Method for determining an efficiency and/or for calibrating a torque of a drivetrain, in particular of a wind turbine

Abstract

A method for determining an efficiency and/or calibrating a torque of a drivetrain comprises two tests. The drivetrain has a motor-side end at a main shaft connectable to a motor and a generator-side end, with a generator arranged between the ends. In a first test, the motor-side end of the drivetrain is driven. A variable dependent on the main shaft torque is determined at the motor-side end of the drivetrain and an electrical power Pelec is determined at the generator-side end of the drivetrain. In a second test, the generator-side end of the drivetrain is driven and the variable dependent on the main shaft torque is determined at the motor-side end and the electrical power is determined at the generator-side end. An efficiency and/or calibration parameters is/are determined from the electrical power values and the variables dependent on the main shaft torque determined in the first test and second tests.

Claims

1. A method for determining an efficiency of a drivetrain of a wind turbine, wherein the drivetrain has a motor-side end on a main shaft connectable to a motor and a generator-side end, between which ends a generator is arranged, the method comprising: performing a first test, wherein in the first test, the motor-side end of the drivetrain is driven and a variable dependent on a main shaft torque is determined at the motor-side end of the drivetrain and an electrical power P.sub.elec is determined at the generator-side end of the drivetrain; performing a second test, wherein in the second test, the generator-side end of the drivetrain is driven, and the variable dependent on the main shaft torque is likewise determined at the motor-side end and the electrical power P.sub.elec is determined at the generator-side end; and determining at least one of an efficiency or one or more calibration parameters from the electrical power values and the variables dependent on the main shaft torque determined in the first test and in the second test, using at least one predetermined assumption.

2. The method according to claim 1, wherein an assumption for determining the efficiency or for calibrating a torque measurement is the assumption that the efficiency of the first test is equal to the efficiency of the second test.

3. The method according to claim 1, wherein an assumption for determining the efficiency or for calibrating a torque measurement is the assumption that a power loss of one of the two tests is a fraction of a total power loss of the two tests.

4. The method according to claim 1, wherein to determine the torque of the main shaft, the variable dependent on the main shaft torque is measured using a sensor, and/or an angular position θ of the main shaft or a rotational speed ω of the main shaft is measured.

5. The method according to claim 1, wherein a mechanical power is determined from the variable dependent on the main shaft torque for the first and for the second test.

6. The method according to claim 1, wherein, to determine the electrical power, a voltage and a current are measured at the generator-side end between the generator and a converter connected to the same, or at the side of the converter facing away from the generator.

7. The method according to claim 1, wherein a test power loss is determined for the first test and for the second test respectively, wherein these two determined test power losses are added together for a total power loss, and wherein the efficiency is calculated using the total power loss.

8. The method according to claim 4, wherein, to determine the variable dependent on the main shaft torque, a measurement signal of the sensor is evaluated, which is arranged at an output shaft of the motor, or at a shaft adapter between the output shaft of the motor and the main shaft.

9. The method according to claim 1, wherein different operating points of the drivetrain are approached during the first test, and that different operating points of the drivetrain are approached during the second test.

10. The method according to claim 1, wherein the generator-side end of the drivetrain is driven in the second test in such a way that the variable dependent on the main shaft torque in the second test is equal to the variable dependent on the main shaft torque determined in the first test.

11. The method according to claim 1, wherein the generator-side end of the drivetrain is driven in such a way in the second test that the electrical power in the second test is equal to the electrical power determined in the first test.

12. The method according to claim 1, wherein, by using the variable dependent on the main shaft torque, a mechanical power P.sub.mechA and P.sub.mechB is determined as a function of two constants a and b for the first test and for the second test, respectively, ${\tilde{P}}_{\mathrm{mech}.A}=\frac{{\int }_{{\theta }_0}^{{\theta }_A}\left(a\epsilon +b\right)d\theta }{t_A}$ ${\tilde{P}}_{\mathrm{mech}.B}=\frac{{\int }_{{\theta }_0}^{{\theta }_B}\left(a\epsilon +b\right)d\theta }{t_B}$ and based on:
P.sub.mech.A−P.sub.elec.A=k(P.sub.elec.B−P.sub.mech.B)
and
P.sub.mech.A+kP.sub.mech.B=kP.sub.elec.B+P.sub.elec.A where k is a correlation between losses of the first test and of the second test, wherein calibration parameters a and b for determining a torque are determined using:
T=aε+b where T is the torque and c is the measured variable dependent on the main shaft torque.

13. The method according to claim 1, wherein each value of the main shaft torque determined in the first test and the second test defines a different power level of the drivetrain.

14. The method according to claim 1, wherein starting from a predetermined efficiency, the one or more calibration parameters are determined, then efficiencies are determined for at least two power levels and subsequently the one or more calibration parameters are determined again, and the last two steps are repeated until the deviation between the last and the preceding determination is smaller than a predetermined value.

15. The method according to claim 1, wherein the one or more calibration parameters are determined using the method of least squares.

16. A method for calibrating a torque measurement of a drivetrain of a wind turbine, on a test rig, wherein the drivetrain has a motor-side end on a main shaft connectable to a motor and a generator-side end, between which ends a generator is arranged, the method comprising: performing a first test, wherein in the first test, the motor-side end of the drivetrain is driven and a variable dependent on a main shaft torque is determined at the motor-side end of the drivetrain and an electrical power P.sub.elec is determined at the generator-side end of the drivetrain; performing a second test, wherein in the second test, the generator-side end of the drivetrain is driven, and the variable dependent on the main shaft torque is likewise determined at the motor-side end and the electrical power P.sub.elec is determined at the generator-side end; and determining calibration parameters a and b from the electrical power values P.sub.elec and the variables dependent on the main shaft torque determined in the first test and in the second test, using at least one predetermined assumption, wherein the first and the second test are carried out for at least two torque settings.

Description

(1) An exemplary embodiment describing the method will subsequently be described and an exemplary test rig structure will be explained in greater detail on the basis of the figures.

(2) As shown in:

(3) FIG. 1 a schematic drawing of a drivetrain of a wind turbine,

(4) FIG. 2 the schematic drawing of the drivetrain from FIG. 1, wherein the power flow is reversed,

(5) FIG. 3 schematic staircase profiles, which illustrate the energy in the first or second test, wherein the mechanical energy is the same at a first measuring point,

(6) FIG. 4 schematic staircase profiles, which illustrate the energy in the first or second test, wherein the electrical energy is the same at a first measuring point,

(7) FIG. 5 staircase profiles from two tests, wherein the generator is driven in the second test in such a way that the energy is different at the respective measuring points for the first and second tests,

(8) FIG. 6 staircase profiles of two tests in terms of the average power, and

(9) FIG. 7 a flowchart for explaining the method for optimizing the torque calibration in greater detail.

(10) FIG. 1 shows a schematic drawing of a drivetrain 1 of a wind turbine having a main shaft 2, which is connected to a generator 4 via a transmission 3. Furthermore, the main shaft is coupled to an output shaft 9 of a test rig motor 8 via an adapter 7. Generator 4 is additionally connected to a converter 5 which in turn is connected to a transformer 6. Three measuring points 10, 11, and 12 are indicated along drivetrain 1. At first measuring point 10, a variable dependent on the torque is measured by means of a sensor, for example, a strain is measured by means of a strain gauge. The sensor is arranged on the main shaft for this purpose in the example shown. The sensor may also be placed, for example, in another embodiment, at adapter 7 or at motor output shaft 9. Second measuring point 11 lies between generator 4 and converter 5. An electric current and an electric voltage of the generator may be measured here, so that an electrical power may be calculated therefrom. Alternatively to measuring point 11, a current and an electric voltage may also be measured at measuring point 12, which lies between converter 5 and transformer 6 or at a side of transformer 6 facing away from the converter, so that any losses of the converter flow into the calculated electrical power.

(11) To determine an efficiency of drivetrain 1, a first test is carried out in which motor 8 drives main shaft 2. The direction of the power flow is indicated by arrow 13A. Gradual losses of the different components of the drivetrain are schematically depicted on the basis of staircase profile 14A. The power at measuring point 10 is thus greater than the power at measuring point 11, and this is greater than the measured power at measuring point 12.

(12) FIG. 2 likewise shows the schematic drawing of drivetrain 1. In a second test, said drivetrain 1 is driven by generator 4 as a motor and the test rig runs in generator operation. The power flow is indicated by arrow 13B. Due to the drive being reversed with respect to the first test, the measured power at measuring point 12 is greater than the measured power at measuring point 11, and the measured power there is in turn greater than the power at measuring point 10, as is illustrated in staircase profile 14B.

(13) Thus, two tests are basically carried out to determine the efficiency. The tests are each carried out in such a way that the drivetrain is driven either at the same output level or at least substantially at the same output level. Different measuring scenarios are additionally possible. In a first measuring scenario, the generator-side end is driven in the second test in such a way that the mechanical power values of the first and second tests are equal. In another measuring scenario, the electrical power values of the first and second tests are kept the same. These scenarios are shown in FIGS. 3 and 4. In another possibility, however, the drivetrain may also be driven in the second test without maintaining the mechanical or electrical power values as constant for tests one and two. This scenario is shown in FIG. 5.

(14) FIG. 3 shows the staircase profiles of first and second tests 14A and 14B according to the first measuring scenario, wherein generator 4 is driven in the second test in such a way that the mechanical power at first measuring point 10 in the first test is equal to the mechanical power at first measuring point 10 in the second test. I.e., the measured value for the measured mechanical torque is kept constant and equal in the two tests. Thus, there arises at measuring point 11 a total power loss ΔP between the first and the second test, which is detected by measuring the current and the voltage in each case in the two tests.

(15) FIG. 4 shows the staircase profiles of first and second tests 14A and 14B, wherein generator 4 is driven in the second test according to the second measuring scenario in such a way that the electrical power at second measuring point 11 in the first test is equal to the electrical power at second measuring point 11 in the second test, i.e., the electrical power is respectively measured in the two tests, wherein the electrical power in the second test is set to that of the first test. Thus, there arises at measuring point 10 a total power loss in mechanical power ΔP between the first and the second test.

(16) Staircase profiles 14A and 14B from two tests are depicted in FIG. 5, wherein in the second test, generator 4 is driven in such a way that the energy is different at respective measuring points 10, 11, and 12 for the first and second tests. The electrical and mechanical power values depicted in the figure are the power values determined across the duration of the test, i.e., the energies divided by the time duration of the tests.

(17) Using an incremental encoder, an angular position is measured for the first test and for the second test, from which a torque T or an energy and/or power

(18) ${\tilde{P}}_{\mathrm{mech}.A}=\frac{{\tilde{E}}_{\mathrm{mech}.A}}{t_A}=\frac{{\int }_{{\theta }_0}^{{\theta }_A}\mathrm{Td}\theta }{t_A}$${\tilde{P}}_{\mathrm{mech}.B}=\frac{{\tilde{E}}_{\mathrm{mech}.B}}{t_B}=\frac{{\int }_{{\theta }_0}^{{\theta }_B}\mathrm{Td}\theta }{t_B}$
may be determined. Furthermore, a current I and a voltage U are each measured for the first and the second tests. From these measured variables, a mechanical power may be calculated
and an electrical power may be calculated for the first test
P.sub.elec.A=E.sub.elec.A/t.sub.A
for the first test, and at
P.sub.elec.B=E.sub.elec.B/t.sub.B

(19) The tildes, which in the above equations indicate averaged values, are left out of the subsequent equations for the sake of simplicity.

(20) The test power loss for the first test results from
P.sub.Loss.A=P.sub.mech.A−P.sub.elec.A
And the test power loss for the second test results from
P.sub.Loss.B=P.sub.elec.B−P.sub.mec.hB.

(21) By adding the test power losses P.sub.Loss.A and P.sub.Loss.B, a total power loss P.sub.Loss.total may be calculated:
P.sub.Loss.total=P.sub.Loss.A+P.sub.Loss.B=P.sub.mech.A−P.sub.mech.B+P.sub.elec.B−P.sub.elec.A.

(22) By using

(23) ${\tilde{P}}_{\mathrm{Loss}.A}=\frac{{\tilde{P}}_{\mathrm{mech}.A}}{{\tilde{P}}_{\mathrm{mech}.A}+{\tilde{P}}_{\mathrm{elec}.B}}{\tilde{P}}_{\mathrm{Loss}.\mathrm{total}}$${\tilde{P}}_{\mathrm{Loss}.B}=\frac{{\tilde{P}}_{\mathrm{elec}.B}}{{\tilde{P}}_{\mathrm{mech}.A}+{\tilde{P}}_{\mathrm{elec}.B}}{\tilde{P}}_{\mathrm{Loss}.\mathrm{total}}$
a test power loss P.sub.Loss.A and P.sub.Loss.B may be determined for each test.

(24) The determined values may then be inserted into

(25) $\mathrm{Effi}=\frac{{\tilde{P}}_{\mathrm{elec}.A}}{{\tilde{P}}_{\mathrm{mech}.A}}=\frac{{\tilde{P}}_{\mathrm{elec}.A}}{{\tilde{P}}_{\mathrm{elec}.A}+{\tilde{P}}_{\mathrm{Loss}.A}}$$\mathrm{or}$$\mathrm{Effi}=\frac{{\tilde{P}}_{\mathrm{mech}.B}}{{\tilde{P}}_{\mathrm{elec}.B}}=\frac{{\tilde{P}}_{\mathrm{elec}.B}-{\tilde{P}}_{\mathrm{Loss}.B}}{{\tilde{P}}_{\mathrm{elec}.B}}$
in order to obtain efficiency Effi. P.sub.mech.A, P.sub.mech.B, P.sub.elec.B and P.sub.elec.A are thereby measured, so that P.sub.Loss.total may be determined therefrom. A division of P.sub.Loss.total yields P.sub.Loss.A or P.sub.Loss.B.

(26) The efficiency may be calculated from P.sub.Loss.A and P.sub.elec.A or from P.sub.Loss.B and P.sub.elec.B. Normally, an electrical measurement is thereby more accurate, so that preferably P.sub.elec.A or P.sub.elec.B may be used to calculate the efficiency.

(27) FIG. 6 shows two staircase profiles 14A and 14B. The electrical power at measuring point 11 in the first test is set to be substantially equal to the power at measuring point 11 in the second test. The power of the first and second test for measuring point 10 may be described as
P.sub.mech.total=k.sub.AP.sub.mech.A+k.sub.BP.sub.mech.B
and may be described for measuring point 11 as
P.sub.elec.total=k.sub.BP.sub.elec.A+k.sub.AP.sub.elec.B
whereby
k.sub.A=Effi/(1+Effi),k.sub.B=1/(1+Effi), and Effi is the efficiency of the drivetrain.

(28) FIG. 6 thereby shows staircase profiles 14A and 14B, which occur when the assumption is made that the efficiency of the first test is equal to the efficiency of the second test (Effi.sub.A=Effi.sub.B=Effi), so that the following equation applies:
k.sub.AP.sub.mech.A+k.sub.BP.sub.mech.B=k.sub.BP.sub.elec.A+k.sub.AP.sub.elec.B.

(29) If the two tests are carried out such that the wind turbine operates at a similar operating point in the first and in the second test, then the power losses of the two tests should likewise be similar or have a certain correlation. The loss in each test may be assumed to be half or a part of the total loss with a certain degree of uncertainty. The uncertainty, which is introduced by the assumption of the losses in the two operating modes, may be analyzed using the behavior of the generator and other important mechanical and electrical components of the turbine.

(30) FIG. 7 shows a flowchart for explaining the method for optimizing the torque calibration in greater detail. The basis for this method are the two tests, which were explained on the basis of the previous figures.

(31) In step I, calibration factors a, b.sub.ini may initially be determined arising from a predetermined efficiency, which was, if necessary, determined from previous experiments and a measured value ε.

(32) Subsequently in step II, the efficiencies (Eff.sub.I, Effi.sub.II, . . . ) may be determined for at least two power levels using a variable dependent on the torque. Two tests are thereby carried out for each power level and the variables dependent on the torque and the electrical power values are measured in the first and in the second test.

(33) In third step III, calibration parameters a, b.sub.ini may be recalculated using the determined efficiency, and this may be carried out for all power levels. If the recalculated calibration parameters a, b.sub.neu deviate strongly from the previously determined calibration parameters, then the efficiencies may be redetermined for the respective power levels and these results compared with the former values. This may be repeated any number of times until the determined calibration parameters only deviate slightly from those previously determined. Thus, a torque measurement may also be calibrated for large torques and carried out with great accuracy.

(34) The described torque calibration may be used, as described above, for drivetrains of wind turbines, but also for other drivetrains of other electrical machines.

(35) It is thereby important that both sides, thus the test rig motor and the generator motor, may run both in motor and in generator operation.

(36) For improved accuracy of the calibration, the machine operated as a generator in the first test may be selected so that the power loss in both modes (thus in the first and in the second test), is as similar as possible.

(37) A specific application may be that two electrical machines of the same type are used. The test rig is replaced by one of the machines in this case. The two machines run in so-called “back to back” mode, wherein a measuring body to be calibrated is installed on the drive train between the machines.