A precise fault diagnosis method for ITSC based on signal compensation

Abstract

This invention relates to the field of fault detection technology for permanent magnet synchronous motors (PMSMs). It discloses a precise ITSC (Intermittently Intermittently Selective Susceptibility) fault diagnosis method based on signal compensation. By establishing a closed-loop dynamic response model of the PSM drive current and voltage based on the system transfer function, the dynamic response of current and voltage under varying operating conditions can be tracked in real time. Then, the response of the established model is used to compensate for the measured current and reference voltage of the motor driver, significantly reducing transient spikes in current and voltage caused by changes in operating conditions, effectively avoiding misdiagnosis of ITSC. Furthermore, the established dynamic response model can be combined with machine control strategies, and the proposed method does not require additional sensors, ensuring its practical implementation.

Description

Technical Field

[0001] This invention relates to the field of fault diagnosis technology for permanent magnet synchronous motors, and in particular to a precise fault diagnosis method for ITSC based on signal compensation. Background Technology

[0002] Inter-turn short circuits are one of the most common faults in permanent magnet synchronous motors (PMSMs), but most diagnostic methods can only be verified under stable operating conditions. In practical applications, the commanded speed or commanded torque of a PMSM drive system needs to be frequently changed to meet certain specific requirements. When a PMSM operates under variable operating conditions (especially transient operating conditions), sudden changes in speed or torque introduce additional harmonics into the machine current and voltage, which may lead to unexpected fault alarms and economic losses.

[0003] Unnecessary downtime due to misdiagnosis of faults can result in significant economic losses. Therefore, accurate detection of ITSC faults is crucial even under variable operating conditions. There are many methods for ITSC fault detection, including model-based methods, search coil (SC)-based methods, artificial intelligence-based methods, and signal feature analysis-based methods. Model-based methods mostly require a state observer to obtain estimates of the motion signal. The residual between the estimated and measured values ​​of the signal is used for fault diagnosis. Search coil (SC)-based methods propose a unique SC layout structure that can significantly reduce costs while meeting fault diagnosis requirements. However, installing the SC requires disassembling the motor, which is inconvenient for motors already in use. Machine learning-based fault diagnosis methods often do not require in-depth research into the fault mechanism but rather sufficient data to train the model; however, their accuracy and versatility for other mechanical transmissions need improvement.

[0004] The above-mentioned fault diagnosis methods are most effective under stable operating conditions. In practical applications, in order to meet certain special requirements, the machine's speed and torque often change frequently, and these changes, especially sudden changes, may lead to misdiagnosis of faults.

[0005] To address the technical problem of misdiagnosis, existing ITSC diagnostic methods under varying operating conditions utilize zero-sequence voltage elements (ZSVs) extracted by the Hilbert-Huang transform under transient conditions for ITSC detection. While faults can be detected during speed changes, this requires additional voltage sensors to measure ZSVs, increasing hardware costs. A Vold-Kalman filter order tracking algorithm is used to extract ITSC features in the angle domain for fault diagnosis; these features are independent of motor speed. However, the performance of this method is sensitive to filter bandwidth; a narrow bandwidth may fail to capture peak values ​​under transient operating conditions. Existing methods have discussed the negative impact of speed changes on ITSC diagnosis and analyzed faults by detecting speed commands in real time and keeping the fault indicator constant during speed changes, thus avoiding misdiagnosis. However, if the fault and speed change occur simultaneously, ITSC diagnosis may be missed. The method proposed in this application effectively solves the problem of misdiagnosis of ITSC caused by the simultaneous occurrence of faults and speed changes.

[0006] Furthermore, signal compensation can be used to address the problem of false fault diagnosis, eliminating spurious fault characteristics in current and voltage introduced by variable operating conditions. For example, one approach utilizes the stator current residual—the difference between the observed and actual current—for ITSC diagnosis; however, this method was validated at variable slope speeds, not under transient operating conditions where false fault diagnosis is more likely. Another approach uses the residual between the measured dq current and the evaluated dq current calculated from the dq voltage to detect ITSC under variable operating conditions such as speed variations. However, at low speeds, the fault-related component in the voltage weakens the fault characteristics, limiting its application. A third approach proposes a model-based residual analysis technique based on a finite element model (FEM) under variable speed conditions to achieve accurate ITSC detection. However, real-time finite element calculations require high-performance processors, and this method was only validated at variable slope speeds.

[0007] The above analysis shows that in order to accurately and reliably diagnose ITSCs in PMSMs (Permanent Magnet Synchronous Motors) under variable operating conditions, especially transient operating conditions, more insightful work is needed. In practical implementation, methods that do not require additional sensors are preferred. Therefore, those skilled in the art need a precise ITSC fault diagnosis method based on signal compensation. Summary of the Invention

[0008] The purpose of this invention is to solve the above problems by designing an accurate fault diagnosis method for ITSC based on signal compensation.

[0009] The technical solution of the present invention to achieve the above objectives is a precise fault diagnosis method for ITSC based on signal compensation, comprising the following processes:

[0010] A dynamic response model of the drive current and voltage of a closed-loop permanent magnet synchronous motor based on the system transfer function is established to track the dynamic response of current and voltage under varying operating conditions in real time.

[0011] The measured current and reference voltage of the motor driver are compensated by using a dynamic response model of the closed-loop permanent magnet synchronous motor drive current and voltage based on the system transfer function, thereby reducing the transient spikes in current and voltage.

[0012] The process of establishing the dynamic response model of the closed-loop permanent magnet synchronous motor drive current and voltage based on the system transfer function is as follows:

[0013] The field-oriented control-based dual closed-loop drive is applied to a permanent magnet synchronous motor, where K ps and K is These are the proportional and integral coefficients of the speed controller, respectively; K pc and K ic These are the proportional and integral coefficients of the current controller, respectively; K pwm T is the equivalent gain of the inverter, typically taken as 1; pwm P is the equivalent delay of the inverter; n and λ pm ω represents the number of pole pairs and flux linkage, respectively; J and B represent the moment of inertia and damping coefficient of the permanent magnet synchronous motor, respectively; r It is the mechanical angular velocity of the motor;

[0014] When considering only the q-axis, the closed-loop permanent magnet synchronous motor drive in transfer function form yields:

[0015]

[0016] In the formula, G s To G emf The definitions are G s The transfer function of the speed regulator element, G c G represents the transfer function of the current regulator element. p G represents the transfer function of the inverter stage. m G represents the transfer function of the PMSM (Permanent Magnet Synchronous Motor) stage. t G represents the transfer function of the torque equation. d G represents the transfer function of a component in the equation of motion of a machine. wn The transfer function G represents the unit conversion of rotational speed. emf L represents the transfer function of the back electromotive force element. s Let be the phase inductance of the motor under healthy conditions, and s be a complex variable representing the independent variable in the Laplace transform;

[0017] Based on equation (1), dynamic response models for q-axis current and q-cycle voltage are established respectively.

[0018] The dynamic response model of the q-axis current is i q i q The transfer function can be derived from the Mason gain formula as follows:

[0019]

[0020] Substituting (1) into (2), then i q The transfer function, denoted as T in , can be represented as:

[0021]

[0022] In the formula, a1, a2, a3, and a4 are all transfer functions T. in The coefficients of the molecule, b1, b2, b3, b4, b5, and b6, are all transfer functions T. in The coefficient of the denominator, where

[0023]

[0024] Then, considering only the load torque variation, assuming the motor speed is zero, and treating the q-axis current as a single-output system, then i q The transfer function can be derived as follows:

[0025]

[0026] Substituting (1) into (6), then i q The transfer function is denoted as T il , can be represented as:

[0027]

[0028] In the formula, c1, c2, c3, and c4 are all transfer functions T. il The coefficients of the molecule, d1, d2, d3, d4, d5, and d6, are all transfer functions T. il The coefficient of the denominator, where

[0029]

[0030] Combining equations (3) and (7), the dynamic response model of the q-axis current based on the transfer function under two inputs, speed and load torque, is as follows:

[0031]

[0032] The q-axis voltage dynamic response model u q The transfer function can be derived as:

[0033]

[0034] Substituting (1) into (11), then u q The transfer function is denoted as T un , can be represented as:

[0035]

[0036] In the formula, e1, e2, e3, e4, e5, and e6 are all transfer functions T. un The coefficients of the molecule, f1, f2, f3, f4, f5, and f6, are all transfer functions T. un The coefficient of the denominator, where

[0037]

[0038] Then, considering only the load torque variation, u q The transfer function can be derived as:

[0039]

[0040] Substituting (1) into (15), then u q The transfer function is denoted as T ul , can be represented as:

[0041]

[0042] In the formula, g1, g2, g3, g4, and g5 are all transfer functions T. ul The coefficients of the molecule, h1, h2, h3, h4, h5, and h6, are all transfer functions T. ul The coefficient of the denominator, where

[0043]

[0044]

[0045] Combining equations (12) and (16), the dynamic response model of the q-axis voltage based on the transfer function under two inputs, speed and load torque, is as follows:

[0046]

[0047] Beneficial effects

[0048] The ITSC accurate fault diagnosis method based on signal compensation, which utilizes the technical solution of this invention, has the following advantages:

[0049] 1. This application establishes a dynamic response model of the drive current and voltage of a closed-loop permanent magnet synchronous motor based on the system transfer function, obtains the dynamic response of current and voltage under varying operating conditions, and compensates for the harmonic transient spikes of current and voltage caused by sudden changes in motor speed and load torque, thereby reducing the possibility of misdiagnosis of ITSC.

[0050] 2. The method proposed in this application effectively solves the problem of misdiagnosis of ITSC caused by the simultaneous occurrence of faults and speed changes, especially the misdiagnosis when speed and load torque change abruptly. The ITSC diagnosis method based on machine current and voltage signal analysis is non-invasive and easy to implement, and does not have the above-mentioned drawbacks. Attached Figure Description

[0051] Figure 1 This is a flowchart of an ITSC accurate fault diagnosis method based on signal compensation as described in this invention;

[0052] Figure 2 This is a block diagram of a closed-loop permanent magnet synchronous motor drive in the form of the transfer function described in this invention;

[0053] Figure 3 This is the invention described Figure 2 A simplified block diagram;

[0054] Figure 4 This is a block diagram of a closed-loop permanent magnet synchronous motor drive in the form of a transfer function, considering only the q-axis, according to the present invention.

[0055] Figure 5 This is a block diagram of the q-axis current transfer function considering only the load torque variation in this invention;

[0056] Figure 6 This is a block diagram of the q-axis voltage transfer function considering only the rotational speed variation in this invention;

[0057] Figure 7 This is a block diagram of the q-axis voltage transfer function of the present invention, which only considers the load torque variation;

[0058] Figure 8 This is a topology diagram of the permanent magnet synchronous motor drive system with fault diagnosis in Embodiment 1 of the present invention. Detailed Implementation

[0059] The present invention will now be described in detail with reference to the accompanying drawings, such as... Figure 1 and Figure 2 As shown, field-oriented control (FOC) based dual closed-loop drive is widely used in permanent magnet synchronous motors, where K... ps and K is These are the proportional and integral coefficients of the speed controller, respectively; K pc and K icThese are the proportional and integral coefficients of the current controller, respectively; K pwm T is the equivalent gain of the inverter, typically taken as 1; pwm P is the equivalent delay of the inverter; n and λ pm ωr represents the number of pole pairs and flux linkage, respectively; J and B represent the moment of inertia and damping coefficient of the permanent magnet synchronous motor, respectively; ωr is the mechanical angular velocity of the motor. Figure 2 Simplified block diagram as follows Figure 3 As shown; for a closed-loop permanent magnet synchronous motor drive in transfer function form, considering only the q-axis, we can derive:

[0060]

[0061] Establish a dynamic response model for the q-axis current;

[0062] Considering only the speed variation, the load torque is treated as zero. The system is reconfigured as shown in the figure below, making it a system with speed as a single input and q-axis current as a single output. Figure 4 A closed-loop permanent magnet synchronous motor drive in transfer function form, considering only the q-axis;

[0063] Then i q The transfer function can be derived from the Mason gain formula as follows:

[0064]

[0065] Substituting (1) into (2), then i q The transfer function, denoted as T in , can be represented as:

[0066]

[0067] In the formula, a1, a2, a3, and a4 are all transfer functions T. in The coefficients of the molecule, b1, b2, b3, b4, b5, and b6, are all transfer functions T. in The coefficient of the denominator, where

[0068]

[0069] Then, considering only the load torque variation, the motor speed is assumed to be zero. Figure 4 The system is restructured into Figure 5 This makes it a system with a single input load torque and a single output q-axis current;

[0070] Then i q The transfer function can be derived as follows:

[0071]

[0072] Substituting (1) into (6), then i q The transfer function, denoted as T il , can be represented as:

[0073]

[0074] In the formula, c1, c2, c3, and c4 are all transfer functions T. il The coefficients of the molecule, d1, d2, d3, d4, d5, and d6, are all transfer functions T. il The coefficient of the denominator, where

[0075]

[0076]

[0077] Combining equations (3) and (7), the dynamic response model of the q-axis current based on the transfer function under two inputs, speed and load torque, is as follows:

[0078]

[0079] Establish a dynamic response model for the q-axis voltage;

[0080] Similarly, based on the additivity of linear systems, a dynamic response model for the q-axis voltage can also be obtained. First, considering only the rotational speed change, ... Figure 3 The system is restructured into Figure 6 ;

[0081] Then u q The transfer function can be derived as:

[0082]

[0083] Substituting (1) into (9), then u q The transfer function, denoted as T un , can be represented as:

[0084]

[0085] In the formula, e1, e2, e3, e4, e5, and e6 are all transfer functions T. un The coefficients of the molecule, f1, f2, f3, f4, f5, and f6, are all transfer functions T. un The coefficient of the denominator, where

[0086]

[0087] Then, considering only the load torque variation, Figure 3 The system is restructured into Figure 7 .

[0088] Then uq The transfer function can be derived as:

[0089]

[0090] Substituting (1) into (13), then u q The transfer function, denoted as T ul , can be represented as:

[0091]

[0092] In the formula, g1, g2, g3, g4, and g5 are all transfer functions T. ul The coefficients of the molecule, h1, h2, h3, h4, h5, and h6, are all transfer functions T. ul The coefficient of the denominator, where

[0093]

[0094]

[0095] Combining equations (12) and (16), the dynamic response model of the q-axis voltage based on the transfer function under two inputs, speed and load torque, is as follows:

[0096]

[0097] Based on the current and voltage models established above, the dynamic response of motor current and voltage under varying operating conditions can be obtained.

[0098] Example 1;

[0099] Using the established q-axis current and voltage dynamic response model, misdiagnosis of ITSC under varying operating conditions can be avoided. The PMSM drive topology for fault diagnosis is as follows: Figure 8 As shown. The input signals include the reference speed n* and the reference load torque T. L The data is fed into the closed-loop permanent magnet synchronous motor driver and the dynamic response model. Figure 8 in,i qt and u qt These are the q-axis current and voltage calculated by the model, respectively; i qm and u qm These are the q-axis current and voltage measured by the closed-loop PMSM driver, respectively. The ITSC fault diagnosis process proposed in this application is as follows: Figure 1 As shown. First, use i in the model. qt and u qt i for the motor qm and u qm Compensation is performed, and the compensated q-axis current and q-axis voltage are denoted as i. qc and u qc

[0100] Under stable operating conditions, i qt and u qt All values ​​are DC values, therefore compensation will not affect i. qm and u qm The harmonic content in [the system]. Under transient operating conditions, n* and T [are related to the harmonic content]. L The step change in i qm and u qm Transient spikes are introduced, containing abundant harmonics, including the second harmonic used for ITSC fault diagnosis. Therefore, false diagnoses of ITSC may occur. Using i qt and u qt Dynamic response to compensate for measured i qm and u qm This can reduce transient spikes, thus avoiding misdiagnosis. Furthermore, based on a healthy PMSM drive, a dynamic response model for q-axis current and voltage was established; therefore, when an ITSC fault occurs, compensation cannot mask the increase in second harmonics caused by the ITSC.

[0101] Fault Indicator FI ic and FI uc Define and display Figure 1 In, i * qc2 and u * qc2 These are the second harmonic reference values ​​for the q-axis current and voltage in a healthy PMSM, stored in a lookup table. These reference values ​​are not zero due to the inherent imbalance of the motor windings. For a normal PMSM, FI... ic and FI uc All are approximately 1, and they increase when an ITSC fault occurs. The principle for determining the fault diagnosis threshold is: take half the difference between the indicator value in the ITSC state and the indicator value in the healthy state, and then add 1, as shown below:

[0102]

[0103] In the formula, "+1" represents the elimination of false alarms that may be caused by noise in the experiment. When FI ic or FI uc A fault alarm is triggered when the corresponding detection threshold is exceeded.

[0104] In summary, the influence of different operating conditions on the harmonic components of motor current and voltage indicates that transient spikes in q-axis current and voltage caused by changes in operating conditions may lead to misdiagnosis of ITSC. These spikes can be viewed as a superposition of several harmonics of different frequencies, including those used for ITSC fault diagnosis. A dynamic response model of q-axis current and voltage based on the system transfer function was established, which can obtain accurate current and voltage responses under varying operating conditions. By using the model's response to compensate for the measured q-axis current and voltage, spikes caused by changes in operating conditions can be significantly reduced, including the second harmonic used for ITSC diagnosis. Furthermore, the compensation does not mask the characteristics of current and voltage caused by ITSC. Experimental results show that the proposed fault diagnosis method can avoid misdiagnosis of ITSC caused by abrupt changes in operating conditions, and can accurately detect ITSC even when the fault occurs during a transient change in operating conditions.

[0105] For other motor fault diagnosis based on the characteristic harmonics of motor current and voltage, false diagnoses may occur due to the large number of harmonics introduced by changes in operating conditions. Since this method can significantly reduce transient spikes in current and voltage under varying operating conditions, it can also be used for reliable detection of other faults under varying operating conditions.

[0106] It should be noted that in this application, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, the phrase "comprising an element defined as..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0107] The above technical solutions only embody the preferred technical solutions of the present invention. Any modifications that may be made by those skilled in the art to certain parts therein embody the principles of the present invention and fall within the protection scope of the present invention.

Claims

1. A signal compensation-based ITSC accurate fault diagnosis method, characterized in that, The method comprises the following steps: A dynamic response model of closed-loop permanent magnet synchronous motor driving current and voltage based on system transfer function is established to track the dynamic response of current and voltage in real time under variable working conditions; The measured current and reference voltage of the motor driver are compensated by using the dynamic response model of closed-loop permanent magnet synchronous motor driving current and voltage based on system transfer function, so as to reduce the transient peak of current and voltage; The establishment process of the dynamic response model of closed-loop permanent magnet synchronous motor driving current and voltage based on system transfer function is as follows: The double closed-loop drive based on field-oriented control is applied to permanent magnet synchronous motor, wherein K ps and K is are the proportional coefficient and integral coefficient of the speed controller respectively; K pc and K ic are the proportional coefficient and integral coefficient of the current controller respectively; K pwm is the equivalent gain of the inverter, usually taken as 1; T pwm is the equivalent delay of the inverter; P n and λ pm are the number of magnetic pole pairs and the flux linkage respectively; J and B are the moment of inertia and damping coefficient of the permanent magnet synchronous motor respectively; ω r is the mechanical angular velocity of the motor; When the closed-loop permanent magnet synchronous motor driving in the form of transfer function only considers the q-axis, the following formula (1) can be obtained: where G s G emf G s represents the transfer function of the speed regulator loop, G c represents the transfer function of the current regulator loop, G p represents the transfer function of the inverter loop, G m represents the transfer function of the PMSM loop, G t represents the transfer function of the torque equation loop, G d represents the transfer function of the mechanical motion equation loop, G wn represents the transfer function of the speed unit conversion loop, G emf represents the transfer function of the back electromotive force loop, L s is the phase inductance under the healthy state of the motor, and s is a complex variable representing an independent variable in Laplace transform. Based on formula (1), the dynamic response model of q-axis current and the dynamic response model of q-axis voltage are respectively established; The q-axis current dynamic response model is i q q The transfer function of the q-axis current dynamic response model can be derived according to the Mason gain formula as​ Substitute (1) into (2), then i q The transfer function of the system, denoted as T in , can be expressed as: wherein a1, a2, a3, a4 are coefficients of the transfer function T in b1, b2, b3, b4, b5, b6 are coefficients of the transfer function T in b1, b2, b3, b4, b5, b6 are coefficients of the transfer function T Then, only considering the load torque variation, taking the motor speed as zero, and taking the q-axis current as a single output system, the transfer function of i q can be derived as: Substitute (1) into (6), then i q The transfer function of i il may be expressed as wherein c1, c2, c3, c4 are coefficients of the transfer function T il wherein d1, d2, d3, d4, d5, d6 are coefficients of the transfer function T il wherein d1, d2, d3, d4, d5, d6 are coefficients of the transfer function T Combined with formula (3) and formula (7), the dynamic response model of q-axis current based on transfer function under the input of speed and load torque is as follows: The q-axis voltage dynamic response model u q The transfer function of the q-axis voltage dynamic response model u can be derived as: Substitute (1) into (11), then u q The transfer function of u un may be expressed as: wherein e1, e2, e3, e4, e5, e6 are coefficients of the transfer function T un wherein f1, f2, f3, f4, f5, f6 are coefficients of the transfer function T un wherein f1, f2, f3, f4, f5, f6 are coefficients of the transfer function T Then, considering only the load torque variation, u q The transfer function of u can be derived as: Substituting (1) into (15), we have q The transfer function of u ul may be expressed as wherein g1, g2, g3, g4, g5 are coefficients of the transfer function T ul wherein h1, h2, h3, h4, h5, h6 are coefficients of the transfer function T ul wherein h1, h2, h3, h4, h5, h6 are coefficients of the transfer function T Combined with formula (12) and formula (16), the dynamic response model of q-axis voltage based on transfer function under the input of speed and load torque is as follows:

Images