Speed ​​control method based on load torque-moment of inertia self-learning model

Abstract

A speed control method based on a load torque-moment of inertia self-learning model is applied to a motor controller. The speed control method includes: establishing a load torque-moment of inertia relationship through self-learning; obtaining the corresponding moment of inertia value based on the load torque value; and adjusting the controller parameters based on the moment of inertia value to control the motor operation.

Description

Technical Field

[0001] This invention relates to a speed control method, and more particularly to a speed control method based on a load torque-moment of inertia self-learning model. Background Technology

[0002] The application of electric motors paired with frequency converters is widely used in systems such as elevators, cranes, and escalators. To improve the operational performance of the drive system, speed controllers are typically designed using the system's relevant mechanical parameters to meet operational performance requirements.

[0003] Taking an elevator as an example, since the moment of inertia of the motor is positively correlated with the mass of the load, the moment of inertia of the motor will change when the weight (mass) of the load is different (i.e., the load capacity is different). Therefore, if the moment of inertia of the motor can be estimated, the parameters of the motor controller can be adjusted according to the estimated value of the moment of inertia, which helps to control the speed of the motor more accurately and improves the control performance of the motor.

[0004] Therefore, how to design a speed control method based on a load torque-moment of inertia self-learning model to achieve the aforementioned technical effects is an important research topic for the inventors of this disclosure. Summary of the Invention

[0005] The purpose of this invention is to provide a speed control method based on a load torque-moment of inertia self-learning model, thereby solving the problems of the prior art.

[0006] To achieve the aforementioned objectives, the speed control method based on a load torque-moment of inertia self-learning model proposed in this invention is applied to a controller for controlling a motor. The speed control method includes: (a) establishing a load torque-moment of inertia relationship through self-learning; (b) obtaining the corresponding moment of inertia value based on the load torque value; and (c) adjusting the controller parameters based on the moment of inertia value to control the operation of the motor.

[0007] In one embodiment, step (a) includes: (a1) obtaining the value of the load torque under zero speed control; (a2) obtaining the value of the corresponding moment of inertia under acceleration control; and (a3) ​​repeating steps (a1) and (a2) to establish the load torque-moment of inertia relationship.

[0008] In one embodiment, step (a2) includes: (a21) calculating the value of the corresponding moment of inertia using integral operations.

[0009] In one embodiment, the load torque-moment of inertia relationship is a lookup table.

[0010] In one embodiment, the load torque-moment of inertia relationship is a curve fitting equation.

[0011] In one embodiment, the speed control method further includes: (d) updating the load torque-moment of inertia relationship.

[0012] In one embodiment, when a new load torque value is determined, the corresponding new moment of inertia value is obtained, and the load torque-moment of inertia relationship is updated.

[0013] In one embodiment, the load torque and the moment of inertia are in a one-to-one relationship.

[0014] In one embodiment, when the relationship between load torque and moment of inertia is one-to-many, an arithmetic average is performed on multiple moments of inertia to obtain an average moment of inertia, so that the relationship between load torque and average moment of inertia is one-to-one.

[0015] In one embodiment, in step (a), the speed information of the motor is obtained through the speed control loop, and the torque information of the motor is obtained through the current control loop, so as to establish the load torque-moment of inertia relationship through self-learning.

[0016] The proposed speed control method based on the load torque-moment of inertia self-learning model can achieve the following technical effects: (1) The value of the corresponding moment of inertia is calculated by integral operation to solve the problem of high-frequency noise caused by the use of differential operation in the prior art, which requires the addition of a filter; (2) The information required for parameter estimation is directly obtained by using the existing speed control loop and current control loop architecture, so the operation of the closed loop and the drive control of the motor are not affected during the parameter (moment of inertia) estimation process; (3) When the value of moment of inertia is obtained by parameter estimation, the parameters of the controller can be adjusted according to the value of moment of inertia to control the operation of the motor, which helps to control the speed of the motor more accurately and improves the control performance of the motor, including acceleration performance, transient response and load rejection capability.

[0017] To gain a deeper understanding of the techniques, means, and effects employed by this invention to achieve its intended purpose, please refer to the following detailed description and accompanying drawings. It is believed that the purpose, features, and characteristics of this invention can be understood in a thorough and specific manner from these drawings. However, the drawings are provided for reference and illustration only and are not intended to limit the scope of this invention. Attached Figure Description

[0018] Figure 1 This is a flowchart of the speed control method based on the load torque-moment of inertia self-learning model of the present invention.

[0019] Figure 2: A flowchart for establishing the load torque-moment of inertia relationship through self-learning in this invention.

[0020] Figure 3 : This is a schematic curve illustrating the speed variation of the motor according to the present invention.

[0021] Figure 4 : This is a schematic curve of the motor output torque and load torque of the present invention.

[0022] Figure 5 This is a schematic diagram of the curve fitting relationship between load torque and moment of inertia of the present invention.

[0023] Figure 6 This is a flowchart illustrating the application of the speed control based on the load torque-moment of inertia self-learning model of the present invention.

[0024] Figure 7 : This is a block diagram of the motor drive system of the present invention.

[0025] Explanation of reference numerals in the attached figures:

[0026] S11~S13: Steps

[0027] S111~S113: Steps

[0028] S100~S520: Steps

[0029] I~III: Interval

[0030] C1~C4: Curves

[0031] T L Load torque

[0032] J m Moment of inertia Detailed Implementation

[0033] The technical content and detailed description of the present invention are explained below with reference to the accompanying drawings.

[0034] Please see Figure 1 The diagram shows a flowchart of the speed control method based on a load torque-moment of inertia self-learning model according to the present invention. The speed control method of the present invention is applied to a controller (also referred to herein as a speed controller) for controlling a motor. This speed control method includes the following steps: First, a self-learning process is used to establish the load torque-moment of inertia relationship (S11).

[0035] See also Figure 2In step (S11), the values ​​of the load torque and moment of inertia are obtained (or estimated) based on the different mechanical motion characteristics of the motor under zero speed and acceleration. Specifically, step (S11) includes the following steps: obtaining (estimating) the load torque value under zero speed control (S111), and then obtaining (estimating) the corresponding moment of inertia value under acceleration control (S112). Furthermore, steps (S111) and (S112) are repeated to achieve self-learning and establish the load torque-moment of inertia relationship (S113), as explained in detail below.

[0036] The equation of motion for an electric motor can be expressed as equation (1):

[0037]

[0038] Among them, J m For rotational inertia, ω m Mechanical angular velocity (after differentiation) (The following is mechanical angular acceleration), T e For motor output torque, T L For load torque, B m is the coefficient of viscous friction.

[0039] Rearranging relation (1), we get:

[0040] J m dω m =(T e -T L -B m ω m )dt (2)

[0041] Integrating both sides of relation (2), we get:

[0042] J m ω m =∫(T) e -T L -B m ω m )dt (3)

[0043] Rearranging equation (3) by terms, we obtain the equation for estimating the moment of inertia:

[0044]

[0045] For elevator systems, J is usually m >>B m Therefore, if J m >>B m Then the equation for estimating the moment of inertia can be simplified to:

[0046]

[0047] See also Figure 3 and Figure 4 The first interval is the load torque T. L The second interval is for the estimation of rotational inertia, and the third interval is for the estimation of the viscous friction coefficient B. m The estimate is explained in detail below.

[0048] As described in the previous step (S111): Under zero-speed control, obtain (estimate) the value of the load torque. In the first interval, as... Figure 3 and Figure 4 The zero-speed control is shown between second 0 and second 2. Figure 3 The motor speed curve shown (C1 is zero) allows us to obtain (estimate) the load torque T. L The value of is equal to the motor output torque T. e The value of is the motor output current (i). q ) and torque constant (K t The product of ) (i.e., T) e =i q *K t ), Figure 4 The load torque curve C2 and the motor output torque curve C3 overlap in the first interval. For example, after the elevator doors close, a mechanical brake is applied. Once the mechanical brake is released, the elevator is controlled to zero speed. When the speed is zero, the viscous friction coefficient B... m It is zero, and the mechanical angular acceleration is zero. It is also zero. Therefore, according to equation (1), the estimated load torque T can be obtained. L That is, the known motor output torque T e (=i q *K t Therefore, the load torque T can be estimated in the first interval (zero speed control interval). L The value, that is, the corresponding Figure 2 Step (S111).

[0049] Then, following the previous step (S112): under acceleration control, the corresponding moment of inertia is obtained (estimated). In the second interval, as... Figure 3 and Figure 4 The interval from the 2nd to the 7th second shown is for acceleration control. Figure 3 The motor speed curve C1 shown gradually increases, from which the (estimated) moment of inertia J can be obtained. m The value of . For example Figure 4 As shown in the initial acceleration phase (i.e., from rest to acceleration), additional motor output torque T is required to overcome the moment of inertia.e That is, the motor output torque curve C3 increases sharply in the initial stage of acceleration. However, after overcoming the moment of inertia, the motor output torque curve C3 decreases significantly. Therefore, according to the relationship (5), in the second interval (acceleration control interval), the motor output torque T is... e With load torque T L Integrate the difference between them, then divide by the mechanical angular velocity ω. m The moment of inertia J can then be obtained (estimated). m The value of .

[0050] Therefore, by repeating steps (S111) and (S112), the load torque-moment of inertia relationship can be established through self-learning.

[0051] Incidentally, in interval III, such as Figure 3 and Figure 4 As shown, between the 7th and 12th seconds, the motor operates at an almost constant angular velocity, therefore, the mechanical angular acceleration... The value is zero. Therefore, according to equation (1), the motor output torque T is zero. e With load torque T L The difference between them, divided by the mechanical angular velocity ω m The viscous friction coefficient B can then be obtained (estimated). m The value of , where, for example Figure 4 Curve C4 shown represents the viscous friction coefficient B. m With mechanical angular velocity ω m The product of.

[0052] Therefore, by executing step (S11), a load torque-moment of inertia relationship can be established through self-learning. In one embodiment, the load torque-moment of inertia relationship is a lookup table, meaning that the relationship is established by mapping a load torque value to at least one moment of inertia value. Incidentally, during the process of establishing the load torque-moment of inertia relationship, one load torque value may correspond to more than two moments of inertia values. Therefore, the values ​​of the two or more moments of inertia can be calculated using an arithmetic average, but this invention is not limited to this. An average moment of inertia value is calculated and used as the corresponding load torque value. Examples are given in Tables 1 and 2 below:

[0053] Table 1

[0054]

[0055] As shown in Table 1, the load torque T L Compared with the estimated moment of inertia It is a one-to-one relationship, that is, when the load torque T is obtained...L The value is T L1 In this case, the estimated moment of inertia can be obtained by looking up a table. The value is J m1 Similarly, when the load torque T is obtained... L The value is T L3 In this case, the estimated moment of inertia can be obtained by looking up a table. The value is J m3 .

[0056] Table 2

[0057]

[0058] As shown in Table 2, due to a load torque T L Corresponding to multiple estimated moments of inertia For example, in the process of establishing the load torque-moment of inertia relationship through self-learning, the same load torque T L3 The estimated moment of inertia obtained from multiple estimations There are many, that is, J m31 J m32 …J m3k Therefore, J can be calculated using an arithmetic mean. m31 J m32 …J m3k Averaging them together yields the estimated moment of inertia. Average J m3-avg This causes the load torque T L Compared with the estimated moment of inertia The average is a one-to-one relationship, which is helpful for finding the load torque T. L The value is T L3 At that time, the estimated moment of inertia is obtained. The value is the estimated moment of inertia. Average J m3-avg Incidentally, in the arithmetic mean of multiple estimated moments of inertia At this time, abnormally high or low (unreasonable) estimates of the moment of inertia can be identified. First delete the components, then perform an arithmetic mean calculation to obtain a more accurate estimate of the moment of inertia. value.

[0059] In another embodiment, the load torque-moment of inertia relationship is a curve-fitting relationship, see [reference needed]. Figure 5 As shown. By sampling (obtaining) several discrete load torques T L The value (for) Figure 5The x-axis is represented by a continuous mathematical function (linear equation) to express the moment of inertia J. m The value (for) Figure 5 (vertical axis) and load torque T L The relationship between the two values, therefore, through the known load torque T L The value can be solved to obtain the moment of inertia J. m The value of .

[0060] Among them, the form or data processing method of the aforementioned lookup table or curve fitting formula can be planned and designed according to hardware conditions such as memory capacity, microprocessor processing speed, and network capability to achieve the best and most real-time estimation performance.

[0061] Based on the established load torque-moment of inertia relationship, after step (S11), the corresponding moment of inertia value is obtained according to the load torque value (S12). Since the load torque-moment of inertia relationship has been established in step (S11), the corresponding moment of inertia value can be found by looking up a table, i.e., based on the load torque value; or, the corresponding moment of inertia value can be calculated by substituting the load torque value into the fitted mathematical function using a curve fitting method. Finally, the controller parameters are adjusted according to the moment of inertia value to control the motor operation (S13).

[0062] Please see Figure 6 The diagram shows a flowchart of the speed control application based on the load torque-moment of inertia self-learning model of the present invention. After the elevator is installed, the load torque T is estimated first through zero-speed control. L The value of the load torque T L Equal to the motor output torque T e (S100) can correspond to the previously disclosed Figure 2 Step (S111) and Figure 3 , Figure 4 The detailed description of the operation in the first interval is omitted here. Then, it is determined whether inertia information for this load is available (S200), i.e., whether information on the rotational inertia of the motor is available. If inertia information for this load is not available, the estimation of rotational inertia is enabled (S210), allowing the motor to operate under preset parameters (S220), and rotational inertia estimation and data collection are performed (S230). According to the aforementioned... Figure 2 Step (S112) and Figure 3 , Figure 4The record of the operation in the second interval is obtained (estimated) through acceleration control. Therefore, through the execution of steps (S210) to (S230), information on the load torque-moment of inertia relationship can be obtained, which can be regarded as establishing the initial load torque-moment of inertia relationship. Therefore, the established load torque-moment of inertia relationship (including the method of lookup table or curve fitting relationship) can update the existing load torque-moment of inertia relationship (S240), making the load torque-moment of inertia relationship more complete.

[0063] Furthermore, if the judgment in step (S200) is "yes," meaning the inertia information of this load is available, then it is further determined whether the model's self-learning is complete (S300). If the model's self-learning has been completed, meaning the judgment in step (S300) is "no," then it can be determined by looking up a table or curve fitting formula, based on different load torques T. L The value of the value (e.g., the number of passengers in the elevator) is used to obtain the corresponding moment of inertia J. m The value (S500). Furthermore, it can be determined based on the moment of inertia J. m The value of the load torque-moment of inertia is used to adjust the parameters of the motor controller, i.e., to control the speed of the motor (S510) to control the operation of the motor (S520). This helps to control the speed of the motor more accurately and improves the control performance of the motor. Therefore, steps (S500) to (S520) can be regarded as adjusting the parameters of the motor controller based on the load torque T without updating the load torque-moment of inertia relationship. L The value of J corresponds to different moments of inertia. m The value of the motor is used to adjust the parameters of the motor controller in order to control the operation of the motor.

[0064] If the model's self-learning is not yet complete, i.e., if the judgment in step (S300) is "yes", then it is also possible to find the relationship by looking up a table or curve fitting formula, based on different load torques T. L The value of the value (e.g., the number of passengers in the elevator) is used to obtain the corresponding moment of inertia J. m The value (S310). Furthermore, it can be determined based on the moment of inertia J. m The value of the load torque-moment of inertia is used to adjust the speed control parameters of the motor controller, i.e., to control the speed of the motor (S320) to control the operation of the motor (S330). Furthermore, the moment of inertia estimation and data collection are continuously performed (S340). Therefore, steps (S310) to (S340) can be considered as a data collection process (load torque information) that still needs to be updated regarding the load torque-moment of inertia relationship. Further, if a new load torque T is found... L Information such as changes in the mass (weight) of new passengers in the elevator can be used to estimate the new moment of inertia J. mTherefore, a new load torque-moment of inertia relationship can be updated, i.e., the judgment in step (S400) is "yes", and the existing load torque-moment of inertia relationship is updated (S240) to make the load torque-moment of inertia relationship more complete. Conversely, if there is no new load torque-moment of inertia relationship to update, i.e., the judgment in step (S400) is "no", then the moment of inertia estimation ends (S250).

[0065] Please see Figure 7 The diagram shown is a block diagram of the motor drive system of the present invention, including the architecture of the hardware and firmware (or software) required for motor drive. The drive system includes outer loop control (i.e., speed control, used to control the motor's rotational speed) and inner loop control (i.e., current control, used to control the motor's output power). In the outer loop control, the speed controller receives the speed command ω from the host computer. m * This means receiving a command to control the elevator speed (corresponding to the motor speed). Combined with position feedback information from the position sensor, the speed calculator calculates the elevator's actual speed (corresponding to the motor's actual speed) and feeds this speed information back to the speed controller. Therefore, based on the speed command ω... m * With speed feedback, a current command, i.e. an equivalent torque command, can be obtained.

[0066] The current controller receives a current command and current feedback from the sensed current measured by the current sensor in the inner loop control circuit (the sensed current is converted into current feedback by a current converter, which can convert the three-phase stationary coordinates abc to the synchronous rotating coordinates dq), and generates a voltage command. The voltage command is modulated by a PWM modulator (pulse width modulator) to generate a gate signal, which then controls the inverter (or frequency converter) to drive the motor.

[0067] In addition, the motor drive system also includes a parameter estimator, which is connected to the outer loop control loop to receive the motor angular velocity ω. m It also connects to the inner loop control circuit to receive the estimated value of the motor output torque. (Based on motor output current (i) q ) and torque constant (K t The product of ()). The parameter estimator calculates the moment of inertia J based on the received motor parameter information. m Value estimation.

[0068] It is worth mentioning that, as mentioned earlier, the information gathering actions required for establishing or updating the load torque-moment of inertia relationship through self-learning do not affect the operation of the outer and inner loop control of the drive system, nor the drive control of the motor. In other words, the parameter estimator only acquires information from the outer and inner loop control to perform calculations on the moment of inertia J.m The estimation of the value did not interfere with the operation of the outer loop control and inner loop control, nor with the drive control of the motor. Furthermore, when the parameter estimator obtains the moment of inertia J... m When the value is , then it can be determined according to the moment of inertia J. m Adjusting the controller parameters to control the motor's operation helps to more accurately control the motor's speed and improves the motor's control performance, including acceleration performance, transient response, and load rejection capability.

[0069] In summary, the present invention has the following features and advantages:

[0070] 1. The value of the corresponding moment of inertia is calculated by integral operation to solve the problem of high-frequency noise caused by the use of differential operation in the existing technology, which requires the addition of a filter.

[0071] 2. By utilizing the existing speed control loop and current control loop architecture, the information required for parameter estimation can be obtained directly. Therefore, the operation of the closed loop and the drive control of the motor are not affected during the parameter (moment of inertia) estimation process.

[0072] 3. When the value of the moment of inertia is obtained by parameter estimation, the parameters of the controller can be adjusted according to the value of the moment of inertia to control the operation of the motor. This helps to control the speed of the motor more accurately and improves the control performance of the motor, including acceleration performance, transient response and load rejection capability.

[0073] The above description is merely a detailed explanation and accompanying drawings of preferred embodiments of the present invention, and the features of the present invention are not limited thereto, nor are they intended to limit the present invention. The full scope of the present invention should be determined by the claims. All embodiments that conform to the concept of the claims of the present invention and similar variations thereof should be included in the scope of the present invention. Any variations or modifications that can be easily conceived by those skilled in the art within the field of the present invention can be covered by the claims disclosed herein.

Claims

1. A speed control method based on a load torque-moment of inertia self-learning model, applied to a controller controlling a motor, the speed control method comprising: (a) Self-learning to establish a load torque-moment of inertia relationship; (b) Obtain the corresponding moment of inertia based on the value of a load torque; and (c) Adjust the parameters of the controller according to the value of the moment of inertia to control the operation of the motor. in, Step (a) includes: (a1) Obtain the value of the load torque under zero speed control; (a2) Under acceleration control, the value of the corresponding moment of inertia is obtained, wherein the viscous friction coefficient is obtained after the acceleration control and the motor operates at a constant angular velocity.

2. The speed control method based on the load torque-moment of inertia self-learning model as described in claim 1, wherein, Step (a) includes: (a3) Repeat steps (a1) and (a2) to establish the load torque-moment of inertia relationship.

3. The speed control method based on a load torque-moment of inertia self-learning model as described in claim 2, wherein, Step (a2) includes: (a21) Calculate the value of the corresponding moment of inertia using integral operations.

4. The speed control method based on a load torque-moment of inertia self-learning model as described in claim 2, wherein, The load torque-moment of inertia relationship is a lookup table.

5. The speed control method based on a load torque-moment of inertia self-learning model as described in claim 2, wherein, The load torque-moment of inertia relationship is a curve fitting equation.

6. The speed control method based on the load torque-moment of inertia self-learning model as described in claim 1 further comprises: (d) Update the load torque-moment of inertia relationship.

7. The speed control method based on a load torque-moment of inertia self-learning model as described in claim 6, wherein, When a new value for the load torque is determined, the corresponding new value for the moment of inertia is obtained, and the load torque-moment of inertia relationship is updated.

8. The speed control method based on the load torque-moment of inertia self-learning model as described in claim 4, wherein, The relationship between the load torque and the moment of inertia is one-to-one.

9. The speed control method based on a load torque-moment of inertia self-learning model as described in claim 4, wherein, When the relationship between the load torque and the moment of inertia is one-to-many, the arithmetic mean of the multiple moments of inertia is calculated to obtain an average moment of inertia, so that the relationship between the load torque and the average moment of inertia is one-to-one.

10. The speed control method based on a load torque-moment of inertia self-learning model as described in claim 1, wherein, In step (a), the speed information of the motor is obtained through a speed control loop, and the torque information of the motor is obtained through a current control loop, so as to establish the load torque-moment of inertia relationship through self-learning.

Images