Electromagnetic bearing gap detection method based on multi-scale gated recurrent unit
By processing the current data of electromagnetic bearings through a multi-scale gated cyclic unit, the problem of accurate measurement of contact points in electromagnetic bearing clearance detection is solved, achieving efficient and accurate clearance detection.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HUANENG NUCLEAR ENERGY TECH RES INST CO LTD
- Filing Date
- 2023-06-06
- Publication Date
- 2026-07-21
AI Technical Summary
In electromagnetic bearings, especially when the auxiliary bearing material is relatively soft, it is difficult to accurately measure the first contact point where the rotor and the auxiliary bearing just make contact, which affects the accuracy of gap detection.
A multi-scale gated loop unit is used to establish a linear relationship by collecting electromagnet current values and spacing data. The pre-trained multi-scale gated loop unit is used to predict contact points at different scales, and multi-scale fusion is performed to determine the electromagnetic bearing clearance value.
Eliminating the need for repeated trial measurements improves the accuracy and convenience of gap measurement, simplifying the process of electromagnetic bearing gap detection.
Smart Images
Figure CN116538942B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of bearing testing technology, and in particular to an electromagnetic bearing clearance detection method based on a multi-scale gated circulation unit. Background Technology
[0002] Electromagnetic bearings are mechatronic devices that rely on electromagnetic force to support a rotor, requiring an air gap between the stator and rotor. Measuring the gap between the auxiliary bearing and the rotor through gap detection is a crucial test for the proper functioning of electromagnetic bearings. Automatic gap testing requires measuring the relationship between the current and the rotor's eccentric position. Due to the compressive deformation between the rotor and the auxiliary bearing, and especially when the auxiliary bearing material is relatively soft, it is difficult to accurately measure the initial contact point where the rotor and auxiliary bearing just make contact. Summary of the Invention
[0003] This application aims to at least partially address one of the technical problems in the related art.
[0004] Therefore, the first aspect of this application proposes a method for detecting the clearance of electromagnetic bearings based on a multi-scale gated loop unit, comprising:
[0005] A test dataset of the target electromagnetic bearing is collected, which includes multiple electromagnet current values and the spacing between multiple target electromagnetic bearing rotors and auxiliary bearings corresponding to each electromagnet current value; and a linear relationship is established between the electromagnet current values and the spacing between the target electromagnetic bearing rotors and auxiliary bearings.
[0006] The multiple electromagnet current values in the test dataset are used as the first current vector, and the first current vector is downsampled based on multiple sampling rates to obtain a second current vector of multiple scales.
[0007] The second current vectors at multiple scales are input to the corresponding pre-trained multi-scale gated recurrent units to obtain multiple sets of first label vectors; the label value in the first label vector represents whether the position between the target electromagnetic bearing rotor and the auxiliary bearing presents a first contact point under the electromagnet current value in the second current vector; wherein, the multi-scale gated recurrent unit has learned the mapping relationship between the electromagnet current value and the label value at different scales;
[0008] The multiple sets of first label vectors are fused to obtain the target label vector;
[0009] In response to the number of tag values in the target tag vector representing the position of the first contact point between the target electromagnetic bearing rotor and the auxiliary bearing satisfying a preset condition, the gap value of the target electromagnetic bearing is obtained based on the target tag vector, the linear relationship, and the distance between the multiple target electromagnetic bearing rotors and the auxiliary bearing corresponding to each of the electromagnet current values.
[0010] The second aspect of this application proposes a training method for multi-scale gated recurrent units, including:
[0011] A training dataset for an electromagnetic bearing is obtained. The training dataset includes multiple electromagnet current values, the spacing between multiple electromagnetic bearing rotors and auxiliary bearings corresponding to each electromagnet current value, and the real label value corresponding to each electromagnet current value. The real label value represents whether the actual position between the electromagnetic bearing rotor and auxiliary bearing presents a first contact point under different electromagnet current values.
[0012] The multiple electromagnet current values in the training dataset are used as the first current vector, and the real label values corresponding to each electromagnet current value in the training dataset are used as the first label vector. The first current vector and the first label vector are downsampled based on multiple sampling rates to obtain a second current vector and a second label vector of multiple scales.
[0013] The second current vectors of various scales are input to the corresponding multi-scale gated loop unit to obtain multiple sets of training label vectors; the training label values in the training label vectors represent whether the position between the electromagnetic bearing rotor and the auxiliary bearing presents a first contact point under different electromagnet current values in the second current vector.
[0014] The corresponding multi-scale gated recurrent unit is trained based on the second label vector corresponding to the second current vector at the multiple scales and the training label vector.
[0015] The third aspect of this application proposes an electromagnetic bearing clearance detection device based on a multi-scale gated loop unit, comprising:
[0016] The first acquisition module is used to collect a test dataset of the target electromagnetic bearing. The test dataset includes multiple electromagnet current values and the spacing between multiple target electromagnetic bearing rotors and auxiliary bearings corresponding to each electromagnet current value; and to establish a linear relationship between the electromagnet current values and the spacing between the target electromagnetic bearing rotors and auxiliary bearings.
[0017] The second acquisition module is used to take the multiple electromagnet current values in the test dataset as a first current vector, and perform downsampling processing on the first current vector based on multiple sampling rates to obtain a second current vector of multiple scales.
[0018] The third acquisition module is used to input the second current vectors of the multiple scales into the corresponding pre-trained multi-scale gated loop unit to obtain multiple sets of first label vectors; the label value in the first label vector represents whether the position between the target electromagnetic bearing rotor and the auxiliary bearing presents a first contact point under the electromagnet current value in the second current vector; wherein, the multi-scale gated loop unit has learned the mapping relationship between the electromagnet current value and the label value at different scales;
[0019] The fourth acquisition module is used to fuse the multiple sets of first label vectors to obtain the target label vector;
[0020] The fifth acquisition module, in response to the number of tag values in the target tag vector representing the position of the first contact point between the target electromagnetic bearing rotor and the auxiliary bearing satisfying a preset condition, is used to obtain the gap value of the target electromagnetic bearing based on the target tag vector, the linear relationship, and the distance between the multiple target electromagnetic bearing rotors and the auxiliary bearing corresponding to each of the electromagnet current values.
[0021] The fourth aspect of this application proposes a training device for a multi-scale gated recurrent unit, comprising:
[0022] The first acquisition module is used to acquire a training dataset of electromagnetic bearings. The training dataset includes multiple electromagnet current values, the distance between multiple electromagnetic bearing rotors and auxiliary bearings corresponding to each electromagnet current value, and the real label value corresponding to each electromagnet current value. The real label value represents whether the actual position between the electromagnetic bearing rotor and auxiliary bearing presents a first contact point under different electromagnet current values.
[0023] The second acquisition module is used to take the multiple electromagnet current values in the training dataset as a first current vector, take the real label values corresponding to each electromagnet current value in the training dataset as a first label vector, and perform downsampling processing on the first current vector and the first label vector based on multiple sampling rates to obtain a second current vector and a second label vector of multiple scales.
[0024] The third acquisition module is used to input the second current vectors of multiple scales into the corresponding multi-scale gated loop unit to obtain multiple sets of training label vectors; the training label values in the training label vectors represent whether the position between the electromagnetic bearing rotor and the auxiliary bearing presents a first contact point under different electromagnet current values in the second current vector.
[0025] The training module is used to train the corresponding multi-scale gated recurrent unit based on the second label vector corresponding to the second current vector at the multiple scales and the training label vector.
[0026] The fifth aspect of this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, when the processor executes the program, it implements the method described in the first aspect above, or implements the method described in the second aspect above.
[0027] The sixth aspect of this application provides a computer-readable storage medium having a computer program stored thereon, characterized in that the program, when executed by a processor, implements the method described in the first aspect above, or implements the method described in the second aspect above.
[0028] The electromagnetic bearing clearance detection method based on a multi-scale gated cyclic unit according to embodiments of this application transforms the traditional mechanical measurement problem into a signal sequence prediction problem. Specifically, it converts electromagnet current data into vectors and the position between the target electromagnetic bearing rotor and the auxiliary bearing into label vectors indicating whether it is the first contact point. Based on current vectors at different scales, a pre-trained multi-scale gated cyclic unit predicts the first contact point at different scales, and then performs multi-scale fusion to determine the first contact point and obtain the electromagnetic bearing clearance value. This application eliminates the need for repeated trial measurements; by collecting current and clearance data and utilizing an offline model multi-scale gated cyclic unit, it accurately obtains the clearance value, making it more convenient in practical applications and improving the accuracy of clearance measurement results.
[0029] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description
[0030] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:
[0031] Figure 1 This is a schematic diagram illustrating the linear relationship between the electromagnet current value and the distance between the electromagnetic bearing rotor and the auxiliary bearing, as provided in the embodiments of this application.
[0032] Figure 2 This is a schematic flowchart illustrating an electromagnetic bearing clearance detection method based on a multi-scale gated loop unit provided in an embodiment of this application.
[0033] Figure 3 This is a schematic diagram illustrating the prediction of label values using a multi-scale gated recurrent unit, provided in an embodiment of this application.
[0034] Figure 4 A flowchart illustrating a training method for a multi-scale gated recurrent unit provided in an embodiment of this application;
[0035] Figure 5 A schematic diagram of an electromagnetic bearing clearance detection device based on a multi-scale gated loop unit provided in an embodiment of this application;
[0036] Figure 6 This is a schematic diagram of a training device for a multi-scale gated loop unit provided in an embodiment of this application. Detailed Implementation
[0037] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.
[0038] It should be noted that when performing clearance testing on an electromagnetic bearing, the electromagnetic bearing rotor needs to be suspended above the center of the auxiliary bearing by the attraction of an electromagnet. By changing the electromagnet current value, the distance between the electromagnetic bearing rotor and the auxiliary bearing is measured accordingly, establishing a... Figure 1 The diagram shows a linear relationship between the electromagnet current value and the distance between the electromagnetic bearing rotor and the auxiliary bearing. This linear relationship, or measurement using a feeler gauge, allows us to determine whether a given location is the first contact point. The inflection point on the linear curve corresponds to the first contact point between the electromagnetic bearing rotor and the auxiliary bearing.
[0039] However, Figure 1 The linear relationship between the electromagnet current value and the distance between the electromagnetic bearing rotor and the auxiliary bearing shown is a linear relationship under ideal conditions. In actual measurement, due to the extrusion deformation between the electromagnetic bearing rotor and the auxiliary bearing, especially when the auxiliary bearing material is relatively soft, it is difficult to accurately measure the first contact point.
[0040] Therefore, this application proposes a method for detecting the clearance of electromagnetic bearings based on a multi-scale gated loop unit. Specifically, the embodiment of the electromagnetic bearing clearance detection method based on a multi-scale gated loop unit of this application is described below with reference to the accompanying drawings.
[0041] Figure 2 This is a schematic flowchart illustrating an electromagnetic bearing clearance detection method based on a multi-scale gated loop unit, provided in an embodiment of this application. Figure 2 As shown, the method includes the following steps:
[0042] Step 201: Collect the test dataset of the target electromagnetic bearing. The test dataset includes multiple electromagnet current values and the corresponding spacing between the target electromagnetic bearing rotor and auxiliary bearing for each electromagnet current value. Establish a linear relationship between the electromagnet current values and the spacing between the target electromagnetic bearing rotor and auxiliary bearing.
[0043] It should be noted that in equipment supported by electromagnetic bearings, such as the main helium blower and helium compressor in high-temperature gas-cooled reactors, gap detection of the electromagnetic bearings requires opening the equipment end cover and measuring the gap using feeler gauges and dial indicators. This process is time-consuming and labor-intensive, and testing cannot be performed when there is pressure inside the pressure vessel. Therefore, in some embodiments of this application, the distance between multiple target electromagnetic bearing rotors and auxiliary bearings corresponding to the electromagnet current values can be collected by displacement sensors, eliminating the need for disassembly and inspection, and making implementation easier.
[0044] Step 202: Take multiple electromagnet current values in the test dataset as the first current vector, and perform downsampling processing on the first current vector based on multiple sampling rates to obtain a second current vector with multiple scales.
[0045] Optionally, in some embodiments of this application, the first current vector can be downsampled based on multiple sampling rates. In the low sampling rate data, the mean of the original data in the sampling region is used as the second current value corresponding to the sampling region in the second current vector.
[0046] As an example, suppose the test dataset includes 12 electromagnet current values i m Given i = 1, 2, ..., 12, obtain the first current vector I1 = [i1, i2, ..., i 12 According to the sampling rate R r =R1 / r The first current vector I1 is downsampled, where R1 is the number of data points in the first current vector I1, and r is a positive integer. When r = 1, the second current vector is the same as the first current vector I1. When r > 1, the second current vector obtained after downsampling is the low sampling rate data. For example, when r = 2, i1-i2 is one sampling region, i3-i4 is another sampling region, and so on. The mean of the original data in each sampling region at the same sampling rate is used as the second current value corresponding to the sampling region in the second current vector of the corresponding scale.
[0047] Step 203 involves inputting second current vectors at multiple scales into corresponding pre-trained multi-scale gated recurrent units to obtain multiple sets of first label vectors. The label value in each first label vector represents whether the target electromagnetic bearing rotor and the auxiliary bearing exhibit a first contact point under the electromagnet current value in the second current vector. The multi-scale gated recurrent unit has already learned the mapping relationship between electromagnet current values and label values at different scales.
[0048] It should be noted that each scale corresponds to a multi-scale gated loop unit, which has learned the mapping relationship between the electromagnet current value and the tag value at the corresponding scale.
[0049] In some embodiments of this application, such as Figure 3 As shown, if the multi-scale gated loop unit predicts that the position between the target electromagnetic bearing rotor and the auxiliary bearing presents a first contact point based on the second current vector of the corresponding scale, then the label value in the first label vector is 1. If the predicted position between the target electromagnetic bearing rotor and the auxiliary bearing does not present a first contact point, then the label value in the first label vector is 0.
[0050] It should also be noted that the training method of the multi-scale gated recurrent unit in this embodiment can be found in the description of the subsequent embodiments of this application, and will not be repeated here.
[0051] Step 204: Merge multiple sets of first label vectors to obtain the target label vector.
[0052] In some embodiments of this application, since the multiple sets of first label vectors have different scales, i.e., different data lengths, it is necessary to unify the lengths of the multiple sets of first label vectors and then fuse them to obtain the target label vector. As one possible implementation, multiple sets of first label vectors can be upsampled based on a preset numerical padding principle to obtain multiple sets of second label vectors with the same length as the first current vector. For example, taking a label value of 0 or 1, the numerical padding principle can be: if the label value in the first label vector is 1, then the padding value of the corresponding region is 1. If the label value in the first label vector is 0, then the padding value of the corresponding region is 0.
[0053] Multiple sets of second-label vectors of the same length are merged to obtain the target label vector. The following formula can be used as a reference:
[0054]
[0055] in, For the target label vector, This is the first label vector of group s.
[0056] Step 205: In response to the number of tag values in the target tag vector representing the position of the first contact point between the target electromagnetic bearing rotor and the auxiliary bearing satisfying the preset condition, the gap value of the target electromagnetic bearing is obtained based on the target tag vector, the linear relationship, and the distance between the multiple target electromagnetic bearing rotors and the auxiliary bearings corresponding to each electromagnet current value.
[0057] In some embodiments of this application, the preset condition may be that the number of label values showing the first contact point between the target electromagnetic bearing rotor and the auxiliary bearing is 2. That is, when the number of label values showing the first contact point between the target electromagnetic bearing rotor and the auxiliary bearing is 2, the gap value of the target electromagnetic bearing is obtained based on the target label vector, the linear relationship, and the distance between the multiple target electromagnetic bearing rotors and the auxiliary bearings corresponding to each electromagnet current value; otherwise, steps 201-205 are repeated to re-detect the gap value of the target electromagnetic bearing.
[0058] like Figure 1 As shown, there is a linear relationship between the electromagnet current value and the distance between the target electromagnetic bearing rotor and the auxiliary bearing, and the electromagnet current value in the first current vector corresponds to the label value in the target label vector. Therefore, the distance between the target electromagnetic bearing rotor and the auxiliary bearing can be determined by the label value representing the position of the first contact point between the target electromagnetic bearing rotor and the auxiliary bearing in the target label vector, thereby obtaining the clearance value of the target electromagnetic bearing.
[0059] For example, when the target label vector satisfies the preset condition that the number of label values that present the first contact point between the target electromagnetic bearing rotor and the auxiliary bearing is 2, it means that there are two first contact points between the target electromagnetic bearing rotor and the auxiliary bearing, and the gap value is the difference between the corresponding positions of the two label values that present the first contact points.
[0060] The electromagnetic bearing clearance detection method based on a multi-scale gated cyclic unit according to embodiments of this application transforms the traditional mechanical measurement problem into a signal sequence prediction problem. Specifically, it converts electromagnet current data into vectors and the position between the target electromagnetic bearing rotor and the auxiliary bearing into label vectors indicating whether it is the first contact point. Based on current vectors at different scales, a pre-trained multi-scale gated cyclic unit predicts the first contact point at different scales, and then performs multi-scale fusion to determine the first contact point and obtain the electromagnetic bearing clearance value. This application eliminates the need for repeated trial measurements; by collecting current and clearance data and utilizing an offline model multi-scale gated cyclic unit, it accurately obtains the clearance value, making it more convenient in practical applications and improving the accuracy of clearance measurement results.
[0061] This application also proposes a training method for multi-scale gated recurrent units. Figure 4This is a schematic flowchart illustrating a training method for a multi-scale gated recurrent unit provided in an embodiment of this application. Figure 4 As shown, the method includes the following steps:
[0062] Step 401: Obtain the training dataset for the electromagnetic bearing. The training dataset includes multiple electromagnet current values, the spacing between the electromagnetic bearing rotor and auxiliary bearing corresponding to each electromagnet current value, and the true label value corresponding to each electromagnet current value. The true label value represents whether the actual position between the electromagnetic bearing rotor and auxiliary bearing presents a first contact point under different electromagnet current values.
[0063] The actual label value corresponding to each electromagnet current value is the actual result measured using a feeler gauge. In some embodiments of this application, if the position between the electromagnetic bearing rotor and the auxiliary bearing presents a first contact point, the actual label value is 1. If the position between the electromagnetic bearing rotor and the auxiliary bearing does not present a first contact point, the actual label value is 0.
[0064] Step 402: Take multiple electromagnet current values in the training dataset as the first current vector, take the real label values corresponding to each electromagnet current value in the training dataset as the first label vector, and perform downsampling processing on the first current vector and the first label vector based on multiple sampling rates to obtain second current vectors and second label vectors of multiple scales.
[0065] In some embodiments of this application, the first current vector can be downsampled based on multiple sampling rates. In the low sampling rate data, the mean of the original data in the sampling region is used as the second current value corresponding to the sampling region in the second current vector. Specific implementation details of this part can be found in step 202 of the above embodiments, and will not be repeated here.
[0066] In some embodiments of this application, taking a true label value of 0 or 1 as an example, the first label vector can be downsampled based on multiple sampling rates. In low sampling rate data, if there is a true label value of 1 in the sampling area, the label value corresponding to the sampling area in the second label vector is marked as 1; if all true label values in the sampling area are 0, the label value corresponding to the sampling area in the second label vector is marked as 0.
[0067] Step 403: Input the second current vectors at multiple scales into the corresponding multi-scale gated recurrent units to obtain multiple sets of training label vectors. The training label values in the training label vectors represent whether the position between the electromagnetic bearing rotor and the auxiliary bearing presents a first contact point under different electromagnet current values in the second current vector.
[0068] Each scale corresponds to a multi-scale gated recurrent unit. As an example, by downsampling, the first current vector is multi-scaled to obtain s sets of second current vectors at different scales, i.e., I→{I1,I2,...,I...} s}. Let {I1,I2,...,I... s The inputs are respectively fed into s multi-scale gated recurrent units {GRU1, GRU2, ..., GRU} at the corresponding scale. s In}, obtain s groups of training label vectors.
[0069] Step 404: Train the corresponding multi-scale gated recurrent unit based on the second label vector corresponding to the second current vector at multiple scales and the training label vector.
[0070] Optionally, in some embodiments of this application, multiple sets of tests can be performed on different electromagnetic bearings, multiple sets of training datasets can be collected, and the multi-scale gated loop unit can be trained multiple times to improve the accuracy of the multi-scale gated loop unit in predicting the first contact point between the electromagnetic bearing and the auxiliary bearing.
[0071] The training method for the multi-scale gated loop unit according to embodiments of this application transforms the traditional mechanical measurement problem into a signal sequence prediction problem. Specifically, it converts electromagnet current data into vectors and the position between the target electromagnetic bearing rotor and the auxiliary bearing into label vectors indicating whether it is the first contact point. The multi-scale gated loop unit is trained based on a test dataset, enabling it to learn the mapping relationship between electromagnet current values at different scales and the label values representing whether it is the first contact point, thus accurately predicting the first contact point. Applying the trained multi-scale gated loop unit to the gap detection of electromagnetic bearings can improve both the gap detection efficiency and accuracy.
[0072] Figure 5 This is a schematic diagram of an electromagnetic bearing clearance detection device based on a multi-scale gated loop unit, provided as an embodiment of this application. Figure 5 As shown, the device includes: a first acquisition module 501, a second acquisition module 502, a third acquisition module 503, a fourth acquisition module 504, and a fifth acquisition module 505.
[0073] in,
[0074] The first acquisition module 501 is used to collect the test dataset of the target electromagnetic bearing. The test dataset includes multiple electromagnet current values and the spacing between the target electromagnetic bearing rotor and auxiliary bearing corresponding to each electromagnet current value. A linear relationship is established between the electromagnet current values and the spacing between the target electromagnetic bearing rotor and auxiliary bearing.
[0075] The second acquisition module 502 is used to take multiple electromagnet current values in the test dataset as the first current vector, and perform downsampling processing on the first current vector based on multiple sampling rates to obtain a second current vector of multiple scales.
[0076] The third acquisition module 503 is used to input second current vectors of multiple scales into corresponding pre-trained multi-scale gated recurrent units to obtain multiple sets of first label vectors. The label value in the first label vector represents whether the position between the target electromagnetic bearing rotor and the auxiliary bearing presents a first contact point under the electromagnet current value in the second current vector. The multi-scale gated recurrent unit has learned the mapping relationship between the electromagnet current value and the label value at different scales.
[0077] The fourth acquisition module 504 is used to fuse multiple sets of first label vectors to obtain the target label vector.
[0078] The fifth acquisition module 505, in response to the number of tag values representing the first contact point between the target electromagnetic bearing rotor and the auxiliary bearing in the target tag vector satisfying a preset condition, is used to obtain the gap value of the target electromagnetic bearing based on the target tag vector, the linear relationship, and the distance between multiple target electromagnetic bearing rotors and auxiliary bearings corresponding to each electromagnet current value.
[0079] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.
[0080] Figure 6 This is a schematic diagram of a training device for a multi-scale gated recurrent unit provided in an embodiment of this application. Figure 6 As shown, the device includes: a first acquisition module 601, a second acquisition module 602, a third acquisition module 603, and a training module 604. Among them,
[0081] The first acquisition module 601 is used to acquire a training dataset for the electromagnetic bearing. The training dataset includes multiple electromagnet current values, the spacing between multiple electromagnetic bearing rotors and auxiliary bearings corresponding to each electromagnet current value, and the actual label value corresponding to each electromagnet current value. The actual label value represents whether the actual position between the electromagnetic bearing rotor and auxiliary bearing presents a first contact point under different electromagnet current values.
[0082] The second acquisition module 602 is used to take multiple electromagnet current values in the training dataset as the first current vector, take the real label values corresponding to each electromagnet current value in the training dataset as the first label vector, and perform downsampling processing on the first current vector and the first label vector based on multiple sampling rates to obtain a second current vector and a second label vector of multiple scales.
[0083] The third acquisition module 603 is used to input the second current vector at multiple scales into the corresponding multi-scale gated recurrent unit to obtain multiple sets of training label vectors. The training label values in the training label vectors represent whether the position between the electromagnetic bearing rotor and the auxiliary bearing presents a first contact point under different electromagnet current values in the second current vector.
[0084] Training module 604 is used to train the corresponding multi-scale gated recurrent unit based on the second label vector corresponding to the second current vector at multiple scales and the training label vector.
[0085] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.
[0086] To implement the above embodiments, this application also proposes an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the electromagnetic bearing clearance detection method based on a multi-scale gated loop unit of any of the foregoing embodiments, or the training method of the multi-scale gated loop unit of any of the foregoing embodiments.
[0087] To implement the above embodiments, this application also proposes a computer-readable storage medium storing a computer program that, when executed by a processor, implements the electromagnetic bearing clearance detection method based on a multi-scale gated loop unit of any of the foregoing embodiments, or implements the training method of the multi-scale gated loop unit of any of the foregoing embodiments.
[0088] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.
[0089] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.
[0090] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.
[0091] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.
[0092] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.
[0093] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.
[0094] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.
[0095] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this application.
Claims
1. A method for detecting the clearance of an electromagnetic bearing based on a multi-scale gated loop unit, characterized in that, Includes the following steps: Collect a test dataset of the target electromagnetic bearing, the test dataset including multiple electromagnet current values and the spacing between multiple target electromagnetic bearing rotors and auxiliary bearings corresponding to each of the electromagnet current values; And establish a linear relationship between the electromagnet current value and the distance between the target electromagnetic bearing rotor and the auxiliary bearing; The multiple electromagnet current values in the test dataset are used as the first current vector, and the first current vector is downsampled based on multiple sampling rates to obtain a second current vector of multiple scales. The second current vectors at multiple scales are input to the corresponding pre-trained multi-scale gated recurrent units to obtain multiple sets of first label vectors; the label value in the first label vector represents whether the position between the target electromagnetic bearing rotor and the auxiliary bearing presents a first contact point under the electromagnet current value in the second current vector; wherein, the multi-scale gated recurrent unit has learned the mapping relationship between the electromagnet current value and the label value at different scales; The multiple sets of first label vectors are fused to obtain the target label vector; In response to the number of tag values in the target tag vector representing the position of the first contact point between the target electromagnetic bearing rotor and the auxiliary bearing satisfying a preset condition, the gap value of the target electromagnetic bearing is obtained based on the target tag vector, the linear relationship, and the distance between the multiple target electromagnetic bearing rotors and the auxiliary bearing corresponding to each of the electromagnet current values.
2. The method according to claim 1, characterized in that, The downsampling process of the first current vector based on multiple sampling rates to obtain a second current vector with multiple scales includes: The first current vector is downsampled based on multiple sampling rates. In the low sampling rate data, the mean of the original data in the sampling region is used as the second current value corresponding to the sampling region in the second current vector.
3. The method according to claim 1, characterized in that, The step of fusing the multiple sets of first label vectors to obtain the target label vector includes: Based on a preset numerical filling principle, the multiple sets of first label vectors are upsampled to obtain multiple sets of second label vectors with the same length as the first current vector. Multiple sets of the second label vectors are fused together to obtain the target label vector.
4. The method as described in claim 1, characterized in that, The distance between the target electromagnetic bearing rotor and the auxiliary bearing, which corresponds to the current value of each of the electromagnets, is collected by a displacement sensor.
5. The method as described in claim 1, characterized in that, The preset condition is that the number of label values for the first contact point between the target electromagnetic bearing rotor and the auxiliary bearing is 2.
6. A training method for a multi-scale gated recurrent unit, characterized in that, Includes the following steps: A training dataset for an electromagnetic bearing is obtained. The training dataset includes multiple electromagnet current values, the spacing between multiple electromagnetic bearing rotors and auxiliary bearings corresponding to each electromagnet current value, and the real label value corresponding to each electromagnet current value. The real label value represents whether the actual position between the electromagnetic bearing rotor and auxiliary bearing presents a first contact point under different electromagnet current values. The multiple electromagnet current values in the training dataset are used as the first current vector, and the real label values corresponding to each electromagnet current value in the training dataset are used as the first label vector. The first current vector and the first label vector are downsampled based on multiple sampling rates to obtain a second current vector and a second label vector of multiple scales. The second current vectors of various scales are input to the corresponding multi-scale gated loop unit to obtain multiple sets of training label vectors; the training label values in the training label vectors represent whether the position between the electromagnetic bearing rotor and the auxiliary bearing presents a first contact point under different electromagnet current values in the second current vector. The corresponding multi-scale gated recurrent unit is trained based on the second label vector corresponding to the second current vector at the multiple scales and the training label vector.
7. The method according to claim 6, characterized in that, The method further includes: If the electromagnetic bearing rotor and the auxiliary bearing form a first contact point, the true label value is 1; if the electromagnetic bearing rotor and the auxiliary bearing do not form a first contact point, the true label value is 0.
8. The method according to claim 7, characterized in that, The process of downsampling the first current vector and the first label vector at multiple sampling rates to obtain a second current vector and a second label vector at multiple scales includes: The first current vector is downsampled based on multiple sampling rates. In the low sampling rate data, the mean of the original data in the sampling region is used as the second current value corresponding to the sampling region in the second current vector. The first label vector is downsampled based on multiple sampling rates. In the low sampling rate data, if there is a true label value of 1 in the sampling region, the label value corresponding to the sampling region in the second label vector is marked as 1; if all the true label values in the sampling region are 0, the label value corresponding to the sampling region in the second label vector is marked as 0.
9. An electromagnetic bearing clearance detection device based on a multi-scale gated loop unit, characterized in that, include: The first acquisition module is used to collect a test dataset of the target electromagnetic bearing. The test dataset includes multiple electromagnet current values and the spacing between multiple target electromagnetic bearing rotors and auxiliary bearings corresponding to each electromagnet current value. And establish a linear relationship between the electromagnet current value and the distance between the target electromagnetic bearing rotor and the auxiliary bearing; The second acquisition module is used to take the multiple electromagnet current values in the test dataset as a first current vector, and perform downsampling processing on the first current vector based on multiple sampling rates to obtain a second current vector of multiple scales. The third acquisition module is used to input the second current vectors of the multiple scales into the corresponding pre-trained multi-scale gated loop unit to obtain multiple sets of first label vectors; the label value in the first label vector represents whether the position between the target electromagnetic bearing rotor and the auxiliary bearing presents a first contact point under the electromagnet current value in the second current vector; wherein, the multi-scale gated loop unit has learned the mapping relationship between the electromagnet current value and the label value at different scales; The fourth acquisition module is used to fuse the multiple sets of first label vectors to obtain the target label vector; The fifth acquisition module, in response to the number of tag values in the target tag vector representing the position of the first contact point between the target electromagnetic bearing rotor and the auxiliary bearing satisfying a preset condition, is used to obtain the gap value of the target electromagnetic bearing based on the target tag vector, the linear relationship, and the distance between the multiple target electromagnetic bearing rotors and the auxiliary bearing corresponding to each of the electromagnet current values.
10. A training device for a multi-scale gated recurrent unit, characterized in that, include: The first acquisition module is used to acquire a training dataset of electromagnetic bearings. The training dataset includes multiple electromagnet current values, the spacing between multiple electromagnetic bearing rotors and auxiliary bearings corresponding to each electromagnet current value, and the real label value corresponding to each electromagnet current value. The real label value represents whether the actual position between the electromagnetic bearing rotor and auxiliary bearing presents a first contact point under different electromagnet current values. The second acquisition module is used to take the multiple electromagnet current values in the training dataset as a first current vector, take the real label values corresponding to each electromagnet current value in the training dataset as a first label vector, and perform downsampling processing on the first current vector and the first label vector based on multiple sampling rates to obtain a second current vector and a second label vector of multiple scales. The third acquisition module is used to input the second current vectors of multiple scales into the corresponding multi-scale gated loop unit to obtain multiple sets of training label vectors; the training label values in the training label vectors represent whether the position between the electromagnetic bearing rotor and the auxiliary bearing presents a first contact point under different electromagnet current values in the second current vector. The training module is used to train the corresponding multi-scale gated recurrent unit based on the second label vector corresponding to the second current vector at the multiple scales and the training label vector.
11. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method as described in any one of claims 1-5, or the method as described in any one of claims 6-8.
12. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1-5, or the method as described in any one of claims 6-8.