CN104280682A - Motor rotor fault diagnosis method based on field-orientated control - Google Patents

Abstract

The invention discloses a motor rotor fault diagnosis method based on field-orientated control. According to the motor rotor fault diagnosis method, calculated torque current of a controller is used for calculating the amplitudes of a direct current component and an alternate current component of the torque current, the number of broken bars of a rotor of a motor is further obtained, and therefore accurate detection and quantization of the broken bars of the rotor of the motor are realized. The motor rotor fault diagnosis method neither needs additional hardware nor increases the system cost; meanwhile, a diagnosis result is not sensitive to motor rotation speed changes or load torque changes, and the diagnosis process does not need parameters except the total number of guide bars of the rotor, so that the diagnosis result is high in robustness, and the motor rotor fault diagnosis method is quite suitable for a rotor-field-orientated control motor system.

Description

A kind of rotor method for diagnosing faults based on Field orientable control

Technical field

The invention belongs to For Diagnosing Faults of Electrical technical field, be specifically related to a kind of rotor method for diagnosing faults based on Field orientable control.

Background technology

Induction motor has simple, the cheap advantage of structure and is widely used in industrial and agricultural production.But due to the problem such as project organization and manufacturing process, motor, after long overload or Fraquent start, braking, may occur that rotor cage bar ruptures.After rotor bar fracture, electric current of its contiguous sliver can be made to strengthen, and stress increases, and broken bar fault will expand further, occur many disconnected bars, motor is exerted oneself reduction, also there will be rotor and sweep thorax and spoil stator, cause complete machine to be scrapped time serious.Therefore, rotor fault occur early stage, just fault detect out and quantize its fault severity level, with on-call maintenance accordingly, can avoid the generation of hang-up and serious accident, tool has very great significance.

When motor hangs over electric online operation, the method such as rotating speed, noise, electric current, electromagnetic torque, some or all of instantaneous power, magnetic flux by detecting motor realizes the Rotor Fault Diagnosis of motor, and its fault detect and quantization method are substantially perfect.

And along with the development of Power Electronic Technique, the frequency control of inverter and motor is widely used.Now, the fault diagnosis of motor also needs to consider the problem such as bandwidth of the change of electric moter voltage power frequency, the strong noise of electric current and voltage, the impact of closed loop and controller.Now, original method for diagnosing faults for mains supply situation can partly or completely lose efficacy.Need to study especially for inverter power supply and closed-loop control situation for this reason.To high performance Electric Machine Control occasion, Field orientable control (Field Oriented Control, FOC) be the most frequently used control method, it passes through coordinate transform, an actual asynchronous machine is transformed into virtual Equivalent DC motor, achieve the dynamic decoupling of magnetic field and torque, obtain excellent torque dynamic control performance.

For this application scenario, have and several authors proposed certain methods, as reference frame coordinate transformation method, Vienna observation procedure (Vienna Monitoring Method, VMM), utilize controller built-in variable (as rotor flux, exciting current etc.) and virtual current technical method.But, first three class methods only achieves qualitative analysis, and do not realize quantizing to rotor fault, and although virtual current technical method can realize fault quantification, but this method not only calculation of complex, and need to use the more parameter of electric machine, and some of them parameter such as rotor time constant is difficult to Obtaining Accurate, this has a certain impact to rotor fault diagnosis result.

Summary of the invention

For the above-mentioned technical matters existing for prior art, the invention provides a kind of rotor method for diagnosing faults based on Field orientable control, the method only needs the rotor bar number of torque current and the motor calculated in electric machine controller, just can realize the Rotor Fault Diagnosis of motor, and the fault severity level of accurate quantification motor.

Based on a rotor method for diagnosing faults for Field orientable control, comprise the steps:

(1) gather the threephase stator electric current of motor, converted exciting current and the torque current of d shaft current and q shaft current and the corresponding motor extracted wherein by dq;

(2) obtained the rotor flux of motor according to described exciting current by flux observation algorithm, and then according to the rotor flux of motor and torque current, calculate the slip frequency f of motor _s;

(3) low-pass filtering is carried out to described torque current, then with time window T, the DC component that moving average filter obtains torque current is carried out to filtered torque current,

(4) described torque current is deducted AC compounent that namely its DC component obtains torque current, and then according to slip frequency f _sgoerzel algorithm process is carried out to torque current AC compounent, obtains the amplitude of torque current AC compounent;

(5) according to the amplitude of torque current DC component and torque current AC compounent, judge whether rotor exists disconnected bar and disconnected bar radical.

Motor slip frequency f is calculated by flux observation algorithm in described step (2) _sspecific implementation process as follows:

First, the rotor flux ψ of motor is calculated according to following formula _r;

${\mathrm{\ψ}}_r=\frac{L_m{i}_{\mathrm{sd}}}{T_{r}s+1},T_r=\frac{L_r}{R_r}$

Wherein: L _mand i _sdbe respectively magnetizing inductance and the exciting current of motor, L _rand R _rbe respectively inductor rotor and the rotor resistance of motor, s is Laplace operator;

Then, the slip angular frequency ω of motor is calculated according to following formula _s;

${\mathrm{\ω}}_s=\frac{L_m{i}_{\mathrm{sq}}}{T_r{\mathrm{\ψ}}_r}$

Wherein: i _sqfor the torque current of motor;

Finally, according to f _s=ω _s/ 2 π determine the slip frequency f of motor _s.

Preferably, in described step (3), by following quadravalence butterworth filter transfer function H (z), low-pass filtering is carried out to torque current, to eliminate switching harmonics in torque current and other noises;

$H\left(z\right)=g_1\frac{1+{2z}^{-1}+z^{-2}}{1+a_1{z}^{-1}+a_2{z}^{-2}}\×g_2\frac{1+{2z}^{-1}+z^{-2}}{1+a_3{z}^{-1}+a_4{z}^{-2}}$

Wherein: g ₁, g ₂, a ₁, a ₂, a ₃and a ₄be filtering parameter, z is transform operator.

Preferably, before cycle average filter being carried out to filtered torque current in described step (3), first carry out down-sampled to this torque current; Torque current is after low-pass filtering, and the highest frequency of its useful signal is very low, and too high sample frequency is nonsensical, simultaneously the down-sampled computation burden that can reduce controller.

In described step (4), the method for Goerzel algorithm process carried out to torque current AC compounent as follows:

First, with the sampled point number in signal length N i.e. this duration of certain time length determination torque current AC compounent;

Then, signal value s (N-1) and the s (N-2) of torque current AC compounent N-2 and N-1 sampled point is tried to achieve according to following formula iteration:

s(i)＝x(i)+2cos(2πω)s(i-1)-s(i-2)?i＝0,1,…N-1

Wherein: s (i), s (i-1) and s (i-2) are respectively the signal value of torque current AC compounent i-th, the i-th-1 and the i-th-2 sampled point, x (i) is the current value of torque current AC compounent i-th sampled point, ω=4 π f _st _s, T _sfor the sampling period of torque current;

Finally, according to the amplitude of following formula calculating torque current alternating component:

${\hat{i}}_{\mathrm{sq}}=\sqrt{s^2(N-1)+s^2(N-2)-2\cos \left(2\mathrm{\π\ω}\right)s(N-1)s(N-2)}$

Wherein: for the amplitude of torque current AC compounent.

Judge whether rotor exists disconnected bar and disconnected bar radical according to following formula in described step (5):

$m=\frac{k}{1+2k}M,k=\frac{{\hat{i}}_{\mathrm{sq}}}{{\stackrel{\‾}{i}}_{\mathrm{sq}}}$

Wherein: m is the disconnected bar radical of rotor, and M is the sliver number of rotor, for the amplitude of torque current AC compounent, for the DC component of torque current.

The torque current of the present invention by having calculated in controller, calculates the amplitude of its DC component and AC compounent thereof, and then obtains the rotor broken bar number of motor, thus realizes rotor and to break the accurate detection of bar and quantification; The method does not need extra hardware, can not increase system cost; Simultaneously, diagnostic result changes insensitive to motor speed change and load torque, and diagnostic procedure does not need all the other parameters except the total sliver number of rotor, thus diagnostic result has higher robustness, is suitable for very much the electric system of orientation on rotor flux.

Accompanying drawing explanation

Fig. 1 is the system schematic of the direct orientation on rotor flux of induction motor and fault diagnosis thereof.

Fig. 2 is the current model schematic diagram calculating rotor magnetic linkage under MT coordinate system.

Fig. 3 is the schematic flow sheet of rotor fault diagnosis of the present invention.

Fig. 4 be rated speed, different loads torque and different disconnected bar radical time diagnostic result figure.

Fig. 5 be a disconnected bar, different loads torque and different given rotating speed time diagnostic result figure.

Embodiment

In order to more specifically describe the present invention, below in conjunction with the drawings and the specific embodiments, technical scheme of the present invention is described in detail.

Fig. 1 gives the schematic diagram that is applicable to the direct orientation on rotor flux of induction motor and fault diagnosis system 100 thereof.This system includes but not limited to DC bus 101, inverter module 102, current sensor 103, voltage sensor 104, induction motor 105, tachogenerator 106, control module 110 and Rotor Fault Diagnosis module 130.

At strong power part, the front end of DC bus sections 101 is generally three-phase alternating-current supply and uncontrollable rectifier circuit.After obtaining more satisfactory DC voltage by 101, powered for threephase asynchronous machine 105 by voltage source inverter module 102.

At motor control module 110, motor adopts the indirect orientation on rotor flux mode of magnetic linkage amplitude closed loop, and this part mainly comprises coordinate transform and flux observation 123, subtracter 111, subtracter 112, subtracter 115, subtracter 116, flux regulating device 113, speed adjusting device 114, regulating current device 117, coordinate transformation module 121 and pwm signal generation module 122.Flux regulating device 113, speed adjusting device 114 generally adopt PI to control with regulating current device 117.PWM generation module often adopts sinusoidal pulse width modulation (SPWM) or Voltage space vector PWM (SVPWM).

In this example, coordinate transform and flux observation module 123 adopt the current model on MT coordinate system.Its structured flowchart as shown in Figure 2.First by 3/2 module, the electric current of motor is transformed to α β coordinate system from three-phase static coordinate system, then two-phase rest frame is transformed in the middle of two-phase synchronous rotating frame, obtain the exciting current i of motor _sdwith torque current i _sq.According to the rotor flux amplitude of motor can be obtained, l _rand R _rbe respectively inductor rotor and the rotor resistance of motor; And the rotating magnetic field angular velocity omega of motor ₁=ω+ω _s, ω _sbe motor slip angular frequency and ω is the angular rate of rotor.To the rotating magnetic field angular velocity omega of motor ₁integration, can obtain the position of rotor magnetic linkage.

What deserves to be explained is, 3/2 module both can input three-phase current, also can input biphase current because three-phase current and always zero, the instantaneous value of third phase electric current can be obtained according to biphase current wherein.Except the current model (as shown in Figure 2) on MT coordinate system obtains except the rotor flux of motor, the current model on α β coordinate system or voltage model etc. can also be adopted to obtain the rotor flux of motor.

The model of the three-phase induction motor of rotor fault is:

$\left[\begin{array}{c}{u}_{\mathrm{sd}}\\ u_{\mathrm{sq}}\\ 0\\ 0\end{array}\right]=\left[\begin{array}{cccc}{R}_s+{{\mathrm{pL}}_s}^{\′}& -{\mathrm{\ω}}_1{L_s}^{\′}& p\frac{L_m}{L_r}& {-\mathrm{\ω}}_1\frac{L_m}{L_r}\\ {\mathrm{\ω}}_1{L_s}^{\′}& R_s+{{\mathrm{pL}}_s}^{\′}& {\mathrm{\ω}}_1\frac{L_m}{L_r}& p\frac{L_m}{L_r}\\ -\frac{L_m}{L_r}({\stackrel{\‾}{R}}_r+{\mathrm{\ΔR}}_r\cos \left({2\mathrm{\θ}}_{\mathrm{s\ω}}\right))& \frac{L_m}{L_r}{\mathrm{\ΔR}}_r\sin \left({2\mathrm{\θ}}_{\mathrm{s\ω}}\right)& p+\frac{{\stackrel{\‾}{R}}_r+{\mathrm{\ΔR}}_r\cos \left({2\mathrm{\θ}}_{\mathrm{s\ω}}\right)}{L_r}& -s{\mathrm{\ω}}_1-\frac{{\mathrm{\ΔR}}_r\sin \left(2{\mathrm{\θ}}_{\mathrm{s\ω}}\right)}{L_r}\\ \frac{L_m}{L_r}{\mathrm{\ΔR}}_r\sin \left({2\mathrm{\θ}}_{\mathrm{s\ω}}\right)& -\frac{L_m}{L_r}({\stackrel{\‾}{R}}_r-{\mathrm{\ΔR}}_r\cos \left({2\mathrm{\θ}}_{\mathrm{s\ω}}\right))& {\mathrm{s\ω}}_1-\frac{{\mathrm{\ΔR}}_r\sin \left({2\mathrm{\θ}}_{\mathrm{s\ω}}\right)}{L_r}& p+\frac{{\stackrel{\‾}{R}}_r-{\mathrm{\ΔR}}_r\cos \left({2\mathrm{\θ}}_{\mathrm{s\ω}}\right)}{L_r}\end{array}\right]\left[\begin{array}{c}{i}_{\mathrm{sd}}\\ i_{\mathrm{sq}}\\ {\mathrm{\ψ}}_{\mathrm{rd}}\\ {\mathrm{\ψ}}_{\mathrm{rq}}\end{array}\right]$

In formula, u _sdand i _sdfor the voltage and current of motor stator d axle, u _sqand i _sqfor the voltage and current of motor stator q axle, ψ _rdand ψ _rqfor the magnetic linkage of rotor d axle and q axle, R _sfor motor stator resistance, L _m, L _rfor magnetic linkage inductance and the inductor rotor of motor, and R _rfor the rotor resistance of motor, n is the rotor bar number that motor ruptures continuously, and N is total rotor bar number.Symbol p=d/dt, ω ₁for the rotational speed of motor in synchrony rotating coordinate system, θ _{s ω}for the slippage angle of motor, the i.e. difference of synchronous rotating frame and rotor electrical angle, s is the revolutional slip of motor.

When motor adopt Field orientable control and magnetic linkage amplitude closed loop time, when its magnetic linkage is positioned d axle, following formula is set up.

${\mathrm{\ψ}}_{\mathrm{rd}}={\mathrm{\ψ}}_r^{*},{\mathrm{\ψ}}_{\mathrm{rq}}=0$

And now d shaft current is also called exciting current, q shaft current is also called torque current.

During motor steady-state operation, in the d axle of motor stator and q shaft current, mainly contain DC component and 2 times of slip-frequency 2f _s=2sf ₁low frequency component, wherein 2f _slow frequency component is introduced due to rotor fault, f ₁for the fundamental frequency (line frequency) of motor stator electric current.Therefore can establish:

$i_{\mathrm{sd}}={\stackrel{\‾}{i}}_{\mathrm{sd}}+{\mathrm{\Δi}}_{\mathrm{sd}},i_{\mathrm{sq}}={\stackrel{\‾}{i}}_{\mathrm{sq}}+{\mathrm{\Δi}}_{\mathrm{sq}}$

Wherein, for motor energization current and torque current mean value and be direct current; △ i _sd, △ i _sqfor the wave component of motor energization current and torque current, its amplitude is less, and its frequency is 2f _s.

The fourth line of the motor model of these formulas substitution above, have:

$\frac{L_m}{L_r}{\mathrm{\ΔR}}_r\sin \left({2\mathrm{\θ}}_{\mathrm{s\ω}}\right)({\stackrel{\‾}{i}}_{\mathrm{sd}}+{\mathrm{\Δi}}_{\mathrm{sd}})-\frac{L_m}{L_r}[{\stackrel{\‾}{R}}_r-{\mathrm{\ΔR}}_r\cos \left({2\mathrm{\θ}}_{\mathrm{s\ω}}\right)]({\stackrel{\‾}{i}}_{\mathrm{sq}}+{\mathrm{\Δi}}_{\mathrm{sq}})+[{\mathrm{s\ω}}_1-\frac{{\mathrm{\ΔR}}_r\sin \left({2\mathrm{\θ}}_{\mathrm{s\ω}}\right)}{L_r}]{\mathrm{\ψ}}_r^{*}=0$

Under rotor failure condition, still set up, substitute into above formula, and ignore second order in a small amount, have:

$\frac{L_m}{L_r}[-{\stackrel{\‾}{R}}_r{\mathrm{\Δi}}_{\mathrm{sq}}+{\mathrm{\ΔR}}_r\cos \left({2\mathrm{\θ}}_{\mathrm{s\ω}}\right){\stackrel{\‾}{i}}_{\mathrm{sq}}]=\frac{L_m}{L_r}{\stackrel{\‾}{R}}_r{\stackrel{\‾}{i}}_{\mathrm{sq}}-{\mathrm{s\ω}}_1{L}_m{\stackrel{\‾}{i}}_{\mathrm{sd}}$

Ignore the change of motor slip ratio s, the left side of above formula equal sign is AC compounent, and the right is then DC quantity.According to the concept of small-signal model, both members is 0, therefore:

$-{\stackrel{\‾}{R}}_r{\mathrm{\Δi}}_{\mathrm{sq}}+{\mathrm{\ΔR}}_r\cos \left({2\mathrm{\θ}}_{\mathrm{s\ω}}\right){\stackrel{\‾}{i}}_{\mathrm{sq}}=0$ Namely $\frac{{\mathrm{\ΔR}}_r}{{\stackrel{\‾}{R}}_r}=\frac{{\mathrm{\Δ}\hat{i}}_{\mathrm{sq}}}{{\stackrel{\‾}{i}}_{\mathrm{sq}}}=k$

In formula, torque current wave component △ i _sqamplitude.

△ R _rwith substitute into, the disconnected number obtaining rotor is:

$n=\frac{k}{1+2k}N$

Because k is general less, so above formula is approximate be written as n ≈ kN.Therefore, by calculating torque current i _sqdC component and the amplitude of AC compounent just can in the hope of the disconnected number of rotor, thus the diagnosis achieved rotor fault and quantification.

Based on above-mentioned theory, the workflow of present embodiment Rotor Fault Diagnosis, its step as shown in Figure 3:

(1) first by the torque current i of motor _sqthrough a low-pass filter 301, to eliminate switching harmonics in electric current and other noises.When motor adopts direct orientation on rotor flux, in its controller, inherently there is torque current i _sq.Its sample frequency is identical with the frequency of electric machine controller electric current loop, is generally several thousand or hertz up to ten thousand, is 8000Hz in this example.Low-pass filter 301 is quadravalence Butterworth lowpass filters.Its cutoff frequency f _cfor 50Hz.The transport function of this low-pass filter is:

$H\left(z\right)=g_1\frac{1+{2z}^{-1}+z^{-2}}{1+a_1{z}^{-1}+a_2{z}^{-2}}\×g_2\frac{1+{2z}^{-1}+z^{-2}}{1+a_3{z}^{-1}+a_4{z}^{-2}}$

In formula, g ₁=3.7978 × 10 ^-4, g ₂=3.7199 × 10 ^-4, a ₁=-1.9689, a ₂=0.9704, a ₃=-1.9285, a ₄=0.9300.

After low-pass filter 301 processes, the signal i obtained _sq1in the 2f that only there is DC component in theory and introduce due to rotor fault _sfrequency component.

(2) to filtered current signal i _sq1carry out down-sampled 302 process, obtain signal i _sq2.The sample frequency of signal is reduced to 200Hz from 8000Hz, and be equivalent to every 40 points in original signal and extract a point, namely the down-sampled factor is 40.Adding down-sampled reason is mainly contain two reasons: one is because torque current i _sqafter low-pass filter 301, very low (generally lower than 10Hz), too high sample frequency is nonsensical for the highest frequency of its useful signal; It two is because the down-sampled computation burden that can reduce controller.

(3) to the signal i after down-sampled _sq2carry out moving average filter, thus obtain the mean value of torque current the cycle of calculating mean value and f _scontroller by motor obtains, and it is by slip angular frequency ω _sobtain divided by 2 π, concrete acquisition process is with reference to figure 2.

(4) the signal i after down-sampled _sq2deduct its mean value (frequency is 2f to obtain only containing AC compounent _s) Δ i _sq, its objective is and prevent larger mean component the 2f that frequency is lower is caused because of spectrum leakage _sfrequency component crest meter is not calculated accurately really.

(5) to Δ i _sqcarry out Goerzel algorithm process, thus extract the 2f in signal _sthe amplitude of frequency component the signal length of Goerzel algorithm process is N=600 point (i.e. the data of 3 seconds).Its computation process is as follows:

s(i)＝x(i)+2cos(2πω)s(i-1)-s(i-2)?i＝0,1,…N-1

Wherein: the current value that s (-1)=s (-2)=0, x (i) is torque current AC compounent i-th sampled point, ω=4 π f _st _s, T _sfor the sampling period of torque current; In this example,

By above formula after N=600 interative computation, s (N-1) and s (N-2) can be obtained, and then can 2f be obtained _sthe amplitude of frequency component, its calculation expression is:

${\mathrm{\Δ}\hat{i}}_{\mathrm{sq}}=\sqrt{s^2(N-1)+s^2(N-2)-2\cos \left(2\mathrm{\π\ω}\right)s(N-1)s(N-2)}$

(6) basis with can try to achieve

(7) according to total rotor bar number of k value and motor, the rotor broken bar radical of motor can be tried to achieve.Its express be for this formula also can be similar to and be written as n ≈ kN.

(8) according to the size of n, the whether disconnected bar of motor is judged, and disconnected bar radical; Corresponding remedial measures is taked according to disconnected bar radical.

Fig. 4 and Fig. 5 gives the simulation result obtained according to present embodiment.Wherein, Fig. 4 is the diagnostic result of motor in rated speed (980rpm), different load torque situations.Can see, the reality of its rotor broken bar number and rotor of calculating number of breaking is coincide, and electric motor load torque change affects smaller on diagnostic result.Fig. 5 has then considered the load torque of motor and given rotating speed to the impact of diagnostic result.The result of Fig. 5 proves that this diagnostic method impact that the change of motor given rotating speed proposes the present invention is very little, thus demonstrates the robustness of the inventive method.

Claims (6)

1., based on a rotor method for diagnosing faults for Field orientable control, comprise the steps:

(1) gather the threephase stator electric current of motor, converted exciting current and the torque current of d shaft current and q shaft current and the corresponding motor extracted wherein by dq;

(2) obtained the rotor flux of motor according to described exciting current by flux observation algorithm, and then according to the rotor flux of motor and torque current, calculate the slip frequency f of motor _s;

(3) low-pass filtering is carried out to described torque current, then with time window T, the DC component that moving average filter obtains torque current is carried out to filtered torque current,

(4) described torque current is deducted AC compounent that namely its DC component obtains torque current, and then according to slip frequency f _sgoerzel algorithm process is carried out to torque current AC compounent, obtains the amplitude of torque current AC compounent;

(5) according to the amplitude of torque current DC component and torque current AC compounent, judge whether rotor exists disconnected bar and disconnected bar radical.

2. rotor method for diagnosing faults according to claim 1, is characterized in that: calculate motor slip frequency f by flux observation algorithm in described step (2) _sspecific implementation process as follows:

First, the rotor flux ψ of motor is calculated according to following formula _r;

${\mathrm{\ψ}}_r=\frac{L_m{i}_{\mathrm{sd}}}{T_{r}s+1},T_r=\frac{L_r}{R_r}$

Wherein: L _mand i _sdbe respectively magnetizing inductance and the exciting current of motor, L _rand R _rbe respectively inductor rotor and the rotor resistance of motor, s is Laplace operator;

Then, the slip angular frequency ω of motor is calculated according to following formula _s;

${\mathrm{\ω}}_s=\frac{L_m{i}_{\mathrm{sq}}}{T_r{\mathrm{\ψ}}_r}$

Wherein: i _sqfor the torque current of motor;

Finally, according to f _s=ω _s/ 2 π determine the slip frequency f of motor _s.

3. rotor method for diagnosing faults according to claim 1, is characterized in that: carry out low-pass filtering by following quadravalence butterworth filter transfer function H (z) to torque current in described step (3):

$H\left(z\right)=g_1\frac{1+{2z}^{-1}+z^{-2}}{1+a_1{z}^{-1}+a_2{z}^{-2}}\×g_2\frac{1+{2z}^{-1}+z^{-2}}{1+a_3{z}^{-1}+a_4{z}^{-2}}$

Wherein: g ₁, g ₂, a ₁, a ₂, a ₃and a ₄be filtering parameter, z is transform operator.

4. rotor method for diagnosing faults according to claim 1, is characterized in that: before carrying out cycle average filter to filtered torque current in described step (3), first carry out down-sampled to this torque current.

5. rotor method for diagnosing faults according to claim 1, is characterized in that: carry out the method for Goerzel algorithm process to torque current AC compounent in described step (4) as follows:

First, with the sampled point number in signal length N i.e. this duration of certain time length determination torque current AC compounent;

Then, signal value s (N-1) and the s (N-2) of torque current AC compounent N-2 and N-1 sampled point is tried to achieve according to following formula iteration:

s(i)＝x(i)+2cos(2πω)s(i-1)-s(i-2)?i＝0,1,…N-1

Wherein: s (i), s (i-1) and s (i-2) are respectively the signal value of torque current AC compounent i-th, the i-th-1 and the i-th-2 sampled point, x (i) is the current value of torque current AC compounent i-th sampled point, ω=4 π f _st _s, T _sfor the sampling period of torque current;

Finally, according to the amplitude of following formula calculating torque current alternating component:

${\hat{i}}_{\mathrm{sq}}=\sqrt{s^2(N-1)+s^2(N-2)-2\cos \left(2\mathrm{\π\ω}\right)s(N-1)s(N-2)}$

Wherein: for the amplitude of torque current AC compounent.

6. rotor method for diagnosing faults according to claim 1, is characterized in that: judge whether rotor exists disconnected bar and disconnected bar radical according to following formula in described step (5):

$m=\frac{k}{1+2k}M,k=\frac{{\hat{i}}_{\mathrm{sq}}}{{\stackrel{\‾}{i}}_{\mathrm{sq}}}$

Wherein: m is the disconnected bar radical of rotor, and M is the sliver number of rotor, for the amplitude of torque current AC compounent, for the DC component of torque current.