t_0)=\left\{\begin{array}{cc}0& t<t_0\\ 1& t\&GreaterEqual;t_0\end{array}\right..$

In the present embodiment, converted by companding, f (t)=s ₁(2t), obtain waveform shown in Fig. 2, its stall frequencies is 1/2 of Fig. 1 waveform.

Step 4: repeat the aerodynamic unstability simulating signal that step 2 ~ step 3 generates the characteristic parameter such as other stall frequencies, surge frequency.

Embodiment 2

In the present embodiment, the analogy method of pneumatic plant aerodynamic unstability signal comprises the following steps:

Step 1: build baseband signal collection { f _i(t) }: baseband signal collection is made up of following 4 signal subsets: (1) is by test, CFD numerical simulation or a small amount of aerodynamic unstability signal { s resolving numerical simulation acquisition _i(t) }; (2) attenuated sinusoidal signal { g _i(t) }; (3) half-sine pulse { h _i(t) }; (4) noise { n _i(t) }.

Attenuated sinusoidal signal { g _i(t) } be expressed as:

ζ in formula _ifor damping ratio, typical value scope-1 ~ 1, f _ifor frequency, typical value scope 0 ~ 300Hz, for initial phase, span-π ~ π.

Half-sine pulse { h _i(t) } be expressed as:

T in formula _ifor the width of half-sine pulse.

Step 2: get the aerodynamic unstability signal of f (t) for generating, and set the characteristic parameter of f (t), as stall frequencies, surge frequency, surge segment signal amplitude etc.; The baseband signal needed for f (t) is generated according to following process choosing:

If 1 f (t) waveform and { s _i(t) } in a certain signal s ₁t () similar, only frequency (with stall frequencies or the definition of surge frequency) or amplitude is different time, then from { s _i(t) } in choose this signal as baseband signal;

If 2 { s _i(t) } in the waveform of no signal similar to f (t) time, choosing following signal is baseband signal:

(1) attenuated sinusoidal signal g ₁(t), wherein ζ ₁=0, f ₁the rotor frequency of=signal f (t),

(2) attenuated sinusoidal signal g ₂(t), wherein ζ ₂=0, f ₂the blade passing frequency of=signal f (t),

(3) attenuated sinusoidal signal g ₃(t), wherein ζ ₃span [-1 0), f ₃the stall frequencies of=signal f (t),

(4) attenuated sinusoidal signal g ₄(t), wherein ζ ₄span [-1 0), f ₄the surge frequency of=signal f (t),

(5) if the stall precursor section of signal f (t) is prominent sharp wave, then half-sine pulse { h is selected _i(t) }, wherein T _ispan [0.1 20] ms; If the stall precursor section of signal f (t) is modal waves, then select attenuated sinusoidal signal g ₅(t), wherein ζ ₅=0, f ₅0.1 ~ 0.9 times of the rotor frequency of=signal f (t),

(6) noise signal n ₁(t), amplitude is 1;

In the present embodiment, the default feature of f (t) is: rotor frequency 65Hz, blade passing frequency 975Hz, and stall precursor is prominent sharp wave pattern, prominent sharp wave pulsewidth 5ms, prominent sharp wave number 6, stall frequencies 54Hz, surge frequency 13Hz, stable section amplitude 0.2, surge section amplitude 1.Select attenuated sinusoidal signal g ₁(t), its ζ ₁=0, f ₁=65Hz, select attenuated sinusoidal signal g ₂(t), its ζ ₂=0, f ₂=975Hz, select attenuated sinusoidal signal g ₃(t), its ζ ₃=-0.002, f ₃=54Hz, select attenuated sinusoidal signal g ₄(t), its ζ ₄=-0.0015, f ₄=13Hz, select h ₁(t), its T ₁=5ms; Select noise signal n ₁t (), amplitude is 1.

Step 3: to be converted by baseband signal waveform and superposition obtains signal f (t):

If baseband signal is s in 1 step 2 ₁t (), then adopt multiple transformation to change signal amplitude; Companding conversion changes frequency, obtains signal f (t):

f(t)＝As ₁(t/b)

Wherein A is number multiplying factor, and b is companding coefficient;

If baseband signal is not s in 2 steps 2 ₁(t), then carry out following waveform conversion:

If the stall precursor section of signal f (t) is prominent sharp wave, then

$\begin{array}{c}f\left(t\right)=[A_1{g}_1\left(t\right)+A_1{g}_2\left(t\right)][U\left(t\right)-U(t-T_{T3})]\\ +\underset{i=0}{\overset{N}{\mathrm{\&Sigma;}}}\left[A_h{h}_i(t-T_{T1}-{\mathrm{iT}}_h)\right]\&CenterDot;[U(t-T_{T1})-U(t-T_{T2})]\\ +A_3{g}_3(t-T_{T2})[U(t-T_{T2})-U(t-T_{T3})]\\ +A_4[g_4(t-T_{T3})-1]U(t-T_{T3})\\ +A_n{n}_1\left(t\right)\end{array}$

If the stall precursor section of signal f (t) is modal waves, then

$\begin{array}{c}f\left(t\right)=[A_1{g}_1\left(t\right)+A_1{g}_2\left(t\right)][U\left(t\right)-U(t-T_{T3})]\\ +A_5{g}_5(t-T_{T1})[U(t-T_{T1})-U(t-T_{T2})]\\ +A_3{g}_3(t-T_{T2})[U(t-T_{T2})-U(t-T_{T3})]\\ +A_4[g_4(t-T_{T3})-1]U(t-T_{T3})\\ +A_n{n}_1\left(t\right)\end{array}$

Wherein T _hfor the cycle that prominent sharp wave occurs, N is the number occurring prominent sharp wave, T _t1for stall precursor section start time, T _t2for stall section start time, T _t3for surge section start time, A ₄for the signal amplitude of surge section, A ₃for the signal amplitude of stall section, A ₁for 1/2, A of the signal amplitude of stable section _hfor the signal amplitude of prominent sharp wave, A _nfor the amplitude of noise signal, A ₅for the signal amplitude of modal waves; U (t) is step function, and function expression is:

$U\left(t\right)=\left\{\begin{array}{cc}0& t<0\\ 1& t\&GreaterEqual;0\end{array}\right.,$ So, $U(t-t_0)=\left\{\begin{array}{cc}0& t<t_0\\ 1& t\&GreaterEqual;t_0\end{array}\right..$

In the present embodiment, following formula is adopted to carry out waveform conversion and superposition:

$\begin{array}{c}f\left(t\right)=[A_1{g}_1\left(t\right)+A_1{g}_2\left(t\right)][U\left(t\right)-U(t-T_{T3})]\\ +\underset{i=0}{\overset{N}{\mathrm{\&Sigma;}}}\left[A_h{h}_i(t-T_{T1}-{\mathrm{iT}}_h)\right]\&CenterDot;[U(t-T_{T1})-U(t-T_{T2})]\\ +A_3{g}_3(t-T_{T2})[U(t-T_{T2})-U(t-T_{T3})]\\ +A_4[g_4(t-T_{T3})-1]U(t-T_{T3})\\ +A_n{n}_1\left(t\right)\end{array}$

Wherein T _hfor the cycle that prominent sharp wave occurs, value 18.5ms, N are the number occurring prominent sharp wave, value 6, T _t1for stall precursor section start time, value 6s, T _t2for stall section start time, value 6.1116s, T _t3for surge section start time, value 6.4821s, A ₄for the signal amplitude of surge section, value 1, A ₃for the signal amplitude of stall section, value 0.2, A ₁for 1/2 of the signal amplitude of stable section, value 0.1, A _hfor the signal amplitude of prominent sharp wave, value 0.01, A _nfor the amplitude of noise signal, value 0.01.

Step 4: repeat the aerodynamic unstability simulating signal that step 2 ~ step 3 generates the characteristic parameter such as other stall frequencies, surge frequency.

Claims (1)

1. an analogy method for pneumatic plant aerodynamic unstability signal, is characterized in that: comprise the following steps:

Step 1: build baseband signal collection { f _i(t) }: baseband signal collection is made up of following 4 signal subsets: (1) is by test, CFD numerical simulation or a small amount of aerodynamic unstability signal { s resolving numerical simulation acquisition _i(t) }; (2) attenuated sinusoidal signal { g _i(t) }; (3) half-sine pulse { h _i(t) }; (4) noise { n _i(t) };

Attenuated sinusoidal signal { g _i(t) } be expressed as:

ζ in formula _ifor damping ratio, f _ifor frequency, for initial phase;

Half-sine pulse { h _i(t) } be expressed as:

T in formula _ifor the width of half-sine pulse;

Step 2: get the aerodynamic unstability signal of f (t) for generating, and set the characteristic parameter of f (t); The baseband signal needed for f (t) is generated according to following process choosing:

1) if f (t) waveform and { s _i(t) } in a certain signal s ₁t () similar, only frequency or amplitude different time, then from { s _i(t) } in choose this signal as baseband signal;

2) if { s _i(t) } in the waveform of no signal similar to f (t) time, choosing following signal is baseband signal:

(1) attenuated sinusoidal signal g ₁(t), wherein ζ ₁=0, f ₁the rotor frequency of=signal f (t),

(2) attenuated sinusoidal signal g ₂(t), wherein ζ ₂=0, f ₂the blade passing frequency of=signal f (t),

(3) attenuated sinusoidal signal g ₃(t), wherein ζ ₃span [-1 0), f ₃the stall frequencies of=signal f (t),

(4) attenuated sinusoidal signal g ₄(t), wherein ζ ₄span [-1 0), f ₄the surge frequency of=signal f (t),

(5) if the stall precursor section of signal f (t) is prominent sharp wave, then half-sine pulse { h is selected _i(t) }, wherein T _ispan [0.1 20] ms; If the stall precursor section of signal f (t) is modal waves, then select attenuated sinusoidal signal g ₅(t), wherein ζ ₅=0, f ₅0.1 ~ 0.9 times of the rotor frequency of=signal f (t),

(6) noise signal n ₁(t), amplitude is 1;

Step 3: to be converted by baseband signal waveform and superposition obtains signal f (t):

1) if baseband signal is s in step 2 ₁t (), then adopt multiple transformation to change signal amplitude; Companding conversion changes frequency, obtains signal f (t):

f(t)＝As ₁(t/b)

Wherein A is number multiplying factor, and b is companding coefficient;

2) if baseband signal is not s in step 2 ₁(t), then carry out following waveform conversion:

If the stall precursor section of signal f (t) is prominent sharp wave, then

If the stall precursor section of signal f (t) is modal waves, then

Wherein T _hfor the cycle that prominent sharp wave occurs, N is the number occurring prominent sharp wave, T _t1for stall precursor section start time, T _t2for stall section start time, T _t3for surge section start time, A ₄for the signal amplitude of surge section, A ₃for the signal amplitude of stall section, A ₁for 1/2, A of the signal amplitude of stable section _hfor the signal amplitude of prominent sharp wave, A _nfor the amplitude of noise signal, A ₅for the signal amplitude of modal waves; U (t) is step function, and function expression is:

Step 4: repeat the aerodynamic unstability simulating signal that step 2 ~ step 3 generates further feature parameter.