KR101218441B1 - Control System of Interior Permanent Magnet Synchronous Motor and Method to Detect Sensor Fault thereof - Google Patents

Abstract

Sensor failure detection method of the control system of the embedded permanent magnet synchronous motor according to an embodiment of the present invention, the modeling step of modeling the model of the embedded permanent magnet synchronous motor; A linear model transformation step of converting the nonlinear model of the embedded permanent magnet synchronous motor into a linearized model by input-output linearization; And an error measuring step of measuring a residual from the linearized model by applying a parity equation. The current sensor and the position sensor of the system using a failure detection algorithm generated based on the steps. The fault can be detected in real time.

Description

Embedded System Permanent Magnet Synchronous Motor and Method to Detect Sensor Fault

An embedded permanent magnet synchronous motor control system and a sensor failure detection method thereof are disclosed. More specifically, a buried permanent magnet synchronous motor control system and a sensor failure detection method thereof, which can accurately detect not only a failure of a current sensor of a buried permanent magnet synchronous motor but also a failure of a position sensor by a designed algorithm, are disclosed.

Embedded Permanent Magnet Synchronous Motors (IPMSMs) have a higher power factor, higher power density, higher torque to current ratio, higher efficiency and greater force to weight ratio than other AC motors. This advantage is essential for automotive applications such as automotive motor systems and hybrid vehicles.

However, the magnetic saturation of the rotor core and the nonlinear coupling between the winding current and the rotor speed make it difficult to design the controller of the motor. Therefore, motor controllers are very different from linear motors such as PI and PID controllers, and some nonlinear controllers have been developed for the application of synchronous motors.

For automotive applications, the motor control system must have the same level of reliability of the machine part as a vehicle breakdown can threaten the safety of the passenger. In order to secure the reliability, various studies on the failure diagnosis algorithm of the motor control system have been conducted.

Most of the research uses a fault detection approach based on observers. One of the advantages of observer-based fault detection is that stiffness to model uncertainty can be easily achieved. However, this method carries the risk that failures with slow rate constants may not be detected because the improvement in stiffness may reduce the sensitivity of detection to slow rate constants.

On the other hand, a parity equation and parameter evaluation approach based on a linear model is a simpler and more direct solution for detecting failure of the control system. However, applying this approach to a motor control system without any correction has the potential for false alarms since the linear model has a large amount of model uncertainty when the controller is in a nonlinear operating state.

In order to solve this problem, it is essential to apply a method based on a nonlinear model for the motor control system.

An object according to an embodiment of the present invention, the embedded permanent magnet synchronous motor control system and its sensor that can accurately detect the failure of the current sensor as well as the position sensor of the embedded permanent magnet synchronous motor in real time by the designed algorithm It is to provide a fault detection method.

Another object according to an embodiment of the present invention, an embedded permanent magnet synchronous motor control system that can detect the failure of the current sensor and the position sensor by the algorithm to reduce the cost of detecting the failure of the sensor compared to the conventional And a sensor failure detection method thereof.

Sensor failure detection method of the control system of the embedded permanent magnet synchronous motor according to an embodiment of the present invention, the modeling step of modeling the model of the embedded permanent magnet synchronous motor; A linear model transformation step of converting a nonlinear model of the embedded permanent magnet synchronous motor into a linearized model by input-output linearization; And an error measuring step of measuring a residual from the linearized model by applying a parity equation. The current sensor of the system using a failure detection algorithm generated based on the steps. And a failure of the position sensor can be detected in real time.

The sensor failure detection method may further comprise an algorithm verification step of verifying the usefulness of the proposed failure detection algorithm by computer simulation.

The computer simulation in the algorithm verification step may be executed by a program including MATLAB or SIMULINK.

및

(여기서,

는 d-축 전압, R은 고정자 저항,

는 d-축 전류,

는 d-축 인덕턴스,

는 모터의 폴 페어(pole pair),

은 모터 속도,

는 q-축 인덕턴스,

는 q-축 전류,

는 q-축 전압,

은 쇄교자속(flux linkage)임)일 수 있다.The model in the modeling step is defined by direct-quadrature (dq) transformation, and the voltage equation within the defined dq sync frame is

And

(here,

Is the d-axis voltage, R is the stator resistance,

D-axis current,

D-axis inductance,

Is the pole pair of the motor,

Silver motor speed,

Is the q-axis inductance,

Is the q-axis current,

Is the q-axis voltage,

May be flux linkage).

(여기서,

는 a축 전류,

는 b축 전류,

은 회전자 위치임)일 수 있다.The current equation in the dq sync frame defined in the modeling step is

(here,

Is the a-axis current,

Is the b-axis current,

Is the rotor position.

In the linear model conversion step, the linearized model may be obtained by applying an equation defining the nonlinear model in the modeling step to a Lie derivative.

The input-output linearization provides a nonlinear equivalent variation that linearizes the dynamic equations of the system, thereby applying the input-output linearization to account for the nonlinear characteristics of the embedded permanent magnet synchronous motor. A model to reflect can be obtained.

On the other hand, the control system of the embedded permanent magnet synchronous motor according to an embodiment of the present invention, after modeling the model of the embedded permanent magnet synchronous motor, the embedding by input-output linearization (input-output linearization) A failure detection algorithm is generated by converting a nonlinear model of the type permanent magnet synchronous motor into a linearized model, and then measuring a residual from the linearized model by applying a parity equation, and generating the failure. Detection algorithms can be used to detect faults in the current and position sensors of the system in real time.

According to the embodiment of the present invention, not only the failure of the current sensor of the embedded permanent magnet synchronous motor but also the failure of the position sensor can be accurately detected in real time by a designed algorithm.

According to the embodiment of the present invention, the failure of the current sensor and the position sensor can be detected by an algorithm, so that the cost of detecting the failure of the sensor can be reduced as compared with the related art.

1 is a block diagram schematically illustrating a configuration of an embedded permanent magnet synchronous motor according to an embodiment of the present invention.
2 is a flowchart illustrating a sensor failure detection method of a control system of an embedded permanent magnet synchronous motor according to an exemplary embodiment of the present invention.
3 to 5 are graphs showing simulation results of a control system of an embedded permanent magnet synchronous motor according to an embodiment of the present invention.

Hereinafter, configurations and applications according to embodiments of the present invention will be described in detail with reference to the accompanying drawings. The following description is one of several aspects of the patentable invention and the following description forms part of the detailed description of the invention.

In the following description, well-known functions or constructions are not described in detail for the sake of clarity and conciseness.

1 is a block diagram schematically showing the configuration of a buried permanent magnet synchronous motor according to an embodiment of the present invention, Figure 2 is a control system of a buried permanent magnet synchronous motor according to an embodiment of the present invention 3 to 5 are graphs illustrating simulation results of a control system of an embedded permanent magnet synchronous motor according to an exemplary embodiment of the present invention.

Referring to FIG. 1, the embedded permanent magnet synchronous motor 100 according to an exemplary embodiment of the present disclosure may convert the current value and the position detection information obtained by the encoder 120 to the failure detection algorithm 101. By substituting, the IPMSM controller 110 can detect the failure of the current sensor and the position sensor provided in the control system in real time. In other words, on the basis of the algorithm 101, not only the failure of the current sensor but also the failure of the position sensor can be detected in real time, thereby improving efficiency and accuracy and reducing costs.

2, the sensor failure detection method of the embedded permanent magnet synchronous motor control system according to an exemplary embodiment of the present disclosure includes a modeling step of modeling a nonlinear model of the embedded permanent magnet synchronous motor 100. (S100), a linear model conversion step (S200) for converting a nonlinear model of the embedded permanent magnet synchronous motor 100 into a linearized model by input-output linearization, and a parity equation (Parity) an error measurement step (S300) of measuring a residual from the linearized model by applying an equation, and an algorithm proof step of verifying the usefulness of the failure detection algorithm 101 proposed by the above-described steps by computer simulation. It may include (S400).

For each step, first, in the modeling step (S100), assuming that the embedded permanent magnet synchronous motor 100 has balanced windings and three phases without saturation, direct quadrature (dq, direct-quadrature) synchronization The voltage equation of the synchronous motor in the frame can be expressed as follows.

......(식 1)

(Equation 1)

......(식 2)

(Equation 2)

는 d-축 전압, R은 고정자 저항,

는 d-축 전류,

는 d-축 인덕턴스,

는 모터의 폴 페어(pole pair),

은 모터 속도,

는 q-축 인덕턴스,

는 q-축 전류,

는 q-축 전압,

은 쇄교자속(flux linkage)이다. In equation 1 and equation 2,

Is the d-axis voltage, R is the stator resistance,

D-axis current,

D-axis inductance,

Is the pole pair of the motor,

Silver motor speed,

Is the q-axis inductance,

Is the q-axis current,

Is the q-axis voltage,

Is the flux linkage.

The electromagnetic torque can be expressed as follows.

......(식 3)

(Equation 3)

On the other hand, the dynamic spinning equation of the speed is as follows.

......(식 4)

(Equation 4)

은 알려지지 않은 외부 하중이고,

,

은 모터의 표준 파라미터이다. In equation 4,

Is an unknown external load,

,

Is the standard parameter of the motor.

The dynamic equation of position is given by

......(식 5)

(Eq. 5)

, b-축 전류인

는 전류 센서에 의해 측정될 수 있고, 회전자 위치인

는 위치 센서에 의해 측정될 수 있다. 이때,

,

로부터

및

를 발생시키기 위해 직접 직교 변환(d-q, direct-quadrature transformation)이 적용될 수 있다. 직접 직교 변환은 다음의 식 6과 같이 정의될 수 있다.Where a-axis current

, b-axis current

Can be measured by a current sensor, and the rotor position

Can be measured by the position sensor. At this time,

,

from

And

Direct quadrature transformation (dq) can be applied to generate Direct orthogonal transformation may be defined as in Equation 6 below.

......(식 6)

(Equation 6)

On the other hand, as described above, after the modeling step (S100), a linear model conversion step (S200) reflecting the non-linear characteristics of the embedded permanent magnet synchronous motor 100 proceeds, where the input output linearization (input-output linearization) Can be made.

In the linear model transformation step (S200), a linear system model can be obtained by applying a nonlinear system model to, for example, a Lie derivative equation. Lee's differential equation is given by the following equation.

......(식 7)

(Eq. 7)

In order to obtain a linearized model here, a new state variable may be defined as in Equation 8 below.

......(식 8)

(Eq. 8)

이고,In Equation 8

ego,

이며,

이고,

Is,

ego,

이며,

Is,

이고,

이다.

ego,

to be.

로 알려졌다고 가정하고, 새로운 상태 변수를 시간에 대해 미분한 후 전술한 식 1 내지 식 5에 적용시키면, 센서 고장 신호(

)를 포함한 시스템의 상태 공간 형성은 다음의 식 9로부터 얻을 수 있다.The system changes slowly and the current state of the system

Suppose it is known that the new state variable is differentiated with respect to time and then applied to equations 1 to 5 as described above.

The state space formation of the system including) can be obtained from Equation 9 below.

......(식 9)

(Eq. 9)

here,

이고,

ego,

,

,

,

,

,

,

이다.

to be.

From these equations, we can obtain a linear system model that reflects the nonlinear part.

Thereafter, an error measuring step S300 of obtaining an error by applying a linear system model to a parity equation is performed.

, residual)를 다음의 식 10에 의해 구할 수 있다.To detect current sensor and position sensor failures in the system,

, residual) can be obtained by the following equation.

......(식 10)

(Eq. 10)

)는

가 0이 아니라면

에 의해 영향을 받고,

가 0이라면 점근적으로 작아진다. 이러한 조건을 만족하는

,

를 찾기 위해, 다음의 식 11이 적용될 수 있다.To check for the presence of a fault, the error (

)

Is not zero

Affected by

Is 0, it becomes asymptotically small. To meet these conditions

,

In order to find, Equation 11 below can be applied.

......(식 11)

(Eq. 11)

이다.here,

to be.

,

의 기초가

의 남은 눌 스페이스(null space)라는 것을 가리킨다. 따라서

,

를 결정하기 위해,

의 눌 스페이스를 구하는 것이 필수적이다.

의 남은 눌 스페이스(

)는 식 9 및 식 11로부터의 매트릭스에 따라 계산될 수 있다.Equation 11

,

The basis of

Indicates the remaining null space of the. therefore

,

To determine,

It is essential to find the space to be pressed.

Remaining pressed space (

) Can be calculated according to the matrices from equations (9) and (11).

......(식 12)

(Eq. 12)

here,

이다.

to be.

,

,

및

는 적절하지 않으며, 이것은 오차의 실현을 불가능하게 만든다. 이에, 저대역 필터(low pass filter)가 적용될 수 있다.As represented by Equation 12,

,

,

And

Is not appropriate, which makes it impossible to realize the error. Thus, a low pass filter may be applied.

......(식 13)

(Eq. 13)

및

는 저대역 필터의 시간 상수이다. 식 13에 식 12을 적용시킴으로써

,

가 계산될 수 있고, 오차가 식 12 및 식 13을 식 10에 적용시킴으로써 계산될 수 있다. 계산된 오차는 다음의 식 14와 같다.In equation 13,

And

Is the time constant of the low pass filter. By applying equation 12 to equation 13

,

Can be calculated and the error can be calculated by applying equations 12 and 13 to equation 10. The calculated error is shown in Equation 14 below.

......(식 14)

(Eq. 14)

)가 알려져 있지 않기 때문에, 식 14를 실행하기 위해서는 그것이 추정되어야만 한다. 모터 속도 추정은 다음의 식 15로부터 구할 수 있다.Current motor speed (

Since is not known, it must be estimated to perform equation 14. The motor speed estimate can be obtained from the following equation (15).

......(식 15)

(Eq. 15)

및

는 저대역 필터의 시간 상수이다. 제안된 고장 검출 알고리즘(101)의 블록 다이아그램은 전술한 바와 같이 도 1과 같다. 그리고 제안된 고장 검출 알고리즘(101)은 식 14 및 식 15에 의해 실행될 수 있다.here,

And

Is the time constant of the low pass filter. The block diagram of the proposed failure detection algorithm 101 is as shown in FIG. 1 as described above. The proposed failure detection algorithm 101 can be executed by equations (14) and (15).

On the other hand, in the algorithm verification step S400 of the present embodiment, the usefulness of the failure detection algorithm 101 determined by the above-described steps can be proved by computer simulation.

The computer simulation that may be applied in the algorithm verification step S400 may be Matlab or Simulink. However, the present invention is not limited thereto.

The parameters of the embedded permanent magnet synchronous motor that can be applied in the algorithm verification step may be given as shown in Table 1 below.

sign Justice Number (unit)

Number of pole pairs 2

Rotor d-axis inductance 42.44 (mH)

Rotor q-axis inductance 79.57 (mH)

Stator resistance 1.93 (Ω)

Magnetic chain coefficient 0.311 (Vs / rad)

Motor inertia 0.003 (kgm ² )

Coefficient of friction 0.001 (Nm / s / rad)

The simulator for simulation is the same as that shown in FIG. 1, but the IPMSM controller 110, the PWM generator 130, and the inverter 140 may not be executed for the convenience of simulation.

To illustrate the effect of external load on error, although not shown, the input signal generator may include a virtual external load generator. The sensor detection generator can generate a fault signal of the current sensor and encoder to confirm the detectability of the proposed algorithm.

Meanwhile, the d-axis voltage and the q-axis voltage are shown in FIG. 3 when the external load of 5 Nm is applied every 10 seconds, and the result of the simulation is shown. As shown in Figure 4 it can be seen that the effect of the external load on the error is very small.

On the other hand, Figure 5 shows the results of the simulation when a 0.5V offset fault signal of the a-axis current sensor is applied every 20 seconds and an external load of 5 Nm is applied every 10 seconds. As shown here, it is easy to see that the change in error is so great that one of the sensors is abnormal after 20 seconds.

As such, in accordance with an embodiment of the present invention, a novel sensor failure detection method of a motor control system is a failure detection algorithm designed based on input-output linearization and parity equation approach to reduce changes in system nonlinearity and load torque effects. By doing so, there is an advantage in that failure detection of both the current sensor and the position sensor can be easily and accurately performed.

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit and scope of the invention. Therefore, such modifications or variations will have to be belong to the claims of the present invention.

100: embedded permanent magnet synchronous motor 101: failure detection algorithm
110: IPMSM controller 120: encoder
130: PWM generator 140: Inverter

Claims (8)

In the sensor failure detection method of the control system of the embedded permanent magnet synchronous motor,
A modeling step of modeling a model of the embedded permanent magnet synchronous motor defined by direct-quadrature (dq) transformation;
A linear model that transforms the nonlinear model of the embedded permanent magnet synchronous motor into a linearized model by input-output linearization providing a nonlinear equivalent deformation that linearizes the dynamic equations of the system Conversion step; And
An error measuring step of measuring a residual from the linearized model by applying a parity equation;
And detecting a failure of the current sensor and the position sensor of the system using a failure detection algorithm generated based on the above steps.
The method of claim 1,
A method for detecting a sensor failure of the control system of the embedded permanent magnet synchronous motor, further comprising an algorithm verification step, which is performed after the error measuring step and proves the usefulness of the proposed failure detection algorithm by computer simulation.
The method of claim 2,
And the computer simulation in the algorithm verification step is executed by a program comprising a matlab or simulink.
The method of claim 1,
In the modeling step, the voltage equation within the defined dq sync frame,

And

(here,

Is the d-axis voltage, R is the stator resistance,

D-axis current,

D-axis inductance,

Is the pole pair of the motor,

Silver motor speed,

Is the q-axis inductance,

Is the q-axis current,

Is the q-axis voltage,

Is a flux linkage. A failure detection method for a control system of an embedded permanent magnet synchronous motor.
5. The method of claim 4,
The current equation in the dq sync frame defined in the modeling step S100 is

(here,

Is the a-axis current,

Is the b-axis current,

Is the rotor position) sensor failure detection method of a control system of an embedded permanent magnet synchronous motor.
The method of claim 1,
In the linear model conversion step, a failure detection method of a control system of an embedded permanent magnet synchronous motor, which obtains the linearized model by applying an equation defining the nonlinear model in the modeling step to a Li derivative.
The method according to claim 6,
And applying the input-output linearization to obtain a model reflecting the nonlinear characteristics of the embedded permanent magnet synchronous motor.
In order to detect the failure of the current sensor and the failure sensor included in the control system of the embedded permanent magnet synchronous motor, a model defined by direct-quadrature (dq) transformation of the embedded permanent magnet synchronous motor is developed. After modeling, the nonlinear model of the embedded permanent magnet synchronous motor is linearized by input-output linearization providing a nonlinear equivalent deformation that linearizes the dynamic equations of the system. Converting the model into a model, and then applying a parity equation to measure the residual from the linearized model to generate a failure detection algorithm that is proved useful by computer simulation, and generates the generated failure detection algorithm. Embedded to detect the failure of the current sensor and position sensor of the system in real time Control system of the permanent magnet synchronous motor.

Images