Method for estimating moment of inertia of electric motor and load system, and electric motor for carrying out method

Abstract

The invention relates to a method for estimating the moment of inertia of an electric motor and load system during normal operation and the invariant load characteristics of the system. The method comprises the steps of: operating the electric motor in a normal operating mode having a constant load characteristic via the electric motor drive; estimating instantaneous mechanical power; calculating a difference in power integration during a speed increase and decrease of the electric motor, wherein the speed increase and decrease correspond to transients and / or disturbances; and calculating the rotational inertia according to the measured power difference integral. The invention also relates to an electric motor drive provided for carrying out the method.

Description

Technical Field

[0001] This invention relates to a method for estimating the moment of inertia of an electric motor and load system during normal operation and the system's constant load characteristics. The method includes the following steps:

[0002] The electric motor is operated in a normal operating mode with constant load characteristics via an electric motor driver;

[0003] Estimate instantaneous mechanical power;

[0004] Calculate the difference between the power integrals during the ramp-up and ramp-down phases of the electric motor's speed, where the ramp-up and ramp-down correspond to transients and / or disturbances; and

[0005] The moment of inertia is calculated by integrating the measured power difference.

[0006] The present invention also relates to an electric motor driver configured to perform the method. Background Technology

[0007] When operating an electric motor and load system, the system's moment of inertia can be determined to allow for optimization of the system's dynamic performance. The moment of inertia can also be used for condition monitoring of applications connected to the system. The system's moment of inertia can refer to the combined inertia of the motor and its load.

[0008] Known methods for providing estimations of the moment of inertia of electric motors and load systems typically rely on accurate torque estimation, which in turn requires precise motor data. Furthermore, most known methods depend on dedicated operating sequences, which is intolerable in applications where uninterrupted normal operation of the system is essential.

[0009] Therefore, these known inertia estimation methods require fairly comprehensive hardware systems, can be time-consuming, and are prone to inaccuracies in system parameters. Furthermore, known inertia estimation methods may cause harmful system downtime during their execution. Summary of the Invention

[0010] The object of this invention is to overcome these problems by providing an improved method for estimating the rotational inertia of an electric motor and load system, and a corresponding electric motor drive. Preferred embodiments of the invention are limited by the dependent claims.

[0011] According to claim 1, a method is provided for estimating the moment of inertia of an electric motor and load system during normal operation and the constant load characteristics of the system. The method includes the following steps:

[0012] The electric motor is operated in a normal operating mode with time-invariant load characteristics via an electric motor driver;

[0013] Estimate instantaneous mechanical power;

[0014] Calculate the difference between the power integrals during the speed increase and decrease of the electric motor, where the speed increase and decrease correspond to transients and / or disturbances; and

[0015] The moment of inertia is calculated by integrating the measured power difference.

[0016] According to this method, the drive output power or instantaneous mechanical power of the motor is measured and integrated during the increase and decrease of the motor speed. Since the motor load maintains the speed as a time-invariant function, the difference in the integrals represents the rotational energy of the system. Inertia is thus calculated.

[0017] The method described herein allows for the use of naturally occurring or induced velocity changes in the system, where motor speed and power are used to determine inertia.

[0018] This method is independent of motor type, load, and control core, and can be used during normal system operation. Furthermore, it does not depend on the availability of precise motor parameters.

[0019] In another preferred embodiment of the invention, the moment of inertia is calculated according to the following equation (2.19).

[0020]

[0021] Where ΔE mech Calculated by the following equation (2.18),

[0022]

[0023] The term x, which describes the ratio of the descent duration to the ascent duration, is defined in the following equation (2.15).

[0024]

[0025] Furthermore, E1 and E2 are calculated by the following equation (2.11),

[0026] .

[0027] In another preferred embodiment of the invention, the speed increase and decrease correspond to the overshoot transient if the overshoot transient during the acceleration of the electric motor to the reference speed has a sufficiently long duration. The overshoot needs to be long enough for the motor driver performing the method to collect sufficient data for calculating the power integral.

[0028] In another preferred embodiment of the invention, estimating instantaneous mechanical power includes measuring the output of the electric motor driver and subtracting the estimated electric motor losses.

[0029] In another preferred embodiment of the invention, increasing and decreasing the speed of the electric motor includes superimposing a preferably symmetrical speed increase and decrease sequence onto the electric motor speed in the normal operating mode of the electric motor, wherein the speed of the electric motor in its normal operating mode is preferably a stable speed. Once the speed increase and decrease sequence has been superimposed, the resulting motor speed is obviously no longer stable. However, the speed increase and decrease sequence can be selected such that the operation driven by the motor is not interrupted.

[0030] In another preferred embodiment of the invention, a speed boost is performed before or after the speed decrease.

[0031] In another preferred embodiment of the invention, the method is performed once after each start of the electric motor and / or periodically at configurable time intervals.

[0032] The present invention also relates to an electric motor driver configured to perform the methods described herein. Attached Figure Description

[0033] Further advantages and details of the invention are described with reference to the embodiments shown in the accompanying drawings, wherein features of the various embodiments can be combined in any logically possible manner within the scope of the invention. The drawings show:

[0034] Figure 1 : A graph showing how the speed and power of a motor change over time;

[0035] Figure 2 : A circuit diagram indicating the current flowing through the motor stator and rotor, indicating losses;

[0036] Figure 3 A graph indicating the motor's lifting speed, showing both reference and actual speeds;

[0037] Figure 4 : A graph indicating the motor overshoot transient simulation used for inertia estimation;

[0038] Figure 5 : A graph indicating motor speed disturbances with reference speed and actual speed; and

[0039] Figure 6 : A graph indicating the motor speed disturbance simulation used for inertial estimation. Detailed Implementation

[0040] In this specification, the following terms will be used:

[0041] T = Torque

[0042] t = time

[0043] J = Moment of inertia

[0044] ω = mechanical velocity

[0045] p = instantaneous power

[0046] E = Energy

[0047] W = Work (energy consumed by the load)

[0048] x = ratio

[0049] u / U, i / I=voltage, current

[0050] R = resistance

[0051] Starting from this point, the familiar law of motion for a rotating element is the following equation (2.1):

[0052] .

[0053] The load can be any function of speed, as described by the following equation (2.2):

[0054] .

[0055] And it can be used with a constant, quadratic, or other relationship with the velocity, or it can be considered zero. The instantaneous mechanical power of the rotating system during acceleration is given by the following equation (2.3):

[0056] .

[0057] The power can be estimated relatively accurately because the drive output power is measured and motor losses can be estimated.

[0058] Integrating the power over a given time span produces the energy given by the following equation (2.4):

[0059] .

[0060] It is evident that energy comprises two parts: T acc The energy change in the rotating system is provided, as well as the energy consumed by the load during that period. If no acceleration occurs, it is clear that the energy stored in the rotating system remains constant when the first integral is zero.

[0061] The following section examines different techniques used to estimate velocity variations.

[0062] In the first embodiment of the present invention, the same ramp-up and ramp-down sequence is applied, such as... Figure 1 As shown. This sequence can be superimposed on the existing operating (steady-state) rate and is therefore considered a small signal.

[0063] Power integration is performed in each ramp sequence, and since the velocity distribution is inverse but symmetrical, the difference between the two integrals will represent the actual acceleration torque, because it is assumed that the load energy is the same and therefore cancels each other out according to the following equation (2.5):

[0064] .

[0065] The extension of these integrals produces the energy difference between the velocity ramp sequences according to the following equation (2.6):

[0066] .

[0067] If the velocity distribution is symmetrical but opposite over the two time series, then the two accelerating torques are equal in magnitude but opposite in sign. The two time spans are zero, therefore t1 - t0 = t3 - t2 = Δt. Thus, the energy difference simplifies to the following equation (2.7):

[0068] .

[0069] The energy difference ΔE represents twice the change in rotational energy from the initial velocity to the final velocity. The velocity trajectory between these points is theoretically unimportant. Generally, the rotational energy at constant velocity and inertia is given by the following equation (2.8):

[0070] .

[0071] By combining it with equation (2.7), we inevitably arrive at (the following formula 2.9):

[0072] ,

[0073] Where ω1 and ω0 are the velocities at the end and beginning of the disturbance slope, respectively.

[0074] In summary, the moment of inertia can be found in the following equation (2.10):

[0075] ,

[0076] Assumption:

[0077] Δω is the speed difference between the rising and falling sequences.

[0078] Δt = t1 - t0 = t3 - t2 represents the duration of the rising and falling sequences.

[0079] • The load torque behavior is considered to be equal in both ramp sequences, i.e., the load is time-invariant.

[0080] As long as these conditions are met, the actual velocity and acceleration torque dynamics during the ramp sequence are irrelevant.

[0081] In another embodiment, a more general case is considered with different rise and fall times, where the speed difference is equal according to the following equation (2.11):

[0082] .

[0083] In this more generalized case, the acceleration torque is not the same in the two sequences. However, the change in stored mechanical energy must be equal, so even if the acceleration torque is not equal, the associated time difference will compensate for it.

[0084] Furthermore, it is assumed that the product of load torque and speed is a general load power function p. l (t). Considering all of these, the expression simplifies to the following equation (2.12):

[0085]

[0086] ,

[0087] Among them W l It is the inverse derivative of the load power function and describes the energy dissipated (or the work done) in the load. It can be arbitrary, but is assumed to be time-invariant. Furthermore, it is assumed that due to time invariance, the difference of the following equation (2.13) It is proportional to the time difference itself, as shown in equation (2.14):

[0088] ,

[0089] Where x describes the ratio of the falling time interval to the rising time interval as defined in the following equation (2.15).

[0090] .

[0091] This will apply to non-oscillatory transitions from one velocity to another (and back again). Substituting equation (2.14) into matrix form yields the following equation (2.16):

[0092] ,

[0093] As shown in equation (2.17), the vector containing load energy and mechanical energy can be solved.

[0094] .

[0095] The change in mechanical energy can be calculated using the following equation (2.18).

[0096] .

[0097] Therefore, after discovering the change in mechanical energy, inertia can be found from the aforementioned formula, that is, inertia can be found according to the following equation (2.19).

[0098] .

[0099] When the rise time and fall time are equal, x = 1, and (2.18) becomes the same as (2.7). In other words, the factor x takes into account the magnitude of the non-uniform energy caused by the non-uniform integral time interval due to the non-uniform slope.

[0100] The inertia estimate depends on the estimate of the shaft power, which can only be indirectly obtained from the measurable electrical output power and the motor losses estimated according to the following equation (3.1).

[0101] .

[0102] Note that power is considered positive when flowing from the driver to the motor. Since losses are always positive, mechanical power is numerically less than electrical power in the case of a motor drive.

[0103] Under generator load conditions, both mechanical power and electrical power are negative. Then, according to equation (3.1), the mechanical power becomes numerically larger than the electrical power expected in this case. Therefore, equation (3.1) holds true in both motor-driven and generator-driven operating modes.

[0104] In the dq coordinate system, the power is calculated accordingly based on the following equation (3.2).

[0105] .

[0106] However, loss calculations require some consideration, especially for induction motors, to differentiate between the currents located in the stator and rotor. Figure 2 The diagram illustrates how stator and rotor currents can be used to predict motor losses. Iron losses are discarded or ignored.

[0107] For synchronous motors, only stator winding losses need to be considered, since iron losses are again discarded or ignored.

[0108] Table 1 below summarizes how to calculate motor losses in each case.

[0109]

[0110] According to the present invention, inertia can be estimated based on a sequence of increasing and decreasing velocities. It does not need to be symmetrical, as long as the sequence starts and ends at the same velocity; that is, the increasing and decreasing components do not need to be the same. Two operating modes for inertia estimation are described below.

[0111] Speed ​​overshoot typically occurs in applications with high inertia and where the speed controller is not actively tuned. Therefore, overshoot transients can be used for both speed increases and decreases, provided they are large enough. A proposal is shown below.

[0112] Figure 3 A schematic diagram is shown illustrating the motor speed increase based on a reference speed and overshoot. Counter 1 represents the duration of the overshoot transient. If this exceeds a lower threshold (Counter 1 min, Counter 1 min), the transient is large enough to estimate the inertia, and the function terminates when this is complete. In this case, the estimation is purely passive.

[0113] Counter 2 is started when the maximum speed is detected, and both counters 1 and 2 terminate when the speed has decreased back to the value at which the counters were started. Therefore, the difference between counter 1 and counter 2 is the duration of the speed increase, and counter 2 is the duration of the speed decrease.

[0114] Figure 4 The simulation of the overshoot transient for inertial estimation is shown. The top curve shows the velocity, the second curve from the top is the power integrator 1 and power integrator 2, the third curve from the top is the counter 1 and counter 2, and the bottom curve is the state and disturbance velocity output, which is zero in this case because the function does not command any velocity disturbance in this case.

[0115] Note that the different speed increases and decreases are used to calculate the ratio x according to equation (2.15).

[0116] If the transient is too short or too small to be determined, a propagation ramp output is superimposed on the velocity reference signal, and inertia is estimated based on this disturbance, such as... Figure 5 As shown.

[0117] Figure 6 The diagram illustrates the functionality. The top curve shows the mechanical speed, the second curve from the top shows power integrator 1 and power integrator 2, the third curve from the top shows counter 1 and counter 2, and the bottom curve shows the status and electrical disturbance speed output.

Claims

1. A method for estimating the moment of inertia of an electric motor and a load system during normal operation and the constant load characteristics of the system, comprising the following steps: - Operate the electric motor in normal operating mode with time-invariant load characteristics via the electric motor driver; - Estimate instantaneous mechanical power; - Calculate the difference in power integrals during the speed increase and decrease of the electric motor, where the speed increase and decrease correspond to transients and / or disturbances; as well as - Calculate the moment of inertia by integrating the measured power difference.

2. The method according to claim 1, characterized in that, Calculate the moment of inertia according to the following equation (2.19). ， Where ΔE mech Calculated by the following equation (2.18), ， The term x, which describes the ratio of the descent duration to the ascent duration, is defined in the following equation (2.15). Furthermore, E1 and E2 are calculated by the following equation (2.11), 。 3. The method according to any one of the preceding claims, characterized in that, If the overshoot transient during the acceleration of the electric motor to the reference speed has a sufficiently long duration, then the speed acceleration and deceleration correspond to the overshoot transient.

4. The method according to any one of the preceding claims, characterized in that, Estimating instantaneous mechanical power involves measuring the output of the electric motor driver and subtracting the estimated electric motor losses.

5. The method according to any one of the preceding claims, characterized in that, Increasing and decreasing the speed of the electric motor includes superimposing a preferred symmetrical speed increase and decrease sequence on the electric motor speed in the normal operating mode of the electric motor, wherein preferably, the electric motor speed in the normal operating mode of the electric motor is a stable speed.

6. The method according to any one of the preceding claims, characterized in that, The speed increase is performed before or after the speed decrease.

7. The method according to any one of the preceding claims, characterized in that, The method is performed once after each start of the electric motor and / or periodically at configurable time intervals.

8. An electric motor driver configured to perform the method according to any one of the preceding claims.

Images