A superconducting motor quench detection method and device based on sliding time filtering

Abstract

This invention provides a method and apparatus for superconducting motor quench detection based on sliding time filtering, relating to the field of superconducting motor quench detection technology. The method includes: differentially processing the terminal voltage signals of adjacent superconducting rotor magnets in the superconducting motor; conducting three trial runs of the superconducting motor before quench detection to obtain interference frequency data; establishing sliding time windows with both long and short time scales during the actual operation of the superconducting motor, and performing notch filtering on the differential voltage signals of adjacent superconducting rotor magnets; performing linear regression fitting on the filtered signals within the sliding time windows to extract effective quench signal feature values; and establishing a quench protection criterion based on the quench feature values. This invention solves the problem of achieving high reliability and rapid response in quench detection technology under the complex interference environment of superconducting motor rotation operation, a problem that is difficult to address in existing technologies.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of superconducting motor quench detection, and particularly relates to a superconducting motor quench detection method and device based on sliding time filtering. BACKGROUND

[0002] Superconducting magnets have advantages of high energy density, low operating loss and compact structure, and have become an important basic component of advanced electromagnetic energy conversion equipment. Large-capacity superconducting motors are in the development stage, and the relatively mature superconducting motor scheme mostly adopts a superconducting magnet as a rotor excitation winding and a conventional copper winding as a stator armature winding structure. In such a system, reliable quench detection technology is a key prerequisite for suppressing local overheating of the superconducting magnet and ensuring safe and high-performance operation of the superconducting motor.

[0003] The quench detection technology based on voltage change has the advantages of fast response speed and accurate positioning, and is being considered for application in superconducting motor quench detection. However, under the rotating working condition of the superconducting motor, the lead voltage from the superconducting magnet will be coupled with multiple different frequency interference sources such as excitation power ripple, rotating mechanical vibration, electromagnetic induction and inductive voltage of the superconducting magnet itself, resulting in the inability to extract the effective quench voltage signal.

[0004] In order to overcome the interference of complex signals on voltage detection, an application No. 202410346510.8 discloses a superconducting motor quench detection method and device based on motor rotational symmetry, which measures the end voltage difference of each magnet of the superconducting motor rotor as a superconducting magnet quench characteristic value, and proposes a detection scheme based on the end voltage difference of multiple magnets, but the above method requires that each superconducting magnet be strictly symmetrical, which puts high requirements on process consistency and limits its practical application. Another application No. 202510139653.6 discloses an arc fault detection method and device based on voltage time difference change, which uses the end voltage difference of adjacent time cycles as a fault signal feature for recognition. The idea is to use a change amount detection method based on time voltage difference to identify fault features, but this method is sensitive to external disturbances, and slight current fluctuations may cause the judgment threshold to deviate, thereby causing protection misoperation and affecting system reliability.

[0005] In summary, the existing superconducting motor quench detection technology shows that the quench detection technology based on voltage difference can eliminate the interference of the inductive voltage of the superconducting magnet itself on the quench detection signal during excitation, but there are still interference sources such as excitation power ripple, mechanical vibration and electromagnetic induction, which affect the accurate judgment of the superconducting motor quench detection. If a conventional low-pass filter is directly used to process the differential voltage signal, in order to filter out the low-frequency mechanical vibration component, the cutoff frequency needs to be set very low, resulting in a large delay of the quench detection signal and affecting the real-time response speed of the quench detection device.

[0006] Therefore, existing queuing detection technology needs further development. Summary of the Invention

[0007] The purpose of this invention is to overcome the above-mentioned technical deficiencies and provide a superconducting motor quench detection method and device based on sliding time filtering, so as to solve the problem in the prior art of achieving quench detection technology that balances high reliability, strong adaptability and fast response capability in the complex interference environment of superconducting motor rotation operation.

[0008] To achieve the above technical objectives, the present invention adopts the following technical solution: A method for superconducting motor quench detection based on sliding time filtering is provided, comprising: differential processing of the terminal voltage signals of adjacent superconducting rotor magnets in the superconducting motor to eliminate inductive voltage interference; before quench detection, the superconducting motor undergoes three trial runs, wherein the three trial runs include static excitation, no-load rotating excitation, and loaded rotating excitation; during the three trial runs, interference frequencies coupled in the differential voltage signals are collected to form an interference frequency data table, wherein the interference frequencies mainly include the excitation power supply ripple frequency, The vibration frequency and electromagnetic induction frequency of rotating machinery are used. During the formal operation of the superconducting motor, the differential voltage signal and excitation current signal of adjacent superconducting rotor magnets are collected in real time. A sliding time window with both long and short time scales is established. Within the sliding time window, based on the interference frequency data table, notch filtering is performed on the real-time collected differential voltage signal. The sliding time window includes a short time window of tens of milliseconds and a long time window of hundreds of milliseconds. Within the sliding time window, a linear regression fitting method is used to extract effective quench feature values ​​from the filtered differential voltage signal. These quench feature values ​​include the equivalent quench resistance. R and equivalent rate of change of superconducting resistance Δ R / Δ t A quench protection criterion is established based on the quench characteristic value. The quench protection criterion includes a fast quench protection criterion and a backup quench protection criterion. The thresholds of the fast quench protection criterion and the backup quench protection criterion are adaptively adjusted in real time. The fast quench protection criterion is set with an interlocking trigger condition. If the fast quench protection criterion is established and the interlocking trigger condition is met, or if the backup quench protection criterion is established, then quench protection is applied to the superconducting motor.

[0009] Furthermore, the superconducting motor quench detection method also includes: real-time monitoring of the superconducting motor's operating status; if the superconducting motor is detected to be in a short-term overload operation under a strong excitation condition, then a quench protection criterion under the strong excitation condition is established based on preset rules.

[0010] Furthermore, the method for extracting quench feature values ​​includes: within the sliding time window, performing a time-based first-order linear regression fitting calculation on the filtered differential voltage signal and the excitation current signal to extract the real-time slidingly updated equivalent quench resistance sample sequence. R [ k ]; For the equivalent quench resistance sample sequence R [ k Repeat the time-based first-order linear regression fitting calculation to extract the equivalent quench resistance change rate sample sequence Δ. R [ k ] / Δ t .

[0011] Furthermore, the time-based linear regression fitting calculation method includes: establishing a first-order linear fitting model between the filtered differential voltage and the excitation current within the sliding time window; and using the least squares method to fit the parameters of the linear model to extract the equivalent quench feature value and bias. The specific method for parameter fitting is as follows: , in, J ( R , b ) represents the least squares objective function. N This represents the total number of sampling points within the sliding time window. i [ k [] indicates excitation current data. v [ k [] indicates the filtered differential voltage data. k Let be the index of each discrete sample within the sliding time window. R This represents the equivalent quench resistance characteristic value obtained from the fitting. b This represents the bias.

[0012] Furthermore, the method for establishing a rapid quench protection criterion based on quench characteristic values ​​specifically involves setting a short sliding time window. T short ,in, T short ≥10 ms; within the short sliding time window T short Within, obtain the updated equivalent quench resistance at each sampling time. R short and equivalent rate of change of superconducting resistance Δ R short / Δ t , as the go-out feature value sample sequence; calculate the mean of the go-out feature value sample sequence. μ Sum of variance Var, construct the quench threshold. R th_short andγ th_short ; in a short sliding time window T short If the following conditions are met R short > R th_short And Δ R short / Δ t > γ th_short If so, the rapid loss of overrun protection criterion is established.

[0013] Furthermore, the method for establishing a backup quench protection criterion based on the quench characteristic value is as follows: setting a long sliding time window. T long ,in, T long ≥100 ms; within the long sliding time window T long Within, obtain the updated equivalent quench resistance at each sampling time. R long and equivalent rate of change of superconducting resistance Δ R long / Δ t , as the go-out feature value sample sequence; calculate the mean of the go-out feature value sample sequence. μ Sum of variance Var, construct the quench threshold. R th_long and γ th_long ; in a long sliding time window T long If the following conditions are met R long > R th_long And Δ R long / Δ t > γ th_long If so, the backup protection criterion is established.

[0014] Furthermore, the mean of the statistical out-of-range characteristic value sample sequence is... μ The specific methods for calculating variance (Var) are as follows: , in, N This represents the total number of sampling points within the sliding time window. k Let be the index of each discrete sample within the sliding time window. r [ k [This refers to the equivalent quench resistance sample sequence extracted at each sampling time point within the sliding time window.] R [ kOr equivalent quench resistance change rate sample sequence Δ R [ k ] / Δt.

[0015] Furthermore, a threshold for determining quench is constructed. R th and γ th The specific method is as follows: , in, R th The threshold value for determining the equivalent quench resistance is [value]. γ th The threshold value for determining the rate of change of the equivalent quench resistance is [value]. λ This is the margin adjustment factor.

[0016] Furthermore, the interlocking triggering condition is specifically as follows: the quench detection device includes multiple detection units, each of which simultaneously monitors the terminal voltage signals of two adjacent superconducting rotor magnets and performs differential processing; the interlocking triggering condition is set as two quench detection units monitoring the terminal voltage of the same superconducting rotor magnet synchronously triggering the fast quench protection criterion.

[0017] A superconducting motor quench detection device based on sliding time filtering is also provided. The superconducting motor quench detection device includes: an analog sampling module for differentially processing the terminal voltage signals of adjacent rotor magnets of the superconducting motor and high-speed sampling of the differential voltage signal and excitation current signal; a frequency analysis module for identifying the interference frequency of the differential voltage signal and generating an interference frequency data table; a sliding filtering module for establishing a sliding time window, constructing a notch filter based on the interference frequency data table, and filtering the differential voltage signal in real time; a quench feature value extraction module for performing linear regression fitting calculations and extracting quench feature value samples within the sliding time window; and a quench diagnosis module for constructing quench protection action criteria, diagnosing quench events, and issuing quench alarm signals or quench protection action commands.

[0018] Beneficial effects: 1. The superconducting motor quench detection method of the present invention filters out interference from excitation power supply ripple, mechanical vibration, and electromagnetic induction layer by layer by establishing an interference signal frequency table. It is applicable to quench voltage detection methods based on the voltage difference of multiple superconducting magnet terminals, and does not require strict symmetry in the winding of each superconducting magnet, resulting in higher quench detection accuracy.

[0019] 2. The superconducting motor quench detection method of the present invention uses a linear regression fitting calculation method to extract the equivalent quench resistance and the rate of change of equivalent quench resistance from the filtered differential voltage signal in real time as quench feature values. This can avoid the operation of frequently adjusting the quench voltage detection threshold according to the excitation current change in conventional superconducting motor quench detection, and the quench detection reliability is higher.

[0020] 3. The superconducting motor quench detection method of the present invention establishes two sliding time windows, a long one and a short one, to execute the fast quench protection criterion and the backup quench protection criterion respectively. The fast quench protection criterion acts within tens of milliseconds to deal with sudden quench and is configured with interlocking trigger conditions to prevent the quench protection circuit from malfunctioning. The backup quench protection criterion acts within hundreds of milliseconds to provide more stable quench state diagnosis, prevent progressive quench, or provide redundant protection when the fast quench protection criterion occasionally misses a report, thereby improving the redundancy of the quench detection system.

[0021] 4. The superconducting motor quench detection method of the present invention realizes adaptive dynamic adjustment of the quench detection threshold. By continuously updating the mean and variance of the quench feature value samples within a sliding window and introducing a margin adjustment coefficient, the quench detection system can still maintain the accuracy of the quench detection threshold when facing fluctuations in operating conditions, equipment aging, and environmental interference. This significantly reduces the failure of the quench protection system caused by improper threshold setting, reduces the dependence on manual experience setting, and improves the intelligence level of the system. Attached Figure Description

[0022] Figure 1 This is a flowchart of the superconducting motor quench detection method based on sliding time filtering used in Embodiment 1 of the present invention; Figure 2 This is a schematic diagram of the superconducting motor quench detection device based on sliding time filtering used in Embodiment 1 of the present invention. Figure 3 This is a schematic diagram of the differential voltage signal at the rotor magnet end of the superconducting motor used in this invention; Figure 4 This is a schematic diagram of the equivalent quench resistance and the rate of change of equivalent quench resistance used in this invention; Figure 5 This is a schematic diagram of the quench protection action waveform used in this invention. Detailed Implementation

[0023] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0024] Example 1: This invention provides a superconducting motor quench detection method based on sliding time filtering. Please refer to [link to relevant documentation]. Figures 1 to 5 ,include: S100 performs differential processing on the terminal voltage signals of adjacent superconducting rotor magnets in the superconducting motor to eliminate inductive voltage interference. Before the quench detection, S200 conducts three trial runs on the superconducting motor, including static excitation, no-load rotary excitation, and loaded rotary excitation of the superconducting motor. S300 collects the interference frequencies coupled in the differential voltage signal during the three trial runs and forms an interference frequency data table, wherein the interference frequencies include the excitation power supply ripple frequency, the rotating machinery vibration frequency, and the electromagnetic induction frequency. During the actual operation of the superconducting motor, the S400 collects the differential voltage signal and excitation current signal of adjacent superconducting rotor magnets in real time. S500 establishes a sliding time window with two time scales, long and short. Within the sliding time window, based on the interference frequency data table, notch filtering is performed on the differential voltage signal acquired in real time. The sliding time window includes a short time window of tens of milliseconds and a long time window of hundreds of milliseconds. Within the sliding time window, S600 extracts effective quench feature values ​​from the filtered differential voltage signal using a linear regression fitting method, wherein the quench feature values ​​include the equivalent quench resistance. R and equivalent rate of change of superconducting resistance Δ R / Δ t ; Specifically, the method for extracting quench feature values ​​includes: within the sliding time window, performing a time-based first-order linear regression fitting calculation on the filtered differential voltage signal and the excitation current signal to extract the real-time sliding updated equivalent quench resistance sample sequence. R [ k ]; It should be noted that the time-based first-order linear regression fitting calculation method includes: Within the sliding time window, a first-order linear fitting model is established between the filtered differential voltage and the excitation current. The linear model is fitted with parameters using the least squares method to extract equivalent hysteresis features and biases. The specific parameter fitting method is as follows: , in, J ( R , b ) represents the least squares objective function. N This represents the total number of sampling points within the sliding time window. i [ k [] indicates excitation current data. v [ k [] indicates the filtered differential voltage data. k Let be the index of each discrete sample within the sliding time window. R This represents the equivalent quench resistance characteristic value obtained from the fitting. b This represents the bias.

[0025] For the equivalent supersonic resistance sample sequence R [ k Repeat the time-based linear regression fitting calculation to extract the equivalent quench resistance change rate sample sequence Δ. R [ k ] / Δ t .

[0026] S700 establishes a quench protection criterion based on the quench feature value, wherein the quench protection criterion includes a fast quench protection criterion and a backup quench protection criterion, the thresholds of the fast quench protection criterion and the backup quench protection criterion are adaptively adjusted in real time, and the fast quench protection criterion is set with an interlocking trigger condition. It should be noted that the method for establishing a rapid quench protection criterion based on quench characteristic values ​​is as follows: Set a short sliding time window T short ,in, T short ≥10 ms; In the short sliding time window T short Within, obtain the updated equivalent quench resistance at each sampling time. R short and equivalent rate of change of superconducting resistance Δ R short / Δ t , as the quench feature value sample sequence; The mean of the sample of the out-of-range feature values ​​is calculated. μ Sum of variance Var, construct the quench threshold. R th_short and γth_short ; In a short sliding time window T short If the following conditions are met R short > R th_short And Δ R short / Δ t > γ th_short If so, the rapid loss of overrun protection criterion is established.

[0027] The method for establishing a backup quench protection criterion based on quench characteristic values ​​is as follows: Set long sliding time window T long ,in, T long ≥100 ms; In the long sliding time window T long Within, obtain the updated equivalent quench resistance at each sampling time. R long and equivalent rate of change of superconducting resistance Δ R long / Δ t , as the quench feature value sample sequence; The mean of the sample of the out-of-range feature values ​​is calculated. μ Sum of variance Var, construct the quench threshold. R th_long and γ th_long ; In a long sliding time window T long If the following conditions are met R long > R th_long And Δ R long / Δ t > γ th_long If the backup overrun protection criterion is met, then the criterion is established. Here, the mean of the statistical overrun characteristic value sample sequence is... μ The specific methods for calculating variance (Var) are as follows: , in, N This represents the total number of sampling points within the sliding time window. k Let be the index of each discrete sample within the sliding time window. r [ k [This refers to the equivalent quench resistance sample sequence extracted at each sampling time point within the sliding time window.]R [ k Or equivalent quench resistance change rate sample sequence Δ R [ k ] / Δ t .

[0028] Specifically, construct the quench determination threshold. R th and γ th The specific method is as follows: , in, R th The threshold value for determining the equivalent quench resistance is [value]. γ th The threshold value for determining the rate of change of the equivalent quench resistance is [value]. λ This is the margin adjustment factor.

[0029] If the fast quench protection criterion is met and the interlocking trigger condition is satisfied, or if the backup quench protection criterion is met, then the superconducting motor is protected against quench.

[0030] It should be noted that, in this embodiment, the interlocking trigger condition is specifically as follows: the quench detection device includes multiple detection units, each of which simultaneously monitors the terminal voltage signals of two adjacent superconducting rotor magnets and performs differential processing; the interlocking trigger condition is set as two quench detection units monitoring the terminal voltage of the same superconducting rotor magnet synchronously triggering the fast quench protection criterion.

[0031] Example 2: This invention also provides a superconducting motor quench detection device based on sliding time filtering, see [link to relevant documentation]. Figure 2 Specifically, it includes: an analog sampling module for differentially processing the terminal voltage signals of adjacent rotor magnets of the superconducting motor and high-speed sampling of the differential voltage signal and excitation current signal; a frequency analysis module for identifying the interference frequency of the differential voltage signal and generating an interference frequency data table; a sliding filter module for establishing a sliding time window, constructing a notch filter based on the interference frequency data table, and filtering the differential voltage signal in real time; a quench feature value extraction module for performing linear regression fitting calculations and extracting quench feature value samples within the sliding time window; and a quench diagnosis module for constructing quench protection action criteria, diagnosing quench events, and issuing quench alarm signals or quench protection action commands.

[0032] Using the above setup, differential processing of the terminal voltage signals of adjacent rotor magnets of the superconducting motor by the analog sampling module can effectively eliminate interference from inductive voltage. Based on this, the sliding filter module utilizes the interference frequency data table pre-built by the frequency analysis module to perform targeted notch filtering on the differential voltage signals, further filtering out specific frequency noises such as excitation power supply ripple and mechanical vibration. This allows for accurate extraction of weak quench characteristic signals even under strong interference. The quench characteristic value extraction module employs a linear regression fitting method to calculate the equivalent quench resistance within sliding time windows of both long and short time scales. R and its rate of change Δ R / Δ t This method can reflect both the instantaneous characteristics of quench occurrence and the cumulative trend of quench resistance changes, effectively overcoming the shortcomings of traditional threshold methods that are susceptible to noise interference and have a single judgment criterion. It improves the accuracy and reliability of quench feature extraction. The quench diagnosis module establishes a dual criterion system that combines fast quench protection criteria and backup quench protection criteria. The fast criterion, in conjunction with interlocking trigger conditions, can respond quickly in the early stage of quench occurrence, meeting the stringent requirements of superconducting motors for the real-time performance of protection actions. The backup criterion provides redundant protection when the fast criterion fails or the signal is abnormal, forming a complete protection mechanism.

[0033] Example 3: Embodiment 3 of the present invention provides a superconducting motor quench detection method based on sliding time filtering, comprising: Step 1: Differential processing of the terminal voltage signals of adjacent superconducting rotor magnets in the superconducting motor is performed using a quench detection device; In practice, the quench detection device includes multiple detection units, each of which simultaneously monitors the voltage signals at the ends of two adjacent superconducting rotor magnets and performs differential processing.

[0034] Step 2: Before the quench test, perform three operation tests on the superconducting motor, including static excitation, no-load rotating excitation, and loaded rotating excitation, and obtain the interference frequency data table; Specifically, in this embodiment, the excitation power supply ripple frequency family {1} in the differential voltage signal is extracted. f sw , 2 f sw ,3 f sw , …}, family of mechanical vibration frequencies {1 f m , 2 f m , 3 f m , …}, and the electromagnetic induction frequency family {1 f e , 2f e , 3 f e , …}, and save it as an interference frequency data table.

[0035] Step 3: Establish sliding time windows with both long and short time scales, and filter the real-time acquired differential voltage signal based on the interference frequency data table; In this embodiment, the differential voltage signal and excitation current signal of adjacent superconducting rotor magnets are acquired in real time, and a short sliding time window is set. T short ,in, T short ≥10 ms, set a long sliding time window T long ,in, T long ≥100ms; Specifically, see Figure 3 In this embodiment, the superconducting motor is energized between 21:50 and 22:25, and the energizing current rises from 0A to 275A. QD6, QD7, and QD8 are the differential voltage signals of adjacent superconducting magnets monitored by the 6th, 7th, and 8th detection units of the quench detection device, respectively. Among them, QD6 and QD7 units detected quench, while the differential voltage signal monitored by QD8 unit remained normal.

[0036] It should be noted that this implementation uses an adaptive notch filter based on an interference frequency data table to achieve fixed-point filtering of discrete interference signals in differential voltage, online tracking of drift, and adding or removing notches as needed. Specifically, the adaptive notch filter uses a high-speed ADC for synchronous sampling, with a sampling rate of over 10 kHz. For interference signals from the excitation power supply ripple frequency family, the notch filter employs a narrow bandwidth, with a quality factor Q set between 60 and 80; for interference signals from the mechanical vibration frequency family and electromagnetic induction frequency family, the bandwidth of the notch filter is widened, and the quality factor Q is reduced to between 20 and 40.

[0037] Step 4: Within the sliding time window, extract effective quench feature values ​​using linear regression fitting. These quench feature values ​​include the equivalent quench resistance. R and equivalent rate of change of superconducting resistance Δ R / Δ t ; In practice, methods for extracting quench feature values ​​include: In this embodiment, when having N Within the sliding time window of a sample sequence ( T = N Δ t ), where the sampling period is fixed at Δ tA time-based first-order linear regression fitting model is established between the filtered differential voltage and the excitation current. The least squares method is used to fit the parameters of the linear model to obtain the equivalent quench characteristic resistance and bias. The parameter fitting calculation is as follows: , in, J ( R , b ) represents the least squares objective function. N The total number of sampling points within the sliding time window, determined by... i [ k [] indicates excitation current data. v [ k [] indicates the filtered differential voltage data. k Represents the sequence number of each discrete sample within the sliding time window ( k =0, 1, 2,…, N ), R This represents the equivalent quench resistance characteristic value obtained from the fitting. b This represents the bias.

[0038] It should be noted that the solution J ( R , b The process of finding the minimum value is related to the equivalent quench resistance. R and bias b We calculate the partial derivatives and set them to zero, and then update them by sampling with each sliding time window. The specific solution results are as follows: , in, This represents the mean of the differential voltage sample sequence within the sliding time window. This is the mean of the excitation current sample sequence within the sliding time window.

[0039] Furthermore, the equivalent quench resistance sample sequence obtained above is analyzed. R [ k Repeat the time-based first-order linear regression fitting to obtain the equivalent quench resistance change rate sample sequence Δ. R [ k ] / Δ t In this embodiment, the solution for the eigenvalue of the equivalent quench resistance change rate is as follows: , in, This represents the mean of the sample indices within the sliding time window. This is the mean of the equivalent overcurrent resistance sample sequence within the sliding time window.

[0040] Specifically, seeFigure 4 This embodiment uses a long sliding time window ( T long Taking 10s as an example, starting from 22:05, the first-order linear regression fitting method is used to extract... Figure 3 The equivalent quench resistance and the rate of change of equivalent quench resistance of the differential voltage signal detected by the QD6 detection unit effectively detect the quench change trend.

[0041] Step 5: Establish a quench protection criterion based on the quench feature value. The quench protection criterion includes a fast quench protection criterion and a backup quench protection criterion. The threshold of the quench protection criterion can be adaptively adjusted in real time. The fast quench protection criterion is set with interlocking trigger conditions. Specifically, the method for establishing a rapid quench protection criterion based on quench characteristic values ​​is as follows: The equivalent quench resistance and the rate of change of equivalent quench resistance within the long and short sliding time windows at each sampling time are obtained as the quench feature value sample sequence. The mean of the sample of the out-of-range feature values ​​is calculated. μ The variance Var is as follows: , in, N The total number of sampling points within the sliding time window is represented by the subscripts short and long, which correspond to the short and long sliding time windows, respectively. The statistical methods for the mean and variance of the equivalent quench resistance change rate are the same as those for the equivalent quench resistance.

[0042] Construct quench determination thresholds based on equivalent quench resistance and the rate of change of equivalent quench resistance, respectively. R th and γ th as follows: in, R th The threshold value for determining the equivalent quench resistance is [value]. γ th The threshold value for determining the rate of change of the equivalent quench resistance is [value]. λ This is the margin adjustment factor.

[0043] In this embodiment, the process of establishing the rapid quench protection criterion includes: Establish a short sliding time window. T short ≥10 ms, preferably, T short =20ms, with a sampling step size of 0.1ms, sliding extraction N short =200 sets of equivalent quench resistance and equivalent quench resistance change rate characteristic values ​​were used as samples, and the mean of the samples was calculated.μ And variance Var; Construct a quench determination threshold based on equivalent quench resistance and the rate of change of equivalent quench resistance. R th_short and γ th_short ; It should be noted that preset rules are set to prevent the quench protection from failing when a forced excitation gate command is encountered or the excitation controller is saturated. R th_short and γ th_short Adaptive updates need to be paused, and the detection program needs to continue running.

[0044] Specifically, in a short sliding time window T short If the following conditions are met R short > R th_short And Δ R short / Δ t > γ th_short If so, the rapid loss of overrun protection criterion is established.

[0045] In this embodiment, the process of establishing the backup overrun protection criterion includes: Establish a long sliding time window T long ,in, T long ≥100ms, preferably, T long =1s, with a sampling step size of 0.1ms, sliding extraction N long =1000 sets of equivalent quench resistance and equivalent quench resistance change rate characteristic values ​​are used as samples, and the mean of the samples is calculated. μ And variance Var; Construct a quench determination threshold based on equivalent quench resistance and the rate of change of equivalent quench resistance. R th_long and γ th_long ; It should be noted that preset rules are set to prevent the quench protection from failing when a forced excitation gate command is encountered or the excitation controller is saturated. R th_long and γ th_long Adaptive updates need to be paused, and the program should be allowed to continue running.

[0046] Specifically, in a long sliding time window T long If the following conditions are met R long > R th_long And Δ R long / Δ t > γ th_long If so, the backup protection criterion is established.

[0047] In this embodiment, the interlocking trigger condition is set as follows: The quench detection device includes multiple detection units, so that the terminal voltage of each superconducting rotor magnet is detected by two detection units at the same time. The interlocking trigger condition is set to allow two quench detection units monitoring the same superconducting rotor magnet terminal voltage to simultaneously trigger the fast quench protection criterion.

[0048] Step 6: Determine whether the quench protection criterion is met. If so, apply quench protection to the superconducting motor.

[0049] In this embodiment, the quench protection action logic is designed to include, but is not limited to: Fast quench protection action logic: If the fast quench protection criterion of only one quench detection unit is met, a quench alarm signal is output; if the fast quench protection criterion is met and the interlock triggering condition is satisfied, the crowbar protection circuit is immediately triggered to demagnetize. Backup overrun protection action logic: If the backup overrun protection criterion is met, the crowbar protection circuit will be immediately triggered to demagnetize; Specifically, when the system detects the quench protection activation command, the superconducting motor excitation power supply system immediately activates, and the process of the excitation current dropping to 0A is described in [reference needed]. Figure 5 .

[0050] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0051] Optionally, specific examples in this embodiment can refer to the examples described in the above embodiments, and will not be repeated here.

[0052] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0053] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0054] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for detecting quench in a superconducting motor based on sliding time filtering, characterized in that, include: Differential processing is performed on the terminal voltage signals of adjacent superconducting rotor magnets in a superconducting motor to eliminate inductive voltage interference; Before the quench detection, the superconducting motor is subjected to three trial runs, including static excitation, no-load rotary excitation, and loaded rotary excitation. In the three trial operation tests, the interference frequencies coupled in the differential voltage signals were collected to form an interference frequency data table, wherein the interference frequencies include the excitation power supply ripple frequency, the rotating machinery vibration frequency, and the electromagnetic induction frequency. During the actual operation of the superconducting motor, the differential voltage signal and excitation current signal of adjacent superconducting rotor magnets are collected in real time. Establish a sliding time window with two time scales, long and short. Within the sliding time window, based on the interference frequency data table, perform notch filtering on the real-time acquired differential voltage signal. The sliding time window includes a short time window of tens of milliseconds and a long time window of hundreds of milliseconds. Within the sliding time window, effective quench feature values ​​are extracted from the filtered differential voltage signal using a linear regression fitting method. These quench feature values ​​include the equivalent quench resistance. R and equivalent rate of change of superconducting resistance Δ R / Δ t ; A quench protection criterion is established based on the quench feature value, wherein the quench protection criterion includes a fast quench protection criterion and a backup quench protection criterion, and the thresholds of the fast quench protection criterion and the backup quench protection criterion are adaptively adjusted in real time, and the fast quench protection criterion is set with an interlocking trigger condition. If the rapid overrun protection criterion is met and the interlocking triggering condition is satisfied, or if the backup overrun protection criterion is met, then the superconducting motor is protected against overrun.

2. The superconducting motor quench detection method based on sliding time filtering according to claim 1, characterized in that, The superconducting motor quench detection method also includes: Real-time monitoring of the operating status of the superconducting motor; If the superconducting motor is detected to be in a short-term overload operation under forced excitation, a quench protection criterion under forced excitation is established based on preset rules.

3. The superconducting motor quench detection method based on sliding time filtering according to claim 1, characterized in that, The method for extracting quench feature values ​​includes: Within the sliding time window, a time-based first-order linear regression fitting calculation is performed on the filtered differential voltage signal and excitation current signal to extract the real-time sliding updated equivalent quench resistance sample sequence. R [ k ]; For the equivalent supersonic resistance sample sequence R [ k Repeat the time-based linear regression fitting calculation to extract the equivalent quench resistance change rate sample sequence Δ. R [ k ] / Δ t .

4. The superconducting motor quench detection method based on sliding time filtering according to claim 3, characterized in that, The time-based first-order linear regression fitting calculation method includes: Within the sliding time window, a first-order linear fitting model is established between the filtered differential voltage and the excitation current. The linear model is fitted with parameters using the least squares method to extract equivalent hysteresis features and biases. The specific parameter fitting method is as follows: ， in, J ( R , b ) represents the least squares objective function. N This represents the total number of sampling points within the sliding time window. i [ k [] indicates excitation current data. v [ k [] indicates the filtered differential voltage data. k Let be the index of each discrete sample within the sliding time window. R This represents the equivalent quench resistance characteristic value obtained from the fitting. b This represents the bias.

5. The superconducting motor quench detection method based on sliding time filtering according to claim 4, characterized in that, The method for establishing a fast quench protection criterion based on quench characteristic values ​​is as follows: Set a short sliding time window T short ,in, T short ≥10 ms; In the short sliding time window T short Within, obtain the updated equivalent quench resistance at each sampling time. R short and equivalent rate of change of superconducting resistance Δ R short / Δ t , as the quench feature value sample sequence; The mean of the sample sequence of the out-of-gain characteristic values ​​is calculated. μ Sum of variance Var, construct the quench threshold. R th_short and γ th_short ; In a short sliding time window T short If the following conditions are met R short > R th_short And Δ R short / Δ t > γ th_short If so, the rapid loss of overrun protection criterion is established.

6. The superconducting motor quench detection method based on sliding time filtering according to claim 4, characterized in that, The method for establishing a backup quench protection criterion based on quench characteristic values ​​is as follows: Set long sliding time window T long ,in, T long ≥100 ms; In the long sliding time window T long Within, obtain the updated equivalent quench resistance at each sampling time. R long and equivalent rate of change of superconducting resistance Δ R long / Δ t , as the quench feature value sample sequence; The mean of the sample sequence of the out-of-gain characteristic values ​​is calculated. μ Sum of variance Var, construct the quench threshold. R th_long and γ th_long ; In a long sliding time window T long If the following conditions are met R long > R th_long And Δ R long / Δ t > γ th_long If so, the backup protection criterion is established.

7. The superconducting motor quench detection method based on sliding time filtering according to claims 5 and 6, characterized in that, The mean of the statistical out-of-range feature value sample sequence μ The specific methods for calculating variance (Var) are as follows: ， in, N This represents the total number of sampling points within the sliding time window. k Let be the index of each discrete sample within the sliding time window. r [ k [This refers to the equivalent quench resistance sample sequence extracted at each sampling time point within the sliding time window.] R [ k Or equivalent quench resistance change rate sample sequence Δ R [ k ] / Δ t .

8. The superconducting motor quench detection method based on sliding time filtering according to claim 7, characterized in that, The construction of the fail-out determination threshold R th and γ th The specific method is as follows: ， in, R th The threshold value for determining the equivalent quench resistance is [value]. γ th The threshold value for determining the rate of change of the equivalent quench resistance is [value]. λ This is the margin adjustment factor.

9. The superconducting motor quench detection method based on sliding time filtering according to claim 1, characterized in that, The interlocking triggering condition is specifically as follows: The quench detection device includes multiple detection units, each of which simultaneously monitors the terminal voltage signals of two adjacent superconducting rotor magnets and performs differential processing. The interlocking trigger condition is set to allow two quench detection units monitoring the same superconducting rotor magnet terminal voltage to simultaneously trigger the fast quench protection criterion.

10. A superconducting motor quench detection device based on sliding time filtering, used to implement the superconducting motor quench detection method based on sliding time filtering as described in any one of claims 1-9, characterized in that, Specifically, it includes: The analog sampling module is used to perform differential processing on the terminal voltage signals of adjacent rotor magnets of the superconducting motor, and to sample the differential voltage signal and excitation current signal at high speed. The frequency analysis module is used to identify the interference frequencies of the differential voltage signal and generate an interference frequency data table. The sliding filter module is used to establish a sliding time window, construct a notch filter based on the interference frequency data table, and filter the differential voltage signal in real time. The quench feature extraction module is used to perform linear regression fitting calculations and extract quench feature sample within the sliding time window; The quench failure diagnosis module is used to construct quench failure protection action criteria, diagnose quench failure events, and issue quench failure alarm signals or quench failure protection action commands.

Images