A motor system current sensor fault comprehensive detection method

Abstract

The present application belongs to the field of motor system fault analysis, and relates to a kind of motor system current sensor fault comprehensive detection method, comprising: the normalized stator current and its absolute value are respectively taken difference, quotient operation to obtain intermediate variable, then pass through average value calculator processing to obtain fault diagnosis variable, and determine adaptive threshold value through the arithmetic mean of diagnostic variable, finally, combined with fault diagnosis criterion, current sensor gain, offset, broken wire and stuck fault can be comprehensively detected.The present application uses phase current as detection variable, avoids the increase of additional detection equipment; through phase current amount design diagnostic variable, multiple current sensor faults can be accurately detected; the adaptive threshold value is proposed to realize the real-time accurate estimation of threshold value required by detection criterion, and provide a more accurate and effective new way for comprehensive diagnosis of ac motor system current sensor gain, offset, broken wire and stuck fault.

Description

Technical fields:

[0001] This invention belongs to the field of motor drive technology and relates to a comprehensive fault detection method for motor system current sensor, which is used to determine the fault type of motor system current sensor and accurately locate the fault point. Background technology:

[0002] During long-term operation of motor systems, harsh working environments such as high temperature and high vibration may cause inaccurate signal sampling by current sensors, leading to faults such as current signal gain deviation, disconnection, and jamming. For applications with high reliability requirements, it is necessary to quickly and accurately identify the type and location of current sensor faults online to facilitate subsequent fault-tolerant operation control. Therefore, comprehensive detection and diagnosis of various current sensor faults in motor systems is a crucial prerequisite for maintaining the safe and reliable operation of the system.

[0003] While most current sensor diagnostic strategies are fast and accurate, their detection range is limited to a single fault type. Different sampling faults in current sensors can cause different types of pulsations in the drive system, which in turn can affect the closed-loop control strategy to varying degrees. Therefore, how to accurately identify and differentiate current sensor fault types is a problem that urgently needs to be solved. Summary of the Invention:

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and propose a comprehensive fault detection method for current sensors in motor systems. This method can effectively distinguish between current sensor gain, offset, open circuit, and jamming faults, providing a more accurate and effective new approach for precisely locating various current sensor fault types and fault points in AC motors.

[0005] The technical solution of this invention is implemented as follows:

[0006] A comprehensive fault detection method for current sensors in motor systems includes the following steps:

[0007] (1) Determination of intermediate variables: The stator phase current is transformed by Park to obtain the current quantity in the synchronous rotating coordinate system to obtain the vector magnitude. The ratio of each stator phase current to the vector magnitude is used as the normalized phase current i. n * , change i n * The intermediate variable d is obtained by subtracting and quotienting its absolute value. n e n ;

[0008] (2) Determination of diagnostic variables: intermediate variable d n e n The diagnostic variable D is then obtained by processing the values ​​using an average calculator. nE n When different types of current sensor failures occur, D n E n Different numerical changes are presented;

[0009] (3) Adaptive threshold determination: The adaptive threshold is determined by the arithmetic mean of the diagnostic variables, thereby realizing the real-time estimation of the reference value of the diagnostic variables;

[0010] (4) Fault location of current sensor: Combining detection criteria, through diagnostic variable D n E n With adaptive thresholds, comprehensive detection, identification, and location of current sensor gain, offset, open circuit, and jamming faults can be achieved.

[0011] Among them, the ratio of each phase current of the stator to the vector magnitude is used as the normalized phase current i. n * , change i n * The intermediate variable d is obtained by subtracting and quotienting its absolute value. n e n The calculation formula is:

[0012]

[0013] Among them, i n * It is the normalized current value of each phase, |i n * | is its absolute value, I am ω is the phase current amplitude, ω is the electric angular velocity, and θ is the phase current amplitude. e ε is the initial phase, ε is a very small positive number, and t is time.

[0014] Among them, the intermediate variable d n e n The diagnostic variable D is then obtained by processing the values ​​using an average calculator. n E n The formula for calculating integrals is:

[0015]

[0016] Where T1 is the electrical period, t is time, and d is the electrical period. n e n It is an intermediate variable; dt is the time integral.

[0017] Optionally, for any number of phases in a motor system current sensor failure condition, D n E n The number of variables has changed, with each phase current corresponding to two diagnostic variables.

[0018] Optionally, the motor is a synchronous motor.

[0019] Optionally, the motor is an asynchronous motor.

[0020] Optionally, the motor can be three-phase.

[0021] Compared with the prior art, the present invention has the following advantages: First, it uses phase current as the detection variable, avoiding the need for additional detection equipment; second, it can comprehensively and accurately detect faults of multiple current sensors; third, the proposed adaptive threshold enables real-time and accurate estimation of the threshold required for the detection criteria. Attached image description:

[0022] Figure 1 This invention relates to schematic diagrams of various current sensor faults;

[0023] Figure 2 Waveforms of intermediate variables during various current sensor malfunctions;

[0024] Figure 3 This is a schematic diagram of the fault diagnosis principle of the current sensor involved in the present invention. Detailed implementation method:

[0025] The present invention will be further described below with reference to the embodiments and accompanying drawings.

[0026] This embodiment uses a nine-phase motor with an H-bridge power supply and an open winding as a prototype. Various current sensor faults in phase a1 are used as case studies. The specific fault detection process is as follows:

[0027] (1) Phase current magnitude calculation: To reduce the impact of stator current amplitude changes on the detection algorithm during load abrupt changes, the current needs to be normalized. First, the stator nine-phase current i in the natural coordinate system is calculated. n (n=a1,…,c3), the current i in the synchronous rotating coordinate system is obtained through Park transformation. d1 i q1 (i d1 i q1 (This represents the current along the d-axis and q-axis in a synchronously rotating coordinate system.) The magnitude of the current can be expressed as:

[0028] (2) Current normalization processing: The ratio of each phase current to the current magnitude is used as the current normalization preprocessing to obtain the normalized values ​​of the nine-phase current. and its absolute value |i n * The normalized phase currents are expressed as:

[0029] i a1_m =βi a1 +C=βI amcos(ωt+θ ori )+C

[0030] i a2_m =i a2

[0031] …

[0032] i c3_m =i c3

[0033] In the formula, i a1 i a2 ...i c3 i represents the actual stator current. a1_m i a2_m ...i c3_m For current sensors measuring current, β is the current sensor gain factor, C is the offset coefficient, and I... am ω is the phase current amplitude, ω is the motor angular frequency, and θ is the phase current amplitude. ori Let t be the initial phase angle, and t be the time.

[0034] The above formula can be used to obtain the phase current reference waveforms for various faults of the current sensor:

[0035]

[0036]

[0037] (3) Determination of intermediate variables: the normalized nine-phase current i n * Its absolute value |i n * By performing subtraction and quotient operations respectively, the intermediate diagnostic variable d of the detection method in this embodiment is obtained. n =|i n * |-i n * e n =i n * / (|i n * |+ε), where ε is a minimal positive number to avoid poles. For different current sensor faults, the intermediate diagnostic variable d... n e n The specific representation method is as follows:

[0038] (3A) When a positive gain fault occurs in the current sensor, the faulty phase i n * and its absolute value |i n * The amplitude increases proportionally, therefore after the fault, d n =|in * |-i n * It becomes a half-wave function with increasing amplitude, and e n =i n * / (|i n * |+ε) remains a square wave signal with an amplitude of one; d during a negative gain fault n The opposite change occurs, e n It remains unchanged.

[0039] (3B) When a positive offset fault occurs in the current sensor, the faulty phase i n * The waveform shifts, causing its absolute value |i n * | becomes a function with different magnitudes for the positive and negative half-cycles, therefore d n =|i n * |-i n * It becomes a half-wave function with increasing amplitude, e n =i n * / (|i n * |+ε) becomes a square wave signal with an increased duty cycle; during a negative offset fault, d n With e n The opposite change occurs.

[0040] (3C) When a current sensor disconnection fault occurs, the entire faulty phase i n * Waveform and its absolute value |i n * | becomes zero, i.e., after the fault |i n * |=i n * =0, therefore, the intermediate variable d n =|i n * |-i n * and e n =i n * / (|i n * |+ε) all become zero after a fault.

[0041] (3D) When a current sensor fails to register a positive current sensor jam, the entire faulty phase i n * Waveform and its absolute value |i n* | becomes a constant value greater than zero, i.e., after the fault, |i n * |=i n * Therefore, the intermediate variable d n =|i n * |-i n * =0, e n =i n * / (|i n * |+ε)=1;D during negative jamming fault n =|i n * |-i n * ≠0、e n =i n * / (|i n * |+ε)=-1.

[0042] (4) Obtaining diagnostic variables: intermediate variable d n e n The final diagnostic variables are then obtained by processing each variable using an average value calculator; the n-phase current then corresponds to 2n diagnostic variables. For different current sensor faults, the diagnostic variable D... n E n The specific representation method is as follows:

[0043] (4A) When a positive gain fault occurs in the current sensor, d n It becomes a half-wave function with increasing amplitude, e n The two signals are square wave signals with an amplitude of one. After averaging, the diagnostic variable D is obtained. n E n Therefore, the faulty phase D n Greater than normal phase, faulty phase E n Equal to normal phase; in case of negative gain fault, faulty phase D n Less than normal phase, faulty phase E n It is equivalent to the normal phase.

[0044] (4B) When a current sensor positive offset fault occurs, d n It becomes a half-wave function with increasing amplitude, e n The signal is transformed into a square wave with an increased duty cycle. After averaging the two signals, the diagnostic variable D is obtained. n E n Then the faulty phase D n Greater than normal phase, faulty phase E nGreater than normal phase; during a negative gain fault, the faulty phase D n Less than normal phase, faulty phase E n Smaller than the normal phase.

[0045] (4C) When a current sensor disconnection fault occurs, d n With e n The result is zero; after averaging the two values, we obtain the diagnostic variable D. n E n It is also zero.

[0046] (4D) When a current sensor fails to register a current sensor, the intermediate variable d n =0, e n =1, therefore D n =0, E n =1; d when there is a negative jamming fault n ≠0、e n =-1, therefore, D n =0, E n =-1.

[0047] (5) Adaptive threshold determination: through diagnostic variable D n E n The arithmetic mean determines the adaptive threshold K. D K E This enables real-time estimation of reference values ​​for diagnostic variables. The specific algorithm is as follows:

[0048]

[0049] The Max and Min functions can filter out the absolute average value of the phase current in the gain fault, where n represents the number of motor phases.

[0050] (6) Current sensor fault location: Based on the following detection criteria, through the diagnostic variable D n E n With adaptive threshold K D K E This enables comprehensive detection, identification, and location of faults such as current sensor gain, offset, open circuit, and jamming.

[0051] The testing criteria are shown in the table below:

[0052]

[0053] This invention uses phase current as the detection variable, which can comprehensively detect and accurately locate faults such as current sensor gain, offset, open circuit, and jamming. Compared with existing technologies, using phase current as the detection variable avoids the need for additional detection equipment and can comprehensively and accurately detect multiple current sensor faults.

Claims

1. A comprehensive fault detection method for current sensors in a motor system, characterized in that, Includes the following steps: (1) Determination of intermediate variables: The stator phase current is transformed by Park to obtain the current quantity in the synchronous rotating coordinate system to obtain the vector magnitude. The ratio of each stator phase current to the vector magnitude is used as the normalized phase current i. n * , change i n * The intermediate variable d is obtained by subtracting and quotienting its absolute value. n e n ; (2) Determination of diagnostic variables: intermediate variable d n e n The diagnostic variable D is then obtained by processing the values ​​using an average calculator. n E n ; (3) Adaptive threshold determination: The adaptive threshold is determined by the arithmetic mean of the diagnostic variables; (4) Fault location of current sensor: Combining detection criteria, through diagnostic variable D n E n With adaptive threshold K D K E To determine the comprehensive detection, identification, and location of faults such as current sensor gain, offset, open circuit, and jamming; The ratio of each phase current of the stator to its vector magnitude is used as the normalized phase current i. n * , change i n * The intermediate variable d is obtained by subtracting and quotienting its absolute value. n e n The calculation formula is: Among them, i n * It is the normalized current value of each phase, |i n * | is its absolute value, I am ω is the phase current amplitude, ω is the electric angular velocity, and θ is the phase current amplitude. e ε is the initial phase, ε is a very small positive number, and t is time; intermediate variable d n e n The diagnostic variable D is then obtained by processing the values ​​using an average calculator. n E n The formula for calculating integrals is: Where T1 is the electrical period, t is time, and d is the electrical period. n e n It is an intermediate variable, and dt is the time integral; The testing criteria are shown in the table below: Where β is the current sensor gain factor and C is the offset coefficient.

2. The comprehensive fault detection method for motor system current sensor according to claim 1, characterized in that, For any phase number motor system current sensor failure condition, D n E n The number of variables has changed, with each phase current corresponding to two diagnostic variables.

3. The comprehensive fault detection method for current sensors in motor systems according to claim 1, characterized in that, The motor is a synchronous motor.

4. The comprehensive fault detection method for motor system current sensor according to claim 1, characterized in that, The motor is an asynchronous motor.

5. The comprehensive fault detection method for motor system current sensor according to claim 1, characterized in that, The motor has three phases.

Images