CN1601294A - Detection method of d.c. permanent magnet dynamo parameter and performance non-loading - Google Patents

Abstract

The invention discloses a test method of DC permanent-magnet machine-DC permanent-magnet machine parameter and perforance loadless test method. Said test method includes the following steps: starting; computer can output signal to make power supply apply a started trigger voltage to DC machine, in the starting time its current sensor can be used for measuring and taking current value is of N sampling points and creating current curve Ps of DC machine; comparing the differences between the current value im if electric machine model curve is every sampling point and measured current value is, and resolving mean square value e of relative value; modifying multidimensional parameter vector relue for deciding Pm, making Pm approach to Ps; the multidimensional parameter vector value correspondent to Pm when e is approached to zero is the resolved actual parameter vector of the DC electric machine; then calculating other electric machine parameter and performance of DC electric machine.

Description

The non-loaded method of testing of direct current permanent magnet motor parameter and performance

Technical field:

The present invention relates to direct current permanent magnet motor and only testing the method for testing that obtains the parameter of electric machine and performance under unloaded sound attitude armature voltage, the current conditions.

Background technology:

The conventionally test of direct current generator parameter and characteristic generally is under rated voltage, output axle head at motor adds a loading moment, measure rotating speed, torque, voltage, the electric current of motor under each loading moment, thereby calculate input, output power, electric efficiency, electromotive force constant, moment constant and other parameter of motor, and obtain the machine operation characteristic on this basis.This method need add steady load to motor, also need measure torque, rotating speed, voltage, electric current equivalent.Finish a lot of hardware devices of these needs of work, and the test macro that constitutes is huger, complicated operation is inappropriate under some occasion, as the on-line testing of motor.

Summary of the invention:

, provide a kind of and need not, and the method for simple direct current permanent magnet motor parameter of equipment needed thereby and the non-loaded test of performance to the motor loading in order to overcome the defective that existing direct current generator parameter test method must can detect for the motor loading.The system that realizes this method of testing by: current sensor, voltage sensor, computing machine, power supply and data acquisition system (DAS) are formed, the present invention is to the test of the parameter and the performance of motor, basically be to be finished by software, the program that operates in the computing machine is as follows to the step of electromechanical testing: start 101; The computing machine output signal makes power supply apply the trigger voltage U 102 of a startup to direct current generator; By current sensor in the start-up time of direct current generator, measure N sampled point be K=1,2,3 ..., N current value i _s(K), obtain the current curve P of direct current generator _s103; According to trigger voltage U find be stored in the computing machine with the corresponding motor model curve of trigger voltage U P _m, sampled point compares motor model curve P one by one _mCurrent value i at each sampled point _m(K) with detected current value i _s(K) difference between, and obtain the mean square value e of relative value, motor model curve P _mThe multidimensional parameter vector P of unique correspondence be known and be stored in the computing machine 104; Modification is determining motor model curve P _mThe value of multidimensional parameter vector P, make motor model curve P _mLevel off to current curve P _s105; When e levels off to motor model curve P 0 time _mThe value of pairing multidimensional parameter vector P promptly is the actual parameter vector 106 of the direct current generator of asking; Other parameter of electric machine and performance 107 according to actual parameter vector calculation direct current generator; Finish 108; So just finished test to the parameter and the operating characteristic of motor.The present invention only needs to gather no-load voltage, the electric current of motor to the test of the parameter and the performance of motor, just can obtain the main sound attitude parameter of motor and the characteristic working curve of motor.Required hardware device is few, and does not need to apply load to tested direct current generator, and this method of testing is compared with the conventionally test method has the advantage that the test duration is short, efficient is high, cost is low, and can on-line testing.The above-mentioned method that measures the parameter of electric machine is different fully with traditional method that measures the parameter of electric machine.The method that tradition measures the parameter of electric machine is to make every effort to the influence of each parameter is separated, and removes to measure the parameter of electric machine one by one.This method workload is big, speed is slow, needs a lot of accurate equipment and instrument.The present invention is all parameters of once measuring motor on the whole.Be the process of a search optimizing, the i of transient process (t) has comprised the information of all parameters and starting condition.Comparison operation by a large amount of makes P _mProgressively approach P _s, i _m(t) progressively approach i _s(t), at last with P _sFind out.Described direct current permanent magnet motor parameter and the non-loaded method of testing of performance, its dynamic and static parameter that can test the direct current generator that obtains comprises: armature resistance, armature inductance, constant of the machine, moment of inertia, moment constant, electromotive force constant, dry friction torque, sticking friction factor, electromechanical time constant, electrical time constant; It can be tested the motor performance index that obtains and comprise: efficient, peak power, mechanical property slope when electric current, maximum output when torque when rotating speed, maximum output, maximum output when output power, maximal efficiency, maximum output when electric current, maximal efficiency when torque when rotating speed, maximal efficiency, maximal efficiency when no-load speed, no-load current, locked rotor current, locked-rotor torque, maximal efficiency; The motor characteristic working curve that it can measure comprises: characteristics of output power, input current characteristic, rotary speed property, efficiency characteristic.

Description of drawings:

Fig. 1 is the process flow diagram of method of testing of the present invention, and Fig. 2 is a parameter of electric machine discrimination method schematic diagram, and Fig. 3 is a parameter of electric machine discrimination method curve synoptic diagram.

Embodiment:

Specify embodiments of the present invention below in conjunction with Fig. 1, Fig. 2 and Fig. 3.The system that realizes this method of testing is by current sensor, voltage sensor, computing machine, power supply and data acquisition system (DAS), the present invention is to the test of the parameter and the performance of motor, basically be to be finished by software, the program that operates in the computing machine is as follows to the step of electromechanical testing: start 101; The computing machine output signal makes power supply apply the trigger voltage U 102 of a startup to direct current generator; By current sensor in the start-up time of direct current generator, measure N sampled point be K=1,2,3 ..., N current value i _s(K), set up the current curve P of direct current generator _s103; According to trigger voltage U find be stored in the computing machine with the corresponding motor model curve of trigger voltage U P _m, sampled point compares motor model curve P one by one _mCurrent value i at each sampled point _m(K) with detected current value i _s(K) difference between, and obtain the mean square value e of relative value, motor model curve P _mThe multidimensional parameter vector P of unique correspondence be known and be stored in the computing machine 104; Modification is determining motor model curve P _mThe value of multidimensional parameter vector P, make motor model curve P _mLevel off to current curve P _s105; When e levels off to motor model curve P 0 time _mThe value of pairing multidimensional parameter vector P promptly is the actual parameter vector 106 of the direct current generator of asking; Other parameter of electric machine and performance 107 according to actual parameter vector calculation direct current generator; Finish 108; Storing the database that comprises a large amount of motor model curves in the computing machine, every pairing multidimensional parameter vector of motor model curve all is known definite.Step 104 to step 106 is processes of the parameter identification of the present patent application, and following mask is stopped the origin that the interpretation parameters identification is handled:

(1) mathematical model of direct current permanent magnet motor: the mathematical model of not having motor in coupling direct current permanent magnet motor parameter and the Performance Test System is promptly to ignore armature reaction with linear as the basis of analyzing direct current permanent magnet motor, and air-gap flux does not change with load.Under light condition, the mathematical model of direct current permanent magnet motor can be described with following two equations:

$u=R_{a}i+L_a\frac{\mathrm{di}}{\mathrm{dt}}+\mathrm{C\Ω}+2\mathrm{\ΔU}--\left(1\right)$

$\mathrm{Ci}=J\frac{\mathrm{d\Ω}}{\mathrm{dt}}+T_f+C_f\mathrm{\Ω}--\left(2\right)$

In the formula

U-is added in the voltage at armature two ends, is step voltage U (V) during test;

I-armature supply (A);

Ω-motor speed (1/s), i.e. mechanical angle speed;

R _a-armature resistance (Ω), armature winding two ends resistance does not comprise the brush pressure drop affects;

L _a-armature winding inductance (H);

C-constant of the machine (Nm/A or VS);

Ignore the influence of armature reaction, can think that C is a constant in whole test specification,

Motor electromotive force e=C Ω then, motor electromagnetic torque T=Ci;

J-rotor moment of inertia (Kg.m ²);

T _f-motor friction torque (Nm), the moment of resistance;

C _f-sticking friction factor (Nm.S)

T _f+ C _fΩ comprises:

Bearing frictional torque,

Brush friction moment,

Wind friction (comprising fan) moment,

The iron loss moment of resistance,

Other added losses moment;

Regard the various moments of resistance linear function of rotating speed as, the moment of resistance is decomposed into two, the constant moment of resistance of the first does not change with rotating speed, and one is the adjustable resistance square, linear with rotating speed;

2 Δ U-brush contact drops (V)

In the starting under no load process, think that the brush pressure drop is a constant, do not change with electric current.

(2) foundation of parameter identification target equation: the target equation of parameter identification is the effective way of contact known quantity and unknown quantity, be wait to ask parameter and can obtainable parameter between relationship expression.According to the direct current permanent magnet motor mathematical model of having set up, at the required parameter of asking for, the mathematical model of motor is handled by mathematical method, can obtain carrying out the target side formula that identification is handled.In this test macro, detectable value is an output current, and the relation of the parameter of electric machine of required identification and output current is very close.Hence one can see that, and the target equation should be able to reflect the relation between the two.

Have by equation (1), (2)

$\frac{{\mathrm{JL}}_a}{C^2+C_f{R}_a}\frac{d^{2}i}{{\mathrm{dt}}^2}+\frac{J{R}_a+C_f{L}_a}{C^2+C_f{R}_a}\frac{\mathrm{di}}{\mathrm{dt}}+i=\frac{C_f(U-2\mathrm{\ΔU})+{\mathrm{CT}}_f}{C^2+C_f{R}_a}--\left(3\right)$

Following formula can be write a Chinese character in simplified form

$a_2\frac{d^{2}i}{{\mathrm{dt}}^2}+a_1\frac{\mathrm{di}}{\mathrm{dt}}+i=a--\left(4\right)$

Wherein

$a_2=\frac{{\mathrm{JL}}_a}{C^2+C_f{R}_a}--\left(5\right)$

$a_1=\frac{{\mathrm{JR}}_a+C_f{L}_a}{C^2+C_f{R}_a}--\left(6\right)$

$a=\frac{C_f(U-2\mathrm{\ΔU})+{\mathrm{CT}}_f}{C^2+C_f{R}_a}--\left(7\right)$

Notice, when zero load and t=∞, i=i ₀, i ₀Be no-load current, i.e. a=i ₀

Equation (4) is a second order differential equation, and it is carried out Laplace transformation can get

$a_2{p}^{2}i\left(p\right)+a_1\mathrm{pi}\left(p\right)+i\left(p\right)=\frac{i_0}{p}--\left(8\right)$

Thus, can solve

$i=i_0+A_1{e}^{p_{1}t}+A_2{e}^{p_{2}t}--\left(9\right)$

Can release equation (9) by the mathematical model of direct current permanent magnet motor.Analysis equation (9) as can be known, electric current is the function of several parameters of electric machine and time.When each parameter was determined in the model, then i was the function of time t.Each parameter will obtain different responses not simultaneously in model, and promptly the waveform of current i is also different, and visible i also is the function of each parameter.It is to be target component with detectable output current, comprises the equation of several unknown parameters.Unknown parameter is wherein all got in touch closely with the parameter of electric machine that needs are found the solution, and obtain the unknown parameter of each parameter rear motor in the equation and also can find the solution, so equation (9) is reduced to following equation:

i＝f(p，t) (10)

And be referred to as the target equation of parameter identification.

(3) discrimination method of direct current permanent magnet motor parameter: after measuring the service data of motor, can carry out identification, estimation to the parameter of motor according to these data.How choosing from numerous parameter identification methods, the method for suitable this test macro is the emphasis of research.Simultaneously, what of the different test condition and the test volume that can get will have influence on and adopt which kind of parameter identification method.

The output current of motor model can be write as

i＝f(p，t) (11)

Being that current i is the function of parameter vector p, is again the function of time t.Corresponding different p, i has different curves.Can infer that corresponding a certain definite p must have an i (t) who determines, p not simultaneously, i (t) is also different.

Will be by measured motor model and mathematical model foundation relation as shown in Figure 2.The motor model has parameter vector P _s, mathematical model has parameter vector P _m, input specific incentives signal u (t), model output current wave i _s(t) come out model output i through sensor test _m(t) can calculate.

Because P _m≠ P _s, promptly model has different parameter vectors with model, so inevitable i _s(t) ≠ i _m(t), that is to say i _s(t) and i _m(t) do not overlap, as shown in Figure 3.In the test duration, get N sampled point, i.e. K=1,2,3 ... N, pointwise is i relatively _s(K) and i _m(K) difference between is obtained the mean square value of relative value

$e=\frac{1}{N}\underset{K=i}{\overset{N}{\mathrm{\Σ}}}{\left|\frac{i_s\left(K\right)-i_m\left(K\right)}{i_s\left(K\right)+i_m\left(K\right)]/2}\right|}^2---\left(12\right)$

Obviously e has represented i _s(t) and i _m(t) degree of mutual difference is also from broadly having represented vectorial P _mWith P _sBetween distance (difference).P _mWith P _sDifference is big more, i _s(t) and i _m(t) difference is big more, and e is also big more.Otherwise, P _mWith P _sDifference is more little, i _s(t) and i _m(t) difference is more little, and e is also more little.Work as P _m=P _sThe time, e=0 must be arranged.

The parameter of tested model is the concrete numerical value of outwardness, is invariant, and the size of e depends on model parameter vector P _mWith P _sApproaching degree.So e can be regarded as vectorial P _mFunction, be expressed as

e＝f(P _m) (13)

So the problem that measures the tested parameter of electric machine has just become asks Function e=f (P _m) minimizing problem.The method of asking for extreme value is a lot, as method of conjugate gradient, Powell method, simple method or the like, all is the searching method of extreme value direction.When adopting simple method to carry out parameter identification, choose six parameter of electric machine Ra, La, C, J, Tf and Cf and debate knowledge as the six-vector of multidimensional parameter vector P.Every one step of search, P _mJust and P _sNear a step, when e reaches enough hour, can think P _m=P _s, then parameter of electric machine identification is finished.Its dynamic and static parameter that can test the direct current generator that obtains comprises: armature resistance, armature inductance, constant of the machine, moment of inertia, moment constant, electromotive force constant, dry friction torque, sticking friction factor, electromechanical time constant, electrical time constant; It can be tested the motor performance index that obtains and comprise: efficient, peak power, mechanical property slope when electric current, maximum output when torque when rotating speed, maximum output, maximum output when output power, maximal efficiency, maximum output when electric current, maximal efficiency when torque when rotating speed, maximal efficiency, maximal efficiency when no-load speed, no-load current, locked rotor current, locked-rotor torque, maximal efficiency; The motor characteristic working curve that it can measure comprises: characteristics of output power, input current characteristic, rotary speed property, efficiency characteristic.

In the operating characteristic that can obtain motor according to the mathematical model of the asking parameter of electric machine and direct current permanent magnet motor:

The input current characteristic $i=\frac{T_{\mathrm{out}}}{C}+i_0--\left(14\right)$

Rotary speed property $n=\frac{60}{2\mathrm{\π}}\·\frac{U-i{R}_a}{C}--\left(15\right)$

Characteristics of output power $P_{\mathrm{out}}=\frac{2\mathrm{\π}}{60}\·T_{\mathrm{out}}n--\left(16\right)$

Efficiency characteristic $\mathrm{\η}=\frac{P_{\mathrm{out}}}{\mathrm{Ui}}\×100\%---\left(17\right)$

When parameter identification is asked for extreme value among the present invention, will add different constraint condition, thereby guarantee to ask for the uniqueness of extreme value according to concrete motor.Speed parameter of the present invention can according to the electric machine structure characteristics, be obtained by FFT or wavelet transformation analysis by the current data of sampling.

Claims (2)

1, a kind of direct current permanent magnet motor parameter and the non-loaded method of testing of performance, the system that realizes this method of testing is made up of current sensor, voltage sensor, computing machine, power supply and data acquisition system (DAS), the present invention is to the test of the parameter and the performance of motor, basically be to be finished by software, it is characterized in that: the program that operates in the computing machine is as follows to the step of electromechanical testing: start (101); The computing machine output signal makes power supply apply the trigger voltage (U) (102) of a startup to direct current generator; By current sensor in the start-up time of direct current generator, measure N sampled point be K=1,2,3 ..., N current value i _s(K), set up the current curve (P of direct current generator _s) (103); According to trigger voltage (U) find be stored in the computing machine with the corresponding motor model curve of trigger voltage (U) (P _m), sampled point compares motor model curve (P one by one _m) at the current value i of each sampled point _m(K) with detected current value i _s(K) difference between, and obtain the mean square value e of relative value, motor model curve (P _m) the multidimensional parameter vector (P) of unique correspondence be known and be stored in (104) in the computing machine; Modification is determining motor model curve (P _m) the value of multidimensional parameter vector (P), make motor model curve (P _m) level off to current curve (P _s) (105); When the mean square value e of relative value levels off to motor model curve (P 0 time _m) value of pairing multidimensional parameter vector (P) promptly is the actual parameter vector (106) of the direct current generator of asking; Other parameter of electric machine and performance (107) according to actual parameter vector calculation direct current generator; Finish (108).

2, direct current permanent magnet motor parameter according to claim 1 and the non-loaded method of testing of performance is characterized in that selecting for use in step (105) Powell method, method of conjugate gradient, simplicial method to carry out parameter identification.

Images