KR101618722B1 - Motor drive system - Google Patents

Abstract

The present invention relates to a motor drive system. In order to achieve the above object, in the motor drive system according to the embodiment of the present invention, the AC induction motor 10, the torque sensor 70, and the powder brake 30 are connected in series , A DSP reference voltage source inverter 50 based on a DSP 320F2833 is designed and the DSP reference voltage source inverter 50 uses a smart power module FSBB30CH60 (IGBT module) as switching devices In the motor drive system 1, the DSP reference voltage source inverter 50 includes indirect field oriented control (IFOC), direct field oriented control (DFOC), extended first Kalman filter (EKF) Design and control according to the second EKF design.

Description

[0001] Motor drive system [0002]

The present invention relates to a motor drive system, and more particularly, to a motor drive system, and more particularly to a motor drive system, and more particularly to a motor drive system that uses a vector reference control method for an induction motor by indirect field oriented control (IFOC) and direct field oriented control To a motor drive system for simulating and obtaining simulation results.

The induction motor is a nonlinear system and there is a coupling between the magnetic flux of the motor and the electromagnetic torque. However, it is easy to show that if a particular flux-oriented reference frame is used, that is, if the flux-based control is employed, the coupling of the magnetic flux of the induction motor and the electromagnetic torque can be obtained.

In this case, the stator current components represented in the stationary reference frame are the new rotating reference frame that rotates the selected flux-linkage space vector .

There are generally three possibilities for the selection of flux vector profiling vectors, including stator flux-flux profiling, rotor flux profiling, and magnetizing flux-flux profiling vectors. Therefore, terminology such as stator flux reference control, rotor flux reference control and magnetic flux reference control are used.

There are basically two different types of vector control techniques, direct and indirect flux-based controls. In indirect flux-based control, slip estimation is required to calculate the synchronous angular speed at the measured or estimated rotor angular velocity. There is no magnetic flux estimation in the system.

In the direct vector method, the synchronous angular velocity is calculated on the basis of the available magnetic flux angles from the magnetic flux estimation or flux sensors. In contrast to direct methods, indirect methods depend heavily on machine parameters. Many applications use indirect methods.

Because they have relatively simpler hardware and better overall performance at lower frequencies, two vector laws, IFOC and DFOC, are presented to interpret vector control methods for AC induction motors.

That is, in order to control the induction motor, the vector reference control method is used in the control system and the most common vector control method is the field oriented control (FOC), basically there are two types of magnet reference control. The first type is indirect field oriented control (IFOC) and the second type is direct field oriented control (DFOC). The reference frame used for magnet reference control is a rotor-flux-oriented reference frame.

Next, minimizing the number of sensors simplifies installation and reduces control and maintenance costs. As a result, during the last decade, there has been considerable interest in driving induction motors without mechanical sensors. Speed sensorless estimation is possible using Extended Kalman filter (EKF).

EFK is a recursive optimal stochastic state estimator. The model uncertainty and nonlinearities inherent in induction motors are well suited to the statistical nature of EFK.

In this way, on-line estimation of states can be made by instantaneously performing identification of relatively short time interval parameters, and also by directly considering system errors or process errors and measurement noise.

EFK is well known for its high convergence rate, which significantly improves transition performance.

Additionally, the accurate estimation and convergence of the steady state requires high frequency signals that are also implicitly (intrinsically) met by the EFKs with the model noise and measurement noise included in the model.

These properties are the main advantages of EFK over other estimators, and the reason why EFK is widely applied in sensorless estimation despite its computational complexity can be solved by development in high performance process technology.

The purpose of the extended Kalman filter is to obtain states that are measurable and states that are not measurable by statistical noise and measurement (eg, rotor speed).

[Related Technical Literature]

1. Power circuit for three-phase motor drive and power module for three-phase motor drive (Patent Application No. 10-2012-0079408)

2. Driving system of multi-phase DC motor, driving system including multi-phase DC motor, and disk drive (Patent Application No. 10-2000-7009579)

SUMMARY OF THE INVENTION The present invention has been made in order to solve the above problems, and an object of the present invention is to provide a DSP source voltage inverter which is connected in series with an AC induction motor (1.5 Kw), a torque sensor and a powder brake, , And the DSP reference voltage source inverter uses indirect power (IFOC) as a vector reference control method for the induction motor by using a smart power module (FSBB30CH60) (IGBT module) as switching devices and And to provide a motor drive system for simulating by direct field oriented control (DFOC) to obtain simulation results.

In addition, the present invention can be applied to a first extended Kalman filter design (first EKF design) and a second extended Kalman filter design (second EKF design) as well as IFOC and DFOC having a 150 MHz system clock as a DSP reference voltage source inverter. To obtain a simulation result by the motor drive system.

The present invention also provides a motor drive system for comparing the induction motor control performance according to the second EKF design and the IFOC and DFOC method according to the simulation and the first EKF design.

However, the objects of the present invention are not limited to the above-mentioned objects, and other objects not mentioned can be clearly understood by those skilled in the art from the following description.

In order to achieve the above object, in the motor drive system according to the embodiment of the present invention, the AC induction motor 10, the torque sensor 70 and the powder brake 30 are connected in series, and the DSP 320F2833 A DSP based voltage source inverter 50 is designed and the DSP reference voltage source inverter 50 is a motor drive system using a smart power module FSBB30CH60 (IGBT module) as switching devices 1, the DSP reference voltage source inverter 50 includes an indirect field oriented control (IFOC), a direct field oriented control (DFOC) for the AC induction motor 10, a first EKF (Extended Kalman filter) Perform control according to EKF design.

At this time, the present invention includes a torque transducer 40 and an encoder 20, which are used for verification of the load torque and speed estimation for the AC induction motor 10 by the DSP reference voltage source inverter 50, ; .

The DSP reference voltage source inverter 50 is a plugged type connected to the socket of the controller board 51 as a first DSP module and a controller board 51 as a second DSP chip, And a controller board 51 and an IGBT driver 52 are preferably provided.

In addition, the DSP reference voltage source inverter 50 samples the analog signal and converts the sampled analog signal into a bit form that can be processed by the controller board (DSP) 51, An analog-to-digital converter (ADC) module having a current sensor 53 for measuring a voltage; .

Further, the current sensor 53 has a supply voltage of ± 15 V as an output voltage and an output voltage of ± 4 V, and due to the analog inputs to the controller board (DSP) 51 being in the range from 0 V to 3 V The first OP-Amp, which is connected as an inverting amplifier after following the signal at the input to reduce the magnitude of the input signals, has two resistors, 1.5KΩ and 12KΩ, to reduce the input voltage of ± 4V to ± 0.5V. A second OP-AM with an input of 1 V to 2 V as an additive amplifier for transferring the input voltage to 1.5 V; And a third OP-AMP which is a voltage following amplifier with two capacitors C ₁ , C ₂ for filtering; .

에 의해 선택되는 것이 바람직하다. In addition, the noise covariance matrices in the first EKF design are:

. ≪ / RTI >

에 의해 선택되는 것이 바람직하다.In addition, the noise covariance matrices in the second EKF design are:

. ≪ / RTI >

Also, it is preferable that the AC induction motor 10 is controlled at a low speed of 20RPM and a high speed of 500RPM in two modes, light load mode and heavy load mode, respectively.

A motor drive system according to an embodiment of the present invention includes an AC induction motor (1.5 Kw), a torque sensor and a powder brake connected in series, a DSP source voltage inverter (DSP) based on a DSP 320F2833, , And the DSP reference voltage source inverter uses indirect power (IFOC) as a vector reference control method for the induction motor by using a smart power module (FSBB30CH60) (IGBT module) as switching devices and It is possible to obtain simulation results by simulating by direct field oriented control (DFOC).

In addition, the motor drive system according to another embodiment of the present invention includes a first extended Kalman filter (first EKF design) and a second extended Kalman filter (second EKF design) as well as IFOC and DFOC having a 150 MHz system clock as a DSP reference voltage source inverter. It is possible to obtain the simulation result by the extended Kalman filter design (the second EKF design).

In addition, the motor drive system according to another embodiment of the present invention can compare the IFOC and DFOC method according to the simulation, and the induction motor control performance according to the first EKF design and the second EKF design.

Fig. 1 is a block diagram showing a motor drive system 1 according to an embodiment of the present invention, and is a view showing the appearance of a motor drive system 1 designed with a simulation device. Fig.
Fig. 2 shows the torque curve of the powder brake 30 in the motor drive system 1 of Fig. 1 with respect to current.
Fig. 3 shows a wire connection of the torque sensor 70 of the motor drive system 1 of Fig.
4 is a diagram showing a DSP source voltage inverter 50 implemented by a voltage source inverter (VSI) based on DSP320F28335 of the motor drive system 1 of FIG.
FIG. 5 is a diagram showing two versions of the controller board 51 among the DSP source voltage inverter 50 (FIG. 4).
6 is a circuit diagram of a current sensor circuit using a current sensor circuit (OP-AMP) 53 in the DSP reference voltage source inverter 50 of FIG.
7 shows a circuit of a voltage follower for obtaining a target voltage V_refs in a circuit for a current sensor circuit using a current sensor 53 (U-phase) in Fig. 6 to be.
8 is a diagram showing a differential amplifier circuit for measuring a DC voltage signal in the DSP reference voltage source inverter 50 of FIG.
Figs. 9A, 9B, 10A and 10B are diagrams for explaining the case where the IFOC method (a) and the DFOC method (b) in the light load mode and the heavy load mode, which are two modes simulated in the motor drive system 1 of Fig. A graph showing the induction motor speed and induction motor speed error.
11a shows the electromagnetic torque in the light load mode using the IFOC method (a) and the DFOC method (b) simulated in the motor drive system 1 of Fig. 1, Fig. 11b shows the electromagnetic torque in the light load mode using the IFOC method (a) Is a graph showing electromagnetic torques in a heavy load mode using a constant-current mode.
12A is a phase current in a light load mode using the IFOC method (a) and the DFOC method (b) simulated in the motor drive system 1 of FIG. 1, FIG. 12B is a phase current in the light load mode using the IFOC method (a) Is a graph showing the phase currents in the heavy load mode using the constant current mode.
13A shows phase voltages in light load mode using the IFOC method (a) and the DFOC method (b) simulated in the motor drive system 1 of FIG. 1, FIG. 13B shows phase voltages in the light load mode using the IFOC method (a) And the phase voltage in the high-load mode using the high-voltage mode.
14A and 14B are graphs showing induction motor speed estimates using the first EKF design and the second EKF design, respectively, in the motor drive system 1 of FIG.
15A and 15B are graphs showing induced motor speed error estimates using the first EKF design and the second EKF design, respectively, in the motor drive system 1 of FIG.
16A and 16B are graphs showing the phase current estimation using the first EKF design and the second EKF design in the motor drive system 1 of Fig. 1, respectively.
17A and 17B are graphs showing the phase current estimation error using the first EKF design and the second EKF design in the motor drive system 1 of FIG.
18 to 24 are reference graphs for explaining the first EFK design and the second EFK design used in the motor drive system 1 according to the embodiment of the present invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a detailed description of preferred embodiments of the present invention will be given with reference to the accompanying drawings. In the following description of the present invention, detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear.

Fig. 1 is a block diagram showing a motor drive system 1 according to an embodiment of the present invention, and is a view showing the appearance of a motor drive system 1 designed with a simulation device. Fig. 1A and 1B, a motor drive system 1 includes an AC induction motor 10, an encoder 20, a powder brake 30, a torque transducer 40, A DSP reference voltage source inverter 50, two couplings 60 and a torque sensor 70.

The AC induction motor 10 used in the motor drive system 1 has the specifications and parameters given in the following [Table 1] and [Table 2] and has three phases, four poles and 2HP / 1.5Kw.

Description Value Unit Power 1.5 Kw Voltage 220 V Peak Current 6.0 A Base Frequency 60 Hz Base speed 1750 rpm Nominal torque 10 Nm Nominal stator flux 0.49 Wb

Symbol Symbol Description Value Unit Zp Number of pole pairs 2 - J Moment of inertia 0.089 Kgm ² Rs Stator resistance 3.285 Ω Rr Rotor resistance 2.715 Ω Ls Stator inductance 0.387117 H Lr Rotor inductance 0.387117 H Lm Mutual inductance 0.374 H

The powder brake 30 operates as a load and can be controlled by adjustment of the voltage from 0 V to 24 V DC. Figure 2 shows the torque curve of the powder brake 30 versus current.

A torque transducer 40 with a rated 10 Kgfm and an encoder 20 with 1024 counts / rev are used for the verification of the load torque and speed estimation. 3 shows a wire connection of the torque sensor 70. Fig. The excitation voltage is preferably set to 10V.

The DSP reference voltage source inverter 50 can be implemented as a voltage source inverter (VSI) based on the DSP 320F28335 as shown in FIG. 4, and the DSP reference voltage source inverter 50 (voltage source inverter) As shown in the figure, the controller board 51 and the IGBT driver 52 are provided in two parts.

In the present invention, the controller board 51 is designed based on DSP 320F28335 of Texas Instrument, and the IGBT driver 52 is designed based on Fairchild's smart power module FSBB30CH60.

Meanwhile, FIG. 5 shows two versions of the DSP reference voltage source inverter 50 used in the present invention. The DSP chip (Fig. 5A) in the first version is plugged into the socket of the controller board 51 and the DSP chip Chip (Fig. 5B) in the second version is connected to the controller board 51 (Solder type).

DSP 320F28335, implemented as a DSP voltage source inverter 50, DSP includes a 32-bit CPU that allows floating-point calculations to be performed in hardware, a floating-point ) Calculation has a single-precision 32-bit floating-point unit (FPU) that can be implemented in hardware. Furthermore, the CPU of the DSP 320F28335 has an 8-stage pipeline structure. This structure allows the CPU to simultaneously execute eight instructions within a system clock cycle. The 150Mhz system clock (150Mhz system clock) is provided by an on-chip oscillator and a phase-locked loop circuit (PPL).

The DSP 320F28335's physical memory consists of a 34Kx16 single-access random access memory (SARAM), a 56Kx16 flash memory, an 8Kx16 read-only memory (ROM), a 1Kx16 OTP memory (OTP memory) . The DSP 320F28335 also has a direct memory access (DAM) architecture. With the DMA bus, data can be transferred from one part of the DSP to another without CPU interactions increasing the data transfer rate. Because the DSP 320F28335 is designed primarily for industrial applications, it has many peripheral circuits.

For example, 16-channel (16-chanels), 12-bit ADC modules, PWM modules and encoder modules can be used for motor control purposes.

The five types of information communication are the controller area network (CAN) module, the serial communication interface module (SCI module), the serial peripheral interface (SPI), the multichannel buffered serial port (McBSP) (inter-integrated circuit module).

96 interrupts are supported by the DSP 320F28335. These interrupts enable or disable several interrupts, determine interrupts' priorities of interrupts, and provide a peripheral interrupt expansion (PIE) to inform the CPU of the occurrence of a new interrupt. Lt; / RTI >

Next, the analog input voltage on the DSP reference voltage source inverter 50 will be described.

An analog-to-digital converter (ADC) module samples an analog signal and converts it into a bit form that can be processed by a controller board (DSP) Since the ADC module is used for measuring the three-phase current and the DC voltage, it is preferable to include the current sensor 53.

The accuracy of this measurement is very important for the performance of the entire vector control systems, and only the currents in the actual phase are required for the calculation of the control algorithm. 6 shows a circuit for a current sensor circuit (current sensor circuit using OP-AMP) 53 in the DSP reference voltage source inverter 50, and other sensors have a similar structure.

The current sensor 53 has a supply voltage of ± 15V and an output voltage of ± 4V while the analog inputs to the controller board (DSP) 51 are in the range of 0V to 3V.

Due to this fact, the input signals are reduced in size and need to be biased with the OP-Amp, as shown in FIG. 6, the first OP-Amp is connected as an inverting amplifier after following the signal at the input.

Two resistors, 1.5KΩ and 12KΩ, reduce the ± 4V input voltage to ± 0.5V. The second OP-AMP is an adder that moves the input voltage to 1.5V. The output of OP-AMP is 1V-2V. The third OP-AMP is a voltage following amplifier with two capacitors C ₁ and C ₂ for filtering. The target voltage V_refs in Fig. 6 can be obtained by using the voltage follower in Fig.

On the other hand, the DC voltage signal can be measured by the ADC module of the DSP by using the differential amplifier circuit shown in Fig.

Next, let's look at the experimental results.

Experiments were performed using an indirect field oriented control (IFOC) via a voltage source inverter (DSP) 50, direct field oriented control (DFOC), a first extended Kalman filter filter design (hereinafter referred to as a first EKF design) and a second extended Kalman filter design (hereinafter referred to as a second EKF design).

In the first EKF design, the basic motor axis angular velocity estimation is performed under the assumption that all the parameters of the AC induction motor 10 are well known. In the second EKF design, not only the motor shaft angular velocity but also the stator resistance and the load torque are also estimated.

The first EFK design has good results if all AC induction motor (10) parameters are well known. For example, if several parameters are not well known, the stator resistance is set to 70% of the actual value, and the estimation of the motor axis angular velocity in the first EKF design becomes poor. However, rotor flux angle estimation is not affected.

When the second EFK design is applied, both motor shaft angular velocity estimation and rotor flux angle estimation are not affected despite stator resistance variations.

The estimated stator resistance converges to the actual value after 1.8 s and the first and second EKF designs are assumed to work well at both the low speed and the high speed of the induction motor 10.

In short, in the present invention, the EFK design method is considered in both cases. A fundamental rotor speed estimation using the first EFK design is performed, and the parameters of the AC induction motor 10 are well known. However, accurate knowledge of the stator resistance is of utmost importance in the correct operation of the speedless sensor induction motor control in the low speed region.

Since stator resistance inevitably varies with operating conditions, stable and accurate operation at nearly zero speed requires an appropriate on-line identification algorithm for stator resistance.

To estimate the stator resistance in speedless sensor control of induction motors, a second EFK design according to another design of the extended Kalman filter is proposed.

On the other hand, the basic rotor speed estimation using the extended Kalman filter (EKF) is one of the major design steps for the speedless sensor induction motor drive implementation using the discrete extended Kalman algorithm. The time-domain induction machine model, 2. discretization of noise and state covariance matrices Q, R and p, 3. implementation of a discrete extended Kalman filter algorithm, and 4. tuning of covariance matrices.

Table 3 below shows the computation times for IFOC, DFOC, first EKF design and second EKF design with DSP S320F28335 having a 150Mhz system clock.

IFOC DFOC First EKP design
(The first EKF) Design of 2nd EKP
(The second EKF) Comp.time (cycle) 1,097 3,801 5,022 11,433 Comp.time (ms) 0.00731 0.02534 0.03348 0.07622

For the realization of the proposed algorithms, all calculations must be performed within the sampling interval. The sampling period of the system is selected as 0.1 ms, and the switching frequency of the PWM inverter is 10 kHz.

AC induction motor 10 is controlled using IFOC and DFOC in the following two modes of light load mode and heavy load mode.

In the light load mode, the powder brake 30 does not operate. The friction and moment of inertia of the system mechanism generate a light load of about 10 Nm. In the heavy load mode, a powder brake 30 is operated and generates an external load of about 10 Nm.

FIGS. 9A, 9B, 10A and 10B show the induction motor speed and induction motor speed error using the IFOC method (a) and the DFOC method (b) in the light load mode and the heavy load mode, which are two modes.

9A and 9B, the dotted line is the target speed and the solid line is the actual speed. The operation of the AC induction motor 10 is considered at a high speed of 500 RPM and a low speed of 20 RPM.

FIGS. 10A and 10B show the induced motor speed error for the IFOC method and the DFOC method in the light load mode and the heavy load mode, and the performances have almost the same values as shown in the figure. The actual induction motor speed tracks the target speed well at both high and low speeds.

Target speed error goes to 0 after 5s. In steady state conditions for both the IFOC and DFOC methods, the velocity error is ramped to ± 3 RPM. However, in the low speed and high load modes, the speed error of the IFOC is increased to ± 8 RPM.

11A shows the electromagnetic torque in the light load mode using the IFOC method (a) and the DFOC method (b). 11B shows the electromagnetic torque in the heavy load mode using the IFOC method (a) and the DFOC method (b). The electromagnetic torque at low speeds is greater than the electromagnetic torque at high speeds due to viscous friction.

 The electromagnetic torque in light load mode is about 1 Nm at high speed and about 1.6 Nm at low speed. The electromagnetic torque in the heavy load mode is about 8.5 Nm at both high speed and low speed.

12A shows the phase current in the light load mode using the IFOC method (a) and the DFOC method (b). In the IFOC method, the magnitude of the phase current is about 4.4 A at high speed and 1.3 A at low speed. In the DFOC method, the magnitude of the phase current is about 0.4 A at both high and low speeds. 12B shows the phase current in the heavy load mode using the IFOC method (a) and the DFOC method (b). The magnitudes of the phase currents are similar in both the light load mode and the heavy load mode.

In the IFOC method, the magnitude of the phase current is about 3.2 A at both high and low speeds. In the DFOC method, the magnitude of the phase current is about 4.2 A at both high and low speeds.

13A shows the phase voltage in the light load mode using the IFOC method (a) and the DFOC method (b). In the IFOC method, the magnitude of the phase voltage is about 70 V at high speed and about 30 V at low speed. In the DFOC method, the magnitude of the phase voltage is about 50 V at high speed and about 20 V at low speed. 13B shows the phase voltage in the heavy load mode using the IFOC method (a) and the DFOC method (b). In the IFOC method, the magnitude of the phase voltage is about 190 V at high speed and about 80 V at low speed. In the DFOC method, the magnitude of the phase voltage is about 180 V at high speed and about 100 V at low speed.

Both models of the EFK design are performed in simulations for the results of Figs. 9-13.

The first EKF design is used to estimate the motor shaft angular velocity. The initialization for the first EFK design is given as follows.

That is, the initial value of the estimated state vector is expressed by the following equation (1).

The initial value of the estimated error covariance is expressed by the following equation (2).

The noise covariance matrices in the first EKF design are selected as shown in Equation (3) below.

The covariance matrix Q of the system noise vector is a 5 × 5 matrix, and the covariance matrix R of the measurement noise vector is a 2 × 2 matrix. Q and P are diagonal matrices and only five elements in Q and two elements in R need to be known.

On the other hand, the second EFK design is used to instantaneously estimate the stator resistance, load torque and motor shaft angular velocity. The initialization for the second EFK design is given by Equation (4) below.

The initialization of the estimated state vector is expressed by the following equation (5).

Then, the noise covariance matrices in the second EKF design are selected as shown in Equation (6) below.

14A and 14B show induction motor speed estimates using a first EKF design and a second EKF design, respectively. The dotted line is the actual speed value and the solid line is the estimated speed value.

15A and 15B show induction motor velocity error estimates using the first EKF design and the second EKF design, respectively. From the results in Figs. 14-15, the estimated induction motor speed in the second EKF design tracks the actual induction motor speeds better than the estimated induction motor speed in the second EKF design.

In the first EKF design, the estimation error is about 5 RPM at high speed and 3.5 RPM at low speed. In the second EKF design, the estimation error is about 3.5 RPM at high speed and 2 RPM at low speed.

16A and 16B show phase current estimates using the first EKF design and the second EKF design, respectively.

17A and 17B show phase current estimation errors using the first EKF design and the second EKF design, respectively. The estimation error tracks the actual value well. Estimation error is within ± 0.5A.

1 to 17, the AC induction motor 10 (1.5 Kw), the torque sensor 70 and the powder brake 30 are connected in series, and the DSP 320F2833 A DSP based voltage source inverter 50 (a voltage source inverter) is designed. The DSP reference voltage source inverter 50 uses a smart power module (FSBB30CH60) (IGBT module) as switching devices.

Experiments are performed on IFOC, DFOC, 1st EKF design and 2nd EKF design. From the experimental results, the conclusions are as follows.

Experimentally, the controller board (DSP) 51 of the proposed DSP reference voltage source inverter 50 using IFOC and DFOC obtained very good results. The AC induction motor 10 is controlled at a low speed of 20RPM and a high speed of 500RPM in two modes of light load mode and heavy load mode, respectively.

The induction motor speed error converges to zero after 5 s. In the steady state in both IFOC and DFOC, the velocity error is ramped within ± 3 RPM. However, in low speed and heavy duty modes, the speed error of the IFOC is increased to ± 8 RMP.

The IFOC is simpler than the DFOC and the computation time of the IFOC is three times less than the computation time of the DFOC.

18 to 24 are reference graphs for explaining the first EFK design and the second EFK design described above. More specifically, Figures 18-21 are for the first EFK design and Figures 22-24 are for the second EFK design.

18A shows the motor axis angular velocity estimate using the first EFK design. The dotted line is the actual value and the solid line is the estimated value. 18B shows the angular velocity estimation error using the first EFK design. From 0 s to 0.2 s, the error is large and changes very quickly. After 0.2 s, the error is small and converges to zero at steady state. In a transition from 1 s to 1.25 s, the error is about 0.25 rad / s.

In Figures 18b and 19a, the estimate tracks well the actual value.

Figure 19A shows estimates of the rotor flux angle using the first EFK design. The dotted line is the actual value and the solid line is the estimated value. FIG. 19B shows a rotor flux angle estimation error. From 0 s to 0.5 s, due to initial conditions, the error is large. From 0.5s to 1s, the load torque is large, so the error is a bit large. After 1.5 s, the error is reduced. In the steady state from 0.4 s to 1 s, the error is about 0.1 rad. However, the error is rapidly reduced in transition from 1 s to 1.25 s.

Figures 20a and 20b show the motor shaft angular velocity and angular velocity estimation error using the first EFK design when the stator resistance is not well known.

The stator resistance is equal to 70% of the actual value used in the EFK design. From the results in Figs. 20A and 20B, the error is too large. The error at the start time is -5 rad / s, the error at 1 s is 2 rad / s and does not go to 0 in steady state.

FIGS. 21A and 21B show rotor angular estimation and rotor flux angular error, assuming that the stator resistance is not well known. The stator resistance is equal to 70% of the actual value used in the EFK design.

Next, FIGS. 22A and 22B show the estimation of the angular velocity of the motor shaft and the angular velocity estimation error using the second EFK design under the condition that the stator resistance is not well known. The stator resistance used in the EFK design is equal to 70% of the actual value. Despite the unknown stator resistance, the estimate tracks the actual value well. From 0 s to 0.2 s, the error is large and changes very quickly. The error is small after 0.2 s and converges to zero in steady state.

FIGS. 23A and 23B show the rotor flux angle estimation and the rotor flux angle error on the assumption that the stator resistance is not well known. The stator resistance used in the EFK design is equal to 70% of its actual value. From the results of Figs. 21A and 21B, rotor flux angle estimation tracks the substance values well.

24A shows stator resistance estimation using the second EFK design. The initial value of the stator resistance is equal to 70% of its actual value. After 0.2 s, the deviation between the estimated value and the actual value becomes small. After 1.4 s, the deviation between the estimated value and the actual value becomes almost zero.

Figure 24B shows the load torque estimation using the second EFK design. The stator resistance used in the EFK design is equal to 70% of its actual value. The dotted line is the actual value and the solid line is the estimated value. From Fig. 19B, the estimated value tracks the target value well. The estimated value and the actual value become 0 in the steady state.

As described above, preferred embodiments of the present invention have been disclosed in the present specification and drawings, and although specific terms have been used, they have been used only in a general sense to easily describe the technical contents of the present invention and to facilitate understanding of the invention , And are not intended to limit the scope of the present invention. It is to be understood by those skilled in the art that other modifications based on the technical idea of the present invention are possible in addition to the embodiments disclosed herein.

1: Motor drive system
10: AC induction motor
20: Encoder
30: Powder brake
40: torque transducer
50: DSP reference voltage source inverter
51: Controller board (DSP)
52: IGBT driver
53: Current sensor
60: Coupling
70: Torque sensor

Claims (8)

An AC induction motor 10 controlled at a low speed of 20RPM and a high speed of 500RPM in two modes of a light load mode and a heavy load mode;
An encoder 20 connected in series through the AC induction motor 10 and the coupling 60 and used for verification of the load torque and speed estimation for the AC induction motor 10 by the DSP reference voltage source inverter 50, Wow;
A torque sensor 70 installed in the encoder 20;
A powder brake 30 operated as a load and controlled by voltage regulation from 0 V DC to 24 V and connected in series through the encoder 20 and the coupling 60;
Based on the DSP 320F2833, a smart power module (FSBB30CH60) (IGBT module) is used as the switching devices, and indirect field oriented control (IFOC), direct field oriented control (DFOC) A plugged type which is connected to a socket of the controller board 51 as a first DSP module as a first version, and a plugged type which is a second version And is formed as a solder type formed by soldering directly on the controller board 51 as a DSP chip and has a controller board 51 and an IGBT driver 52 while sampling an analog signal, An analog-to-digital converter (ADC) module having a current sensor 53 for converting a three-phase current and a DC voltage into a bit form that can be processed by a board (DSP) 51 , The current sensor 53 outputs an output voltage The analog input to the controller board (DSP) 51 has a supply voltage of ± 15 V and an output voltage of ± 4 V and follows the signal at the input so that the magnitude of the input signals decreases due to the range of 0 V to 3 V The first OP-Amp, which is then connected as an inverting amplifier and allows the two resistors, 1.5KΩ and 12KΩ, to reduce the ± 4V input voltage to ± 0.5V. A second OP-AM with an input of 1 V to 2 V as an additive amplifier for transferring the input voltage to 1.5 V; And a third OP-AMP which is a voltage following amplifier with two capacitors C ₁ , C ₂ for filtering; A voltage source inverter 50,
A torque transducer (40) used for verification of the load torque and speed estimation for the AC induction motor (10) by the DSP reference voltage source inverter (50)
Basic rotor speed estimation using an extended Kalman filter (EKF) is a key design step for implementing a speedless sensor induction motor drive using a discrete extended Kalman algorithm. The time-domain induction machine model ), 2. discretization of noise and state covariance matrices Q, R and p, 3. implementation of a discrete extended Kalman filter algorithm, and 4. tuning of covariance matrices,
The first EKF (Extended Kalman filter) design, in which the basic motor axis angular velocity estimation is performed under the assumption that all the parameters of the AC induction motor 10 are well known, is obtained by multiplying the initial value of the estimated state vector by

Lt; / RTI &
The initial value of the estimated error covariance is

, And the noise covariance matrices in the first EKF design are

(The covariance matrix Q of the system noise vector is a 5 × 5 matrix, the covariance matrix R of the measurement noise vector is a 2 × 2 matrix, Q and P are diagonal matrices, only 5 elements in Q and 2 elements in R has exist)
, ≪ / RTI >
The second EKF design, in which the stator resistance and the load torque are estimated as well as the motor shaft angular velocity, is initialized for the second EFK design

, And the initialization of the estimated state vector is

, And the noise covariance matrices in the second EKF design are

Is selected and executed. delete delete delete delete delete delete delete

Images