A method and device for dual closed-loop fault prediction and health management of permanent magnet synchronous motors

Abstract

The present invention discloses a dual-closed-loop fault prediction and health management method and device for permanent magnet synchronous motors, belonging to the field of fault prediction and health management based on digital twin technology. The digital twin model is self-corrected through an abnormal trigger mechanism, achieving dynamic self-correction and updating of the inherent parameters of the digital twin model, ensuring that the operating state of the digital twin model remains consistent with the actual equipment and that parameters with different degrees of change urgency maintain synchronization. The operating data of the digital twin model of the permanent magnet synchronous motor is used to replace the physical data information collected to perform fault prediction and health management on the monitoring equipment, thereby improving the accuracy of the prediction results.

Description

Technical Field

[0001] The present invention relates to the field of fault prediction and health management based on digital twin technology, and in particular to a method and device for dual closed-loop fault prediction and health management of a permanent magnet synchronous motor. Background Art

[0002] Digital twin-based modeling technology boasts high fidelity (multidisciplinary, multi-physics), real-time performance, and multi-scale capabilities. Due to its high degree of digitization and low development and application costs, it is widely used in fault prediction and health management, design optimization, and manufacturing for complex electromechanical equipment. However, traditional fault prediction and health management approaches face numerous disadvantages for equipment with increasingly complex structural integration, increasing interdisciplinary integration, and high design and manufacturing costs. First, spatial structure and sensor limitations hinder the volume and quality of data collected on key equipment parameters. Second, during data collection on target equipment, the fault prediction and health management phase is susceptible to random noise interference. Finally, data collection and maintenance for physical equipment are expensive and costly. Given these challenges, the risk and cost of failure or malfunction are substantial if effective maintenance is not performed on complex equipment. Therefore, research on fault prediction and health management using digital twin technology that can reflect the operating status of physical equipment is necessary and meaningful.

[0003] The hallmark of digital twins is their ability to accurately and in real time reflect the operating status of physical equipment. Unlike traditional simulation technologies, digital twin models can update and adjust based on changes in the physical equipment, rather than simply simulating transient conditions within a specific state. Therefore, digital twin-based fault prediction and health management must reflect the operating status of the digital twin at various stages and provide ongoing maintenance guidance.

[0004] Permanent magnet synchronous motors (PMSMs) are highly interdisciplinary, and their parameters vary in speed. When using digital twin technology for PMSM fault prediction and health management, ensuring that their operating status remains consistent with the actual equipment and that the twin system's parameters, varying in speed, remain synchronized are crucial indicators of the digital twin's high fidelity and real-time performance. Summary of the Invention

[0005] The purpose of the present invention is to provide a dual-closed-loop fault prediction and health management method and device for a permanent magnet synchronous motor. By combining a dynamically updated digital twin model with fault prediction and health management, the operating status is kept consistent with the actual equipment and the parameters are kept synchronized, thereby improving the accuracy of the prediction results.

[0006] To achieve the above object, the present invention provides the following solutions:

[0007] A dual closed-loop fault prediction and health management method for a permanent magnet synchronous motor, comprising:

[0008] A digital twin model of the permanent magnet synchronous motor is established based on first principles, and parameters used for diagnosing faults of the permanent magnet synchronous motor are determined as inherent parameters of the digital twin model, and the inherent parameters are preset;

[0009] Real-time acquisition of key characteristic parameters of permanent magnet synchronous motors;

[0010] Based on the key characteristic parameters collected in real time, the anomaly detection trigger mechanism of the digital twin model is used to detect anomalies in the current digital twin model;

[0011] If an anomaly is detected in the current digital twin model, the inherent parameters of the current digital twin model are updated based on the key characteristic parameters collected in real time;

[0012] If it is detected that there is no abnormality in the current digital twin model, the inherent parameters of the current digital twin model are used to perform fault prediction and health management on the permanent magnet synchronous motor.

[0013] Optionally, the method of establishing a digital twin model of the permanent magnet synchronous motor based on first principles and determining parameters for diagnosing permanent magnet synchronous motor faults as inherent parameters of the digital twin model specifically includes:

[0014] Based on the first principles, a digital twin model of the permanent magnet synchronous motor is established.

[0015]

[0016] Y=(i d ,i q ,U d ,U q ,ω e )

[0017]

[0018] Among them, Y is the output of digital twin, U d is the voltage component parallel to the NS pole direction of the magnet, U q is the voltage component perpendicular to the NS pole direction of the magnet, R is the resistance, i d 、i q are the current components parallel and perpendicular to the NS pole direction of the magnet, L d 、L q are the inductance components parallel and perpendicular to the NS pole direction of the magnet, ω e is the speed, Ψ is the rotor flux, T eis the electromagnetic torque, T m is the output torque, B is the motor friction coefficient, ω r is the mechanical speed, J is the motor moment of inertia;

[0019] will i d ,i q ,U d ,U q ,T m ,ω r ,ω e Divided into time-varying parameters, P n ,B,J are time-invariant parameters, ψ,R,L d ,L q is the intrinsic parameter, U=(ω r ) is external input, c j =(T m ) corresponds to the operating conditions.

[0020] Optionally, the real-time acquisition of key characteristic parameters of the permanent magnet synchronous motor specifically includes:

[0021] Selecting key characteristic parameters that can reflect the operating characteristics of the permanent magnet synchronous motor; the key characteristic parameters include line voltage, line current, mechanical speed and output torque;

[0022] According to the selected key characteristic parameters, corresponding sensors are arranged on the permanent magnet synchronous motor;

[0023] The synchronous acquisition card collects the detection data of the sensor, filters and upsamples the detection data to obtain the pre-processed detection data;

[0024] The pre-processed inspection data is received by the computer containing the digital twin model.

[0025] Optionally, the method of performing anomaly detection on the current digital twin model based on the key characteristic parameters collected in real time by utilizing the anomaly detection trigger mechanism of the digital twin model specifically includes:

[0026] Input power is selected as the anomaly detection metric for the digital twin model;

[0027] According to the real-time collected line voltage U L and line current I L , using the formula P 实体 =U L *I L *cosθ, calculate the input power P of the permanent magnet synchronous motor 实体 ; where cosθ is the power factor;

[0028] Using formula P 数字孪生 =1.5*U q *iq , calculate the input power P of the digital twin model 数字孪生 ;

[0029] According to formula D 偏差 =|P 实体 -P 数字孪生 |, calculate the input power P of the permanent magnet synchronous motor 实体 The input power P of the digital twin model 数字孪生 Deviation;

[0030] If the deviation is greater than the anomaly detection threshold, the current digital twin model is determined to be abnormal;

[0031] If the deviation is less than or equal to the anomaly detection threshold, it is determined that the current digital twin model has no anomaly.

[0032] Optionally, updating the inherent parameters of the current digital twin model according to the key characteristic parameters collected in real time specifically includes:

[0033] Based on the key characteristic parameters collected in real time, the least squares estimation method is used to update the inherent parameters of the current digital twin model;

[0034] Perform consistency measurement on the updated digital twin model.

[0035] Optionally, the consistency measurement of the updated digital twin model specifically includes:

[0036] The U in the digital twin model d 、U q and i q The state vector is composed and the state vector of the digital twin model is expressed as S = {u dE ,u qE ,i qE}, the state vector of the permanent magnet synchronous motor is expressed as S = {u dD ,u qD ,i qD};

[0037] Using the formula Normalize the state vector of the digital twin model and use the formula Normalize the state vector of the permanent magnet synchronous motor;

[0038] According to the formula Calculate the normalized state vector X of the digital twin model EN and the normalized state vector X of the permanent magnet synchronous motor DN The Euclidean distance D E ;

[0039] If the Euclidean distance D EIf the distance is less than or equal to the distance threshold, the digital twin model and the permanent magnet synchronous motor are judged to be consistent in state, and the updated digital twin model is output;

[0040] If the Euclidean distance D E If the distance is greater than the distance threshold, it is determined that there is a deviation between the digital twin model and the permanent magnet synchronous motor state, and the updated digital twin model is corrected again until the corrected digital twin model and the permanent magnet synchronous motor state are consistent.

[0041] Optionally, the performing fault prediction and health management on the permanent magnet synchronous motor according to the acquired inherent parameters of the current digital twin model specifically includes:

[0042] Select rotor flux from inherent parameters to perform fault prediction and health management on permanent magnet synchronous motors;

[0043] Performing up-and-down sampling on the rotor magnetic flux to obtain a corrected rotor magnetic flux;

[0044] Perform data fitting on the corrected rotor flux to obtain the demagnetization trend curve of the rotor flux;

[0045] In the demagnetization trend curve, if the rotor flux is less than or equal to the flux threshold, it is predicted that the permanent magnet synchronous motor is about to fail or has failed, and maintenance advice is given based on the failure prediction result.

[0046] A dual-closed-loop fault prediction and health management device for a permanent magnet synchronous motor, comprising: a synchronous acquisition card, a computer, and a key characteristic parameter detection sensor;

[0047] The key characteristic parameter detection sensor is arranged on the permanent magnet synchronous motor;

[0048] The signal output end of the key characteristic parameter detection sensor is connected to the signal input end of the synchronous acquisition card, and the signal output end of the synchronous acquisition card is connected to the computer;

[0049] The synchronous acquisition card is used to collect the key characteristic parameter values ​​of the permanent magnet synchronous motor detected by the key characteristic parameter detection sensor, and filter and up-sample the key characteristic parameter values, and transmit the obtained pre-processed key characteristic parameter values ​​to the computer;

[0050] The computer is used to establish a digital twin model of the permanent magnet synchronous motor, and dynamically self-correct and update the digital twin model based on the preprocessed key characteristic parameter values, and then use the dynamically updated digital twin model to perform fault prediction and health management on the permanent magnet synchronous motor.

[0051] Optionally, the computer includes: a permanent magnet synchronous motor double closed loop fault prediction and health management system;

[0052] The permanent magnet synchronous motor dual closed-loop fault prediction and health management system specifically includes:

[0053] The anomaly detection module is used to detect anomalies in the current digital twin model based on the key feature parameters collected in real time and using the anomaly detection trigger mechanism of the digital twin model;

[0054] An update module is used to update the inherent parameters of the current digital twin model based on the key characteristic parameters collected in real time if an anomaly is detected in the current digital twin model;

[0055] An intrinsic parameter acquisition module is used to obtain the intrinsic parameters of the current digital twin model if it is detected that there is no abnormality in the current digital twin model;

[0056] The model application module is used to perform fault prediction and health management on the permanent magnet synchronous motor based on the inherent parameters of the current digital twin model;

[0057] The human-machine interactive visualization platform interface is used to display the operating status and maintenance instructions of the permanent magnet synchronous motor according to the electrical, mechanical power, temperature, power and fault prediction and health management sections.

[0058] Optionally, the key characteristic parameter detection sensors include: a current sensor, a voltage sensor, a rotation speed sensor and a torque sensor.

[0059] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0060] The present invention discloses a method and device for dual-closed-loop fault prediction and health management of a permanent magnet synchronous motor. The method performs self-correction of a digital twin model through an abnormal trigger mechanism, realizes dynamic self-correction and update of the inherent parameters of the digital twin model, ensures that the operating state of the digital twin model is consistent with the actual equipment, and maintains synchronization with parameters with different degrees of change urgency. The operating data of the digital twin model of the permanent magnet synchronous motor is used to replace the physical collected data information to perform fault prediction and health management on the monitoring equipment, thereby improving the accuracy of the prediction results. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0062] Figure 1 A flowchart of a method for double closed-loop fault prediction and health management of a permanent magnet synchronous motor provided by an embodiment of the present invention;

[0063] Figure 2 A schematic diagram of a dual-closed-loop fault prediction and health management method for a permanent magnet synchronous motor provided by an embodiment of the present invention;

[0064] Figure 3 Schematic diagram of parameter definition and division of the digital twin model provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0065] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0066] The purpose of the present invention is to provide a dual-closed-loop fault prediction and health management method and device for a permanent magnet synchronous motor. By combining a dynamically updated digital twin model with fault prediction and health management, the operating status is kept consistent with the actual equipment and the parameters are kept synchronized, thereby improving the accuracy of the prediction results.

[0067] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0068] The present invention provides a dual closed-loop fault prediction and health management method for permanent magnet synchronous motors, which realizes the self-correction and dynamic update of the digital twin model through an abnormal trigger mechanism, wherein the dynamic update mainly refers to the ability to adjust the main parameters of the digital twin model according to external sensor information. Figure 2 As shown in the figure, the dual closed loop refers to the inner closed loop formed by the digital twin model with dynamic self-correction and update, and the outer closed loop including the fault prediction and health management method based on the digital twin model. It clarifies that the fault prediction and health management based on digital twins must revolve around the two links of "establishing the model" and "applying the model".

[0069] The rotor flux has the greatest impact on the performance of permanent magnet synchronous motors. If demagnetization occurs, the rotor flux value decreases, the motor's armature reaction will accelerate rotor demagnetization and increase the motor temperature, endangering the motor's operation. However, due to the limitations of the motor structure and the expensive magnetic field sensors, it is inconvenient to carry out real-time monitoring of the motor. Therefore, a dual closed-loop fault prediction and health management method based on a dynamically updated digital twin model can enable operators to better understand the motor's operating conditions.

[0070] The embodiment of the present invention provides a method for double closed-loop fault prediction and health management of a permanent magnet synchronous motor, such as Figure 1 As shown, including:

[0071] Step 1: Establish a digital twin model of the permanent magnet synchronous motor based on first principles, determine parameters used to diagnose permanent magnet synchronous motor faults as inherent parameters of the digital twin model, and preset the inherent parameters.

[0072] The specific process is as follows:

[0073] 1.1 Divide the parameters in the first principle model

[0074] First principles model Model output Model intrinsic parameters (x1,…,x m ,x m+1 ,…,x n ), where x1,…,x m is the time-varying parameter in the intrinsic parameters of the model, x m+1 ,…,x n It is a time-invariant parameter among the intrinsic parameters, and the external input is U(t). There are many time-varying parameters, and the rates of change vary. When monitoring the state of the digital twin model, they need to be divided to facilitate the clear definition of their meaning and centralized parameter management.

[0075] In the digital twin model, the equipment's digital twin model parameters are divided into "time-varying parameters" and "time-invariant parameters" based on their dynamic characteristics. Time-varying parameters can be further divided into sudden-varying parameters and slowly-varying parameters.

[0076] A time-varying parameter is one whose value changes as the running time increases, while a time-invariant parameter is one whose value does not change as the running time increases.

[0077] Among the time-varying parameters, the parameters that change quickly with time are classified as "suddenly varying parameters", and the parameters that change slowly with time are classified as "slowly varying parameters".

[0078] Generally, the internal operating status parameters of the equipment will change with the change of working conditions, and the change is relatively fast, which are sudden-changing parameters; the inherent parameters inside the equipment generally have the law of degenerative change and change slowly, which are slow-changing parameters.

[0079] The monitoring of sudden-change parameters is real-time, while the monitoring of slow-change parameters should be triggered by conditions.

[0080] Get the actual measurement value U d实 is the voltage component U parallel to the NS pole direction of the magnet d实 =U L *(1-cos 2 θ) 1 / 2 ; Actual measured value U q实 =UL *cosθ is the voltage component perpendicular to the NS pole direction of the magnet; the current component i is parallel to the NS pole direction of the magnet d实 =0, Actual measured value motor speed ω e =P n* ω r , P n is the number of motor pole pairs. Magnetic flux Ψ, resistance R, inductance L q , L d , torque T e and the speed ω e .

[0081] In order to clarify the nature and meaning of each parameter in the motor digital twin model, the internal parameters of the motor, magnetic flux Ψ, resistance R, inductance L q , L d Divided into slowly varying parameters within the digital twin model, the voltage U d 、U q , the current i of the digital twin model d 、i q Divided into sudden change parameters, speed ω r is the input parameter, torque T m For operating conditions.

[0082] 1.2 Constructing a digital twin model based on the divided parameters

[0083] Based on the establishment of the digital twin model of the first principle, the digital twin model of the permanent magnet synchronous motor is

[0084]

[0085]

[0086] Y=(i d ,i q ,U d ,U q ,ω e ) (4)

[0087] Where Y is the output of the digital twin, and the voltage component U of the digital twin model parallel to the NS pole direction of the magnet is d , the voltage component U perpendicular to the NS pole direction of the digital twin model q , B is the motor friction coefficient, J is the motor moment of inertia, T m is the output torque, T e is the electromagnetic torque.

[0088] The above formula is based on the first principle Digital twin model of permanent magnet synchronous motor.

[0089] In the digital twin model of permanent magnet synchronous motor, different parameters are divided into detailed categories based on the properties and change characteristics of the parameters, and different types of parameters in the digital twin model are defined. Figure 3 As shown, i q ,U d ,U q ,ψ,R,L d ,L q ,T e ,ω e ,ω m The time-varying parameters correspond to x1,…,x in the digital twin model. M , P n ,B,J are time-invariant parameters, corresponding to x in the digital twin model M+1 ,…,x N , c j =(T m ) corresponds to the working condition, U=(ω r ) is external input, x m =(ψ,R,L d ,L q ) is the inherent parameter of the digital twin model, so the model parameters can be divided into time-varying parameters d r =(i d ,i q ,U d ,U q ,T m ,ω r ,ω e ).

[0090] Step 2: Collect key characteristic parameters of the permanent magnet synchronous motor in real time.

[0091] First, the key parameter values ​​of the permanent magnet synchronous motor that need to be monitored are clarified, and the sensor layout of the physical measurement points on the permanent magnet synchronous motor is optimized. By fusing the sensor information at a certain measurement point, the credibility of the key parameter status collected from the entity is guaranteed.

[0092] Select appropriate acquisition parameters (key characteristic parameters) based on the characteristics of the equipment's operating conditions. To ensure data reliability, deploy an appropriate number of redundant sensors for data fusion. The appropriate number is to ensure information accuracy and facilitate signal fusion, so the specific location and number of sensors must be determined for each device.

[0093] As a typical electromechanical device, the permanent magnet synchronous motor has a line voltage U L , line current I L , mechanical speed ω r and output torque Tm It can reflect the running status of the motor, so it is clear here that the focus should be on the motor's line voltage U L , line current I L , mechanical speed ω r and output torque T m Therefore, current sensors, voltage sensors, speed sensors and torque sensors are arranged to monitor the motor.

[0094] After sensors deployed on physical equipment collect key characteristic parameters, they transmit them to a synchronous acquisition card. The synchronous acquisition card performs data processing, such as filtering and upsampling, to ensure the quality of the data collected from the physical equipment. This ensures that the data frequency is lower than the digital twin model's solution frequency. The resulting data is high-quality key characteristic parameter data that meets the model's solution frequency after data fusion. The synchronous acquisition card transmits this high-quality key characteristic parameter data to the computer containing the digital twin model, which serves as the operational data source for the digital twin model.

[0095] The synchronous data acquisition card is the medium for transmitting data between the sensor and the computer, and its main body is not on the actual equipment.

[0096] That is, by utilizing industrial Internet of Things technology, the data collected by the sensor is transmitted to the computer end that establishes a digital twin model of the permanent magnet synchronous motor through a wired or wireless information interaction mechanism.

[0097] Step 3: Based on the key feature parameters collected in real time, use the anomaly detection trigger mechanism of the digital twin model to perform anomaly detection on the current digital twin model.

[0098] According to the parameter properties divided in the digital twin model of the permanent magnet synchronous motor, a key parameter performance indicator is selected as the trigger indicator for abnormal operation of the digital twin of the permanent magnet synchronous motor. After the deviation between the selected indicator of the digital twin model and the actual permanent magnet synchronous motor exceeds the set threshold, the digital twin model will self-update according to the status of the actual permanent magnet synchronous motor and enter the inner closed-loop circulation stage; if the deviation of the selected indicator is less than the set threshold, the digital twin model is considered to have high fidelity, that is, the solution result of the digital twin can reflect the operating status of the actual equipment, and its operating results can be used for fault prediction and health management, entering the outer closed-loop.

[0099] Select input power P, that is, line voltage U L , line current I L The product of θ and power factor cosθ is used as the detection indicator of digital twin model anomaly.

[0100] 3.1 Calculate the input power P of the physical motor based on the parameters monitored in step 1 实体

[0101] Input power P of the physical motor 实体 , where the formula is used to calculate P 实体 =U L *I L *cosθ.

[0102] 3.2 Calculating the input power P of the digital twin model 数字孪生

[0103] Input power P of the digital twin model 数字孪生 , where P is calculated by the formula 数字孪生 =1.5*U q *i q , where U q From formula (2), we can get q It is obtained by formula (2).

[0104] 3.3 According to P 实体 and P 数字孪生 Calculation deviation

[0105] Calculate the deviation D 偏差 =|P 实体 -P 数字孪生 |, to compare the status differences between the two

[0106] If D 偏差 >Trigger: If the difference between the input power of the actual equipment and the digital twin exceeds the established abnormality detection threshold Trigger, the digital twin model needs to be updated.

[0107] If D 偏差 ≤Trigger means that the difference between the input power of the actual equipment and the digital twin does not exceed the anomaly detection threshold, then the digital twin model is considered to meet the accuracy requirements and does not need to be updated.

[0108] Step 4: If an abnormality is detected in the current digital twin model, the inherent parameters of the current digital twin model are updated according to the key characteristic parameters collected in real time.

[0109] After completing the inner closed-loop digital twin self-update stage, it is necessary to measure the state consistency of the digital twin model. The key state combination in the digital twin model is selected to form a state vector, which is compared with the vector in the actual permanent magnet synchronous motor. The comparison result is compared with the set consistency threshold. If the requirements are met, it will run according to the updated model parameters. Otherwise, it will be updated again or the model will be adjusted.

[0110] 4.1 Update the digital twin model

[0111] When the digital twin model needs to be updated, the line voltage U of the motor collected from the physical model of the digital twin motor is used. L , line current I L , mechanical speed ω r The inductance L(L d =L q =L), flux linkage Ψ, and resistance R are identified to obtain the actual internal parameters of the motor: inductance L, resistance R, and flux linkage Ψ:

[0112] The calculation formula is as follows:

[0113] ①[U d ] m*1 =[-ω e *i q ] m*1 *L

[0114] ②

[0115] The recursive least squares method is used to find parameters as follows:

[0116] K m+1 =P m+1 *h(m+1)*[λ+h(m+1)*P m *h T (m+1)] -1

[0117]

[0118]

[0119] where θ m are the inherent parameters of the physical motor, which are represented by the inductance L in formula ① (for surface-mount synchronous motors) and the resistance R and flux linkage Ψ in formula ②. The subscript m is the number of observations, and K m is the gain correction parameter, for θ m and K m Assign a sufficiently small initial value; Z m is the observed quantity, i.e. the current and voltage signals collected from the physical device; h m Respectively expressed as ω in formula ① e *i q and i in formula ② q ,ω e .

[0120] These three parameters can be detected by sensors. Through multiple iterations, the inductance L, resistance R, and flux linkage Ψ are calculated using the least squares method from the voltage, current, and speed signals collected from the actual system. The identification results are then updated to the corresponding parameters in the digital twin model.

[0121] 4.2 Consistency measurement of the updated digital twin model

[0122] After the digital twin model is updated, it is necessary to measure the state consistency between the digital twin model and the actual equipment to ensure the fidelity of the digital twin model.

[0123] Taking Euclidean distance as an example, the u d ,u q ,i q As the key parameters, the state vector is composed and the state vector in the digital twin model is compared with the state vector of the physical motor, which are expressed as S = {u dE ,u dE ,i qE}, S={u dD ,u dD ,i qD}.

[0124] 4.21 First, the state vectors from the motor entity and the digital twin model are normalized.

[0125]

[0126]

[0127]

[0128]

[0129] 4.22 Calculate the Euclidean distance and compare it with the threshold

[0130]

[0131] According to experience, the threshold is set. E If the value is less than the set threshold, the digital twin is considered to be consistent with the actual device state. E If it is greater than the set threshold, it is considered that there is a large deviation between the digital twin and the actual device status, and the model needs to be corrected again.

[0132] Step 5: If it is detected that there is no abnormality in the current digital twin model, the inherent parameters of the current digital twin model are used to perform fault prediction and health management on the permanent magnet synchronous motor.

[0133] The inherent parameter data obtained from the digital twin model requires further data processing, such as data classification, filtering, upsampling and downsampling, and storage. Before entering the fault prediction and health management links, the data generated by the digital twin model needs to be preprocessed.

[0134] Based on the processed digital twin model data, relevant fault prediction and health management work is carried out, including degradation monitoring of key parameters and equipment fault diagnosis. Finally, based on the guidance feedback from fault prediction and health management, human-computer interaction technology is used to conduct relevant equipment condition monitoring and maintenance work.

[0135] Specifically, based on the processed flux data, the flux obtained from the digital twin model can be fitted to obtain the rotor flux demagnetization trend curve Ψ=g(t), where the function g is the demagnetization time function fitted from the digital twin model. This curve is used to predict the rotor demagnetization situation. Based on the calculated results, the threshold Ψ' is set. When Ψ≤Ψ', the motor needs to be repaired as soon as possible to ensure its healthy operation.

[0136] The dual closed-loop fault prediction and health management method for permanent magnet synchronous motors of the present invention is not only applicable to permanent magnet synchronous motors, but also to other electromechanical equipment. The process for other electromechanical equipment is as follows:

[0137] Identify the key parameter values ​​of the physical equipment that needs to be monitored, and optimize the sensor layout of the physical measurement points on the equipment. By fusing the sensor information at a certain measurement point, ensure the credibility of the key parameter status collected from the entity;

[0138] Utilize industrial Internet of Things technology to transmit the data collected by sensors to the computer end with a digital twin model through wired or wireless information interaction mechanism;

[0139] The multidisciplinary digital twin model established based on first principles will generate relevant output results based on the data transmitted by the sensors. In the digital twin model, different parameters are divided into different categories based on their properties and change characteristics, and different types of parameters in the digital twin model are defined;

[0140] Based on the parameter properties divided in the digital twin model, a key parameter performance indicator is selected as the trigger indicator for digital twin operation anomalies. After the deviation between the digital twin model and the actual target operation exceeds the set threshold, the digital twin model will self-update according to the status of the actual equipment and enter the inner closed loop cycle stage. If the indicator deviation is less than the set threshold, the digital twin model is considered to have high fidelity and the solution result of the digital twin model can reflect the operating status of the actual equipment. Its operating results can be used for fault prediction and health management, entering the outer closed loop.

[0141] After completing the internal closed-loop digital twin self-update phase, the state consistency of the digital twin model needs to be measured. The key state combinations in the digital twin model are selected to form a state vector. This is then compared with the vector in the actual equipment. The resulting distance is then compared with the consistency threshold. If the requirements are met, the system will run according to the updated model parameters. Otherwise, another update or model adjustment will be performed.

[0142] The data obtained from the digital twin model requires further data processing, such as data classification, filtering, upsampling and downsampling, and storage. The data generated by the digital twin model must be preprocessed before entering the fault prediction and health management stages.

[0143] Based on the processed digital twin model data, relevant fault prediction and health management processing are carried out, and degradation monitoring and fault diagnosis of key model parameters are carried out;

[0144] Finally, based on the guidance information fed back by fault prediction and health management, relevant maintenance work is carried out on the equipment entity using human-computer interaction.

[0145] Based on the fundamental characteristics of the physical equipment and utilizing Industrial Internet of Things (IIoT) technology, a digital twin system based on a first-principles mathematical model is established. Specific state parameters are selected as indicators of abnormal deviations between the digital twin model and the actual equipment. If the selected indicators exceed a threshold, the digital twin model parameters are updated using sensor detection or linear and nonlinear parameter estimation methods. Finally, using consistency metrics such as Euclidean distance or Mahalanobis distance, the key state parameters of the actual equipment are compared with the corresponding parameters in the digital twin model to evaluate the similarity of the technical status of the digital twin with the actual equipment. This completes the internal closed-loop process to ensure high-fidelity modeling of the digital twin model and guarantee the accuracy of the digital twin.

[0146] Using the operating condition data generated by the digital twin model for fault prediction and health management, or locating and determining fault patterns based on the specific values ​​of the parameters, this process can be applied to a wide range of equipment types and fault prediction and health management algorithms. After signal filtering, methods such as signal feature extraction, Bayesian estimation, support vector machines, artificial neural networks, fault trees, and expert systems are used. The specific method selected needs to be combined with the model characteristics and operating conditions of the actual equipment.

[0147] Based on the equipment operating status results obtained during the fault prediction and health management phase, this information is fed back to operators via a human-computer interaction visualization platform interface, building an information exchange bridge between the digital environment and the real world, helping professionals fully understand the operating status of the target equipment. The human-computer interaction interface is divided into sections based on their subject scope and function, clearly representing the diverse and complex data types within the digital twin system. Based on their subject scope, the modules are divided into electrical, mechanical power, temperature, power, and fault prediction and health management, each displaying the operating status and maintenance guidance of the permanent magnet synchronous motor.

[0148] The present invention determines the sensor measurement points that need to be collected and the number of arranged sensors based on the characteristics of the actual monitoring equipment; utilizes the Internet of Things technology to transmit the key parameters obtained from the actual equipment to the computer end containing the digital twin model for the operation and solution of the digital twin model; constructs the mathematical model of the digital twin according to the first principles, divides the parameter categories in the digital twin system, and clarifies the key parameters in the digital twin system; selects reasonable digital twin model abnormality trigger parameters as indicators of the deviation between the digital twin model and the actual equipment, and when the indicator exceeds the set threshold, the internal parameters of the digital twin model need to be updated and adjusted; the internal parameter update of the digital twin can be carried out by direct sensor acquisition or various parameter estimation methods to adjust the internal inherent parameters of the digital twin model. After the update, the key operating status parameters of the actual equipment and the corresponding status parameters in the digital twin model are selected for consistency measurement comparison to ensure the accuracy of the digital twin model. After completing the above steps, the inner closed-loop stage of dynamically updating the digital twin model is completed, and the "model establishment" work of the digital twin system is completed; while the digital twin model ensures its accuracy, the data generated in its digital twin model is pre-processed for calculation and reference in the fault prediction and health management links; based on the solution results of fault prediction and health management, the operating status of the equipment is presented through a visual interface through the human-computer interaction system, which helps professional and technical personnel to carry out inspection and maintenance of the target equipment, completing the outer closed-loop work of "application model".

[0149] The present invention proposes a dual-closed-loop fault prediction and health management method for permanent magnet synchronous motors, which clearly divides and defines model establishment and information interaction based on digital twin systems; realizes self-correction of the digital twin model through an abnormal trigger mechanism; finally, the operating data of the permanent magnet synchronous motor digital twin model replaces the physical collection data information to perform fault prediction and health management on the monitoring equipment, solving the current problems of incompleteness, lack of systematization and lack of detail in fault prediction and health management based on digital twins.

[0150] The present invention also provides a permanent magnet synchronous motor double closed loop fault prediction and health management device, comprising: a synchronous acquisition card, a computer and a key characteristic parameter detection sensor.

[0151] A key characteristic parameter detection sensor is arranged on the permanent magnet synchronous motor. The signal output end of the key characteristic parameter detection sensor is connected to the signal input end of the synchronous acquisition card, and the signal output end of the synchronous acquisition card is connected to a computer. The synchronous acquisition card is used to collect the key characteristic parameter values ​​of the permanent magnet synchronous motor detected by the key characteristic parameter detection sensor, filter and up-sample the key characteristic parameter values, and transmit the pre-processed key characteristic parameter values ​​to the computer. The computer is used to establish a digital twin model of the permanent magnet synchronous motor, and dynamically self-correct and update the digital twin model based on the pre-processed key characteristic parameter values, and then use the dynamically updated digital twin model to perform fault prediction and health management on the permanent magnet synchronous motor.

[0152] Exemplarily, the key characteristic parameter detection sensors include: a current sensor, a voltage sensor, a rotation speed sensor, and a torque sensor.

[0153] The computer includes: a permanent magnet synchronous motor dual closed-loop fault prediction and health management system. The permanent magnet synchronous motor dual closed-loop fault prediction and health management system specifically includes:

[0154] The anomaly detection module is used to detect anomalies in the current digital twin model based on the key feature parameters collected in real time and using the anomaly detection trigger mechanism of the digital twin model;

[0155] An update module is used to update the inherent parameters of the current digital twin model based on the key characteristic parameters collected in real time if an anomaly is detected in the current digital twin model;

[0156] An intrinsic parameter acquisition module is used to obtain the intrinsic parameters of the current digital twin model if it is detected that there is no abnormality in the current digital twin model;

[0157] The model application module is used to perform fault prediction and health management on the permanent magnet synchronous motor based on the inherent parameters of the current digital twin model;

[0158] The human-machine interactive visualization platform interface is used to display the operating status and maintenance instructions of the permanent magnet synchronous motor according to the electrical, mechanical power, temperature, power and fault prediction and health management sections.

[0159] After obtaining the diagnostic results of fault prediction and health management, the human-computer interaction link, as a bidirectional information exchange link within the digital twin system, can provide feedback and display the status and current health status of the equipment, helping professionals participate in equipment monitoring and maintenance. They can always understand the equipment's operating status and provide maintenance advice and related support work through the fault prediction and health management link. The human-computer interaction system for the target equipment serves as an information exchange bridge between the digital environment and the real world, helping professionals fully understand the operating status of the target equipment. When dividing the human-computer interaction interface, the area is divided according to its subject scope and function, which can clearly represent the various types of data in the digital twin system.

[0160] Human-computer interaction and monitoring and maintenance of actual equipment. In the human-computer interaction interface, the data information of the digital twin model of the permanent magnet synchronous motor is divided into electrical, mechanical power, temperature, power and fault prediction and health management modules, which respectively display the operating status and maintenance instructions of the permanent magnet synchronous motor. Among them, the electrical module includes internal parameters that can reflect the main performance of the motor, such as resistance, inductance flux, and voltage and current that describe the operating status of the motor; in the mechanical power state and interactive control module, a three-dimensional model of the permanent magnet synchronous motor of the detection object and a real-time three-dimensional temperature distribution diagram of the motor are displayed, and the degree of consistency between the current digital twin model of the motor and the actual equipment is given. Below the module, the motor operating conditions of speed and torque are also included. Finally, the operator can use different buttons to start, pause, shut down, record data, and turn on sensors for the digital twin system; the temperature and energy module displays the temperature and motor power distribution of each key node of the motor; finally, there is a fault prediction and health management module based on digital twins, which includes the total running time of the motor, the number of digital twin model corrections, the remaining life estimation results, the demagnetization degree of the monitored motor rotor, and the collected motor voltage and current status. Through the human-computer interaction system consisting of an electrical module, a mechanical power status and control module, a temperature and energy module, and a fault prediction and health management module, the monitored motor equipment status and operation and maintenance results are presented to the staff, completing the transfer of information from the digital environment to the real world.

[0161] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0162] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only intended to help understand the method and core concept of the present invention. At the same time, those skilled in the art will find that the specific implementation methods and application scopes may vary based on the concept of the present invention. In summary, the contents of this specification should not be construed as limiting the present invention.

Claims

1. A method for double closed-loop fault prediction and health management of a permanent magnet synchronous motor, characterized in that: The dual closed loop refers to an inner closed loop formed by a digital twin model that dynamically self-corrects and updates, and an outer closed loop that includes a fault prediction and health management method based on the digital twin model; the method includes: A digital twin model of a permanent magnet synchronous motor is established based on first principles, and parameters used for diagnosing faults of the permanent magnet synchronous motor are determined as inherent parameters of the digital twin model, and the inherent parameters are preset; the inherent parameters include rotor flux; Real-time acquisition of key characteristic parameters of permanent magnet synchronous motors; Based on the key characteristic parameters collected in real time, the anomaly detection trigger mechanism of the digital twin model is used to detect anomalies in the current digital twin model; If an anomaly is detected in the current digital twin model, the inherent parameters of the current digital twin model are updated based on the key characteristic parameters collected in real time; If it is detected that there is no abnormality in the current digital twin model, the inherent parameters of the current digital twin model are used to perform fault prediction and health management on the permanent magnet synchronous motor; specifically, the method includes: selecting the rotor flux from the inherent parameters to perform fault prediction and health management on the permanent magnet synchronous motor; performing up and down sampling on the rotor flux to obtain a corrected rotor flux; performing data fitting on the corrected rotor flux to obtain a demagnetization trend curve of the rotor flux; in the demagnetization trend curve, if the rotor flux is less than or equal to the flux threshold, it is predicted that the permanent magnet synchronous motor is about to fail or has already failed, and maintenance advice is given based on the fault prediction results.

2. The method for double closed-loop fault prediction and health management of a permanent magnet synchronous motor according to claim 1, characterized in that: The digital twin model of the permanent magnet synchronous motor is established based on the first principles, and the parameters used to diagnose the fault of the permanent magnet synchronous motor are determined as the inherent parameters of the digital twin model, specifically including: Based on the first principles, a digital twin model of the permanent magnet synchronous motor is established. Y=(i d ,i q ,U d ,U q ,ω e ) Among them, Y is the output of digital twin, U d is the voltage component parallel to the NS pole direction of the magnet, U q is the voltage component perpendicular to the NS pole direction of the magnet, R is the resistance, i d 、i q are the current components parallel and perpendicular to the NS pole direction of the magnet, L d , L q are the inductance components parallel and perpendicular to the NS pole direction of the magnet, ω e is the speed, Ψ is the rotor flux, T e is the electromagnetic torque, T m is the output torque, B is the motor friction coefficient, ω r is the mechanical speed, J is the motor moment of inertia; will i d ,i q ,U d ,U q ,T m ,ω r ,ω e Divided into time-varying parameters, P n ,B,J are time-invariant parameters, ψ,R,L d ,L q ) is the intrinsic parameter, U=(ω r ) is external input, c j =(T m ) corresponds to the operating conditions.

3. The method for double closed-loop fault prediction and health management of a permanent magnet synchronous motor according to claim 2, characterized in that: The real-time acquisition of key characteristic parameters of the permanent magnet synchronous motor specifically includes: Selecting key characteristic parameters that can reflect the operating characteristics of the permanent magnet synchronous motor; the key characteristic parameters include line voltage, line current, mechanical speed and output torque; According to the selected key characteristic parameters, corresponding sensors are arranged on the permanent magnet synchronous motor; The synchronous acquisition card collects the detection data of the sensor, filters and upsamples the detection data to obtain the pre-processed detection data; The pre-processed inspection data is received by the computer containing the digital twin model.

4. The method for double closed-loop fault prediction and health management of a permanent magnet synchronous motor according to claim 3, characterized in that: The method of performing anomaly detection on the current digital twin model based on the key characteristic parameters collected in real time and utilizing the anomaly detection trigger mechanism of the digital twin model specifically includes: Input power is selected as the anomaly detection metric for the digital twin model; According to the real-time collected line voltage U L and line current I L , using the formula P 实体 =U L *I L *cosθ, calculate the input power P of the permanent magnet synchronous motor 实体 ; where cosθ is the power factor; Using the formula P 数字孪生 =1.5*U q *i q , calculate the input power P of the digital twin model 数字孪生 ; According to formula D 偏差 =|P 实体 -P 数字孪生 |, calculate the input power P of the permanent magnet synchronous motor 实体 The input power P of the digital twin model 数字孪生 Deviation; If the deviation is greater than the anomaly detection threshold, the current digital twin model is determined to be abnormal; If the deviation is less than or equal to the anomaly detection threshold, it is determined that the current digital twin model has no anomaly.

5. The method for double closed-loop fault prediction and health management of a permanent magnet synchronous motor according to claim 2, characterized in that: The updating of the inherent parameters of the current digital twin model based on the key characteristic parameters collected in real time specifically includes: Based on the key characteristic parameters collected in real time, the least squares estimation method is used to update the inherent parameters of the current digital twin model; Perform consistency measurement on the updated digital twin model.

6. The method for double closed-loop fault prediction and health management of a permanent magnet synchronous motor according to claim 5, characterized in that: The consistency measurement of the updated digital twin model specifically includes: The U in the digital twin model d 、U q and i q The state vector is composed and the state vector of the digital twin model is expressed as S = {u dE ,u qE ,i qE }, the state vector of the permanent magnet synchronous motor is expressed as S = {u dD ,u qD ,i qD }; Using the formula Normalize the state vector of the digital twin model and use the formula Normalize the state vector of the permanent magnet synchronous motor; According to the formula Calculate the normalized state vector X of the digital twin model EN and the normalized state vector X of the permanent magnet synchronous motor DN The Euclidean distance D E ; If the Euclidean distance D E If the distance is less than or equal to the distance threshold, the digital twin model and the permanent magnet synchronous motor are judged to be consistent in state, and the updated digital twin model is output; If the Euclidean distance D E If the distance is greater than the distance threshold, it is determined that there is a deviation between the digital twin model and the permanent magnet synchronous motor state, and the updated digital twin model is corrected again until the corrected digital twin model and the permanent magnet synchronous motor state are consistent.

7. A dual closed-loop fault prediction and health management device for a permanent magnet synchronous motor, characterized in that: The device is used to implement the method according to any one of claims 1 to 6, and the device comprises: a synchronous acquisition card, a computer, and a key characteristic parameter detection sensor; The key characteristic parameter detection sensor is arranged on the permanent magnet synchronous motor; The signal output end of the key characteristic parameter detection sensor is connected to the signal input end of the synchronous acquisition card, and the signal output end of the synchronous acquisition card is connected to the computer; The synchronous acquisition card is used to collect the key characteristic parameter values ​​of the permanent magnet synchronous motor detected by the key characteristic parameter detection sensor, and filter and up-sample the key characteristic parameter values, and transmit the obtained pre-processed key characteristic parameter values ​​to the computer; The computer is used to establish a digital twin model of the permanent magnet synchronous motor, and dynamically self-correct and update the digital twin model based on the preprocessed key characteristic parameter values, and then use the dynamically updated digital twin model to perform fault prediction and health management on the permanent magnet synchronous motor.

8. The dual closed-loop fault prediction and health management device for a permanent magnet synchronous motor according to claim 7, characterized in that: The computer includes: a permanent magnet synchronous motor double closed loop fault prediction and health management system; The permanent magnet synchronous motor dual closed-loop fault prediction and health management system specifically includes: The anomaly detection module is used to detect anomalies in the current digital twin model based on the key feature parameters collected in real time and using the anomaly detection trigger mechanism of the digital twin model; An update module is used to update the inherent parameters of the current digital twin model based on the key characteristic parameters collected in real time if an anomaly is detected in the current digital twin model; An intrinsic parameter acquisition module is used to obtain the intrinsic parameters of the current digital twin model if it is detected that there is no abnormality in the current digital twin model; The model application module is used to perform fault prediction and health management on the permanent magnet synchronous motor based on the inherent parameters of the current digital twin model; The human-machine interactive visualization platform interface is used to display the operating status and maintenance instructions of the permanent magnet synchronous motor according to the electrical, mechanical power, temperature, power and fault prediction and health management sections.

9. The dual closed-loop fault prediction and health management device for a permanent magnet synchronous motor according to claim 7, characterized in that: The key characteristic parameter detection sensors include: a current sensor, a voltage sensor, a rotation speed sensor and a torque sensor.

Images