Generator bearing monitoring method and device, electronic equipment and storage medium

Abstract

The invention discloses a generator bearing monitoring method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring current vibration data of a generator bearing at a current operation moment; performing frequency band division processing on the current vibration data to obtain current frequency band division vibration data of different frequency bands, and determining a root-mean-square observation value of the current frequency band division vibration data; different fitting modes are adopted in different frequency bands, and a root-mean-square predicted value of the current sub-frequency-band vibration data is obtained through fitting based on root-mean-square observed values of historical sub-frequency-band vibration data at multiple historical operation moments; determining a current statistical index based on a current residual matrix formed by the root-mean-square observed value and the root-mean-square predicted value and a historical residual matrix of a plurality of historical operation moments; and according to the current statistical index and the statistical index threshold value, vibration abnormity monitoring of the generator bearing is carried out. According to the scheme, the false alarm rate can be reduced, and the accuracy of generator bearing vibration abnormity monitoring is improved.

Description

Technical Field

[0001] The embodiments of the present invention relate to the technical field of generator monitoring, and in particular, to a method, device, electronic device and storage medium for monitoring a generator bearing. Background Art

[0002] The generator bearing is one of the key components of a wind turbine unit, and its operating state directly affects the safety of the entire wind turbine unit. It is necessary to detect the abnormal conditions of the generator bearing in time through vibration monitoring, so as to discover potential faults in advance and ensure the safe and stable operation of the wind turbine unit.

[0003] In the prior art, the root mean square (RMS) value corresponding to the vibration data of the generator bearing can be determined, and the vibration monitoring of the generator bearing can be directly carried out based on the RMS value. For example, an abnormal warning is given when the RMS value exceeds the recommended threshold in the set standard. The recommended threshold in the set standard is applicable to wind turbine units with a capacity not exceeding 3 MW, but most wind turbine units have exceeded 3 MW, and it is no longer appropriate to continue using this set standard. Another example is to perform abnormal monitoring based on the RMS value through the normal distribution hypothesis. Data outside the range of the mean ± 3 times the standard deviation is considered an abnormal value. However, the RMS value itself does not belong to the normal distribution, and there may be cases of abnormal false alarms. Summary of the Invention

[0004] The present invention provides a method, device, electronic device and storage medium for monitoring a generator bearing, which can improve the accuracy of vibration abnormal monitoring of the generator bearing.

[0005] In a first aspect, an embodiment of the present invention provides a method for monitoring a generator bearing, including:

[0006] Obtaining current vibration data of the generator bearing at the current operating moment;

[0007] Performing frequency band division processing on the current vibration data to obtain current frequency band division vibration data in different frequency bands, and determining the root mean square observation value of the current frequency band division vibration data;

[0008] Adopting different fitting methods in different frequency bands, and fitting based on the root mean square observation values of the historical frequency band division vibration data at multiple historical operating moments to obtain the root mean square prediction value of the current frequency band division vibration data;

[0009] Determining a current statistical index based on the current residual matrix formed by the root mean square observation value and the root mean square prediction value, and the historical residual matrices at multiple historical operating moments;

[0010] Performing vibration abnormal monitoring of the generator bearing according to the current statistical index and the statistical index threshold.

[0011] Second aspect, an embodiment of the present invention provides a generator bearing monitoring device, including:

[0012] An acquisition module, configured to acquire current vibration data of a generator bearing at a current operation moment;

[0013] A processing module, configured to perform frequency band division processing on the current vibration data to obtain current frequency band division vibration data of different frequency bands, and determine a root mean square observation value of the current frequency band division vibration data;

[0014] A fitting module, configured to adopt different fitting methods in different frequency bands, and fit based on root mean square observation values of historical frequency band division vibration data at multiple historical operation moments to obtain a root mean square prediction value of the current frequency band division vibration data;

[0015] A determination module, configured to determine a current statistical index based on a current residual matrix formed by the root mean square observation value and the root mean square prediction value, and historical residual matrices at multiple historical operation moments;

[0016] A monitoring module, configured to perform vibration anomaly monitoring on the generator bearing according to the current statistical index and a statistical index threshold.

[0017] Third aspect, an embodiment of the present invention provides an electronic device, including:

[0018] At least one processor; and

[0019] A memory communicatively connected to the at least one processor; wherein,

[0020] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the method described in the first aspect.

[0021] Fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the method described in the first aspect is implemented.

[0022] The technical solution of the embodiment of the present invention processes the current vibration data of the generator bearing in different frequency bands to obtain the current sub-band vibration data in different frequency bands, and determines the root mean square observation value of the current sub-band vibration data; different fitting methods are used in different frequency bands to determine the root mean square prediction value of the current sub-band vibration data; based on the current residual matrix of the root mean square observation value and the root mean square prediction value of the current sub-band vibration data, and multiple historical residual matrices, the current statistical index is determined; the vibration anomaly monitoring of the generator bearing is carried out according to the current statistical index and the statistical index threshold. This solution can monitor the vibration data in different frequency bands at the same time, determine the statistical index according to the residual fusion of the root mean square observation value and the root mean square prediction value of the vibration data in each frequency band, and carry out vibration anomaly monitoring based on the determined statistical index. Compared with the technical solution that does not divide frequency bands and directly monitors vibration anomalies based on the root mean square value, it can effectively reduce the false alarm rate and improve the accuracy of vibration anomaly monitoring of the generator bearing.

[0023] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present invention, nor is it used to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. Brief Description of the Drawings

[0024] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0025] Figure 1 is a flowchart of a method for monitoring a generator bearing according to Embodiment 1 of the present invention;

[0026] Figure 2 is a flowchart of a method for monitoring a generator bearing according to Embodiment 2 of the present invention;

[0027] Figure 3 is a schematic structural diagram of a device for monitoring a generator bearing according to Embodiment 3 of the present invention;

[0028] Figure 4 is a schematic structural diagram of an electronic device for implementing the embodiments of the present invention. Detailed Embodiments

[0029] In order to enable those skilled in the art to better understand the solution of the present invention, the following will clearly and completely describe the technical solution in the embodiments of the present invention in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0030] It should be noted that the terms "first", "second", etc. in the present invention are used to distinguish similar objects, and do not necessarily need to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances, so that the embodiments of the present invention described here can be implemented in an order other than those illustrated or described here. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0031] Embodiment 1

[0032] Figure 1 It is a flowchart of a method for monitoring a generator bearing according to Embodiment 1 of the present invention. This embodiment is applicable to the situation of monitoring abnormal vibration of a generator bearing. This method can be executed by a generator bearing monitoring device, and the device can be implemented in the form of software and / or hardware and integrated in an electronic device. Further, the electronic device includes but is not limited to: a computer, a laptop, a server, etc.

[0033] As Figure 1 shown, the method includes:

[0034] S110. Obtain the current vibration data of the generator bearing at the current operating moment.

[0035] The generator bearing can be a generator bearing in a wind turbine, which is a key component to ensure the normal operation of the wind turbine. The current vibration data can be the vibration data of the generator bearing at the current operating moment. The vibration data can be data reflecting the vibration state of the generator bearing, and the health status of the generator bearing can be monitored through the vibration data.

[0036] During the operation of the generator bearing, vibration data of the generator bearing at different operating times can be sampled by a hardware sensor at a certain sampling frequency. The hardware sensor can transmit the sampled vibration data to an electronic device, enabling the electronic device to obtain the vibration data at different operating times. In this way, the current vibration data of the generator bearing at the current operating time can be obtained.

[0037] Among them, the hardware sensor can be any sensor capable of obtaining the vibration data of the generator bearing. For example, it can be the hardware sensor used for vibration data acquisition in a Condition Monitoring System (CMS), which is not limited here; the sampling frequency of the hardware sensor is not limited. For example, it can be 12,800 Hz or 25,600 Hz.

[0038] Exemplarily, when the current operating time is the t-th moment, the current vibration data X of the generator bearing at the current operating time t can be expressed as X t ={X t1 , X t2 ,..., X tn}, where n is the number of data included in the current vibration data, and X t1 , X t2 ,..., X tn are the different data included in the current vibration data. According to the different sampling frequencies of the hardware sensor, n can be 12,800 or 25,600.

[0039] S120. Perform band - division processing on the current vibration data to obtain the current band - divided vibration data in different frequency bands, and determine the root - mean - square observation value of the current band - divided vibration data.

[0040] In this step, the current vibration data can be subjected to band - division processing, and the data after band - division processing of the current vibration data is regarded as the current band - divided vibration data. The band - division processing can use any algorithm that can achieve the band - division function, which is not limited here.

[0041] In practical applications, different frequency bands can at least include a low - frequency band, a middle - frequency band, and a high - frequency band. The basis for dividing different frequency bands can be determined according to actual application needs, which is not limited here.

[0042] That is, the current vibration data is band - divided according to the low - frequency band, the middle - frequency band, and the high - frequency band to obtain the current band - divided vibration data in three different frequency bands, which can be expressed by the following formula: X t =S t +T t +R t . Among them, the current band - divided vibration data of the low - frequency band is St = {S t1 , S t2 ,..., S tn}, the current sub - band vibration data in the medium frequency band is T t = {T t1 , T t2 ,..., T tn}, the current sub - band vibration data in the high frequency band is R t = {R t1 , R t2 ,..., R tn}.

[0043] Determine the root - mean - square (RMS) observed values of the current sub - band vibration data respectively. The root - mean - square observed value is the RMS (Root Mean Square) value calculated based on the current sub - band vibration data. The root - mean - square observed value RMSS of the current sub - band vibration data in the low - frequency band t can be expressed as The root - mean - square observed value RMST of the current sub - band vibration data in the medium - frequency band t can be expressed as The root - mean - square observed value RMSR of the current sub - band vibration data in the high - frequency band t can be expressed as

[0044] S130. Adopt different fitting methods in different frequency bands. Based on the root - mean - square observed values of the historical sub - band vibration data at multiple historical operating times, fit to obtain the root - mean - square predicted value of the current sub - band vibration data.

[0045] The historical sub - band vibration data can be the data obtained by dividing the historical vibration data into sub - bands. The root - mean - square observed value of the historical sub - band vibration data can be the RMS value calculated based on the historical sub - band vibration data. The historical vibration data can be the vibration data of the generator bearing at historical operating times. The determination methods of the historical vibration data, historical sub - band vibration data, and root - mean - square observed value of the historical sub - band vibration data at historical operating times can refer to the determination methods of the corresponding data at the current operating time, which will not be elaborated here.

[0046] The root - mean - square predicted value of the current sub - band vibration data can be understood as the RMS value of the current sub - band vibration data predicted based on the root - mean - square observed values of the historical sub - band vibration data at multiple historical operating times.

[0047] In this step, the fitting methods for each frequency band of the low-frequency band, the middle-frequency band, and the high-frequency band can be preset. For example, the fitting method one is used for the low-frequency band, the fitting method two is used for the middle-frequency band, and the fitting method three is used for the high-frequency band. For the current sub-band vibration data of each frequency band, the corresponding fitting method for the frequency band is adopted, and the root mean square observed value of the multiple historical sub-band vibration data of the frequency band is used as the input for fitting, and the root mean square predicted value of the current sub-band vibration data of the frequency band is obtained by fitting.

[0048] There is no limitation on the above fitting methods. For example, it can include but is not limited to the Support Vector Regression (SVR) algorithm, the ensemble learning algorithm, and the probabilistic neural network. Among them, the SVR algorithm can be understood as a supervised learning algorithm based on the support vector machine theory, which is used to solve regression analysis problems. The ensemble learning algorithm can be understood as a machine learning technique that combines multiple weak learners into a strong learner, aiming to improve the generalization ability, stability, and prediction performance of the model. The probabilistic neural network can be understood as a feedforward neural network based on probability density function estimation.

[0049] S140. Determine the current statistical index based on the current residual matrix formed by the root mean square observed value and the root mean square predicted value, and the historical residual matrices at multiple historical operating times.

[0050] In this step, for the current sub-band vibration data of each frequency band, the residual between the corresponding root mean square predicted value and the root mean square observed value can be determined, and the current residual matrix is constructed according to the residuals corresponding to the current sub-band vibration data of each frequency band. Among them, the current residual matrix can be the residual matrix formed by the root mean square observed value and the root mean square predicted value of the current sub-band vibration data, and can be expressed as follows:

[0051]

[0052] Among them, the current residual matrix P t may include the following elements: the root mean square predicted value of the current sub-band vibration data of the low-frequency band and the residual of the root mean square observed value RMSS t the root mean square predicted value of the current sub-band vibration data of the middle-frequency band and the residual of the root mean square observed value RMST the root mean square predicted value of the current sub-band vibration data of the high-frequency band t and the residual of the root mean square observed value RMSR The root mean square predicted value of the current sub-band vibration data of the high-frequency band and the residual of the root mean square observed value RMSR t The current residual matrix P ~N3(0,Σ) can be understood as P t t ​Subject to the standard three-dimensional normal distribution, and Σ is a matrix of all 1s.

[0053] The historical residual matrix can be a residual matrix formed by the root mean square observation value and the root mean square predicted value of the historical vibration data in frequency bands. The determination method can refer to the determination method of the current residual matrix, which will not be elaborated here.

[0054] In this step, the correlation between the current residual matrix and the historical residual matrices at multiple historical operating times can be comprehensively considered to determine the current statistical index. The current statistical index can be the statistical index currently used for abnormal vibration monitoring of the generator bearing. Optionally, the current statistical index can be Hotelling's T 2 statistic. Hotelling's T 2 statistic can be a statistic used for multivariate statistical analysis, which can be understood as a measure of the degree of difference between the sample and the population in multivariate data, that is, a measure of the degree of difference between the current residual matrix and the overall historical residual matrices.

[0055] S150. According to the current statistical index and the statistical index threshold, perform abnormal vibration monitoring of the generator bearing.

[0056] The statistical index threshold can be understood as the threshold used to judge whether the vibration state of the generator bearing is abnormal during the abnormal vibration monitoring of the generator bearing. The statistical index threshold can be determined based on the historical statistical indices at multiple historical operating times. Among them, the historical statistical index can be the statistical index corresponding to the historical operating time, and its determination method can refer to the determination method of the current statistical index, which will not be elaborated here.

[0057] In one embodiment, the statistical index threshold includes an abnormal warning threshold and an abnormal alarm threshold. The abnormal warning threshold is the first multiple of the maximum value among the historical statistical indices at multiple historical operating times, and the abnormal alarm threshold is the second multiple of the maximum value. The first multiple is less than the second multiple. Exemplarily, if the maximum value among the multiple historical statistical indices is Hmax, the abnormal warning threshold can be 2 times Hmax, and the abnormal alarm threshold can be 5 times Hmax.

[0058] In this step, compare the current statistical index with the statistical index threshold to perform abnormal vibration monitoring of the generator bearing. Specifically, if the current statistical index exceeds the abnormal warning threshold, perform an abnormal warning for the vibration of the generator bearing to remind relevant personnel to pay attention to the abnormality in time; if the current statistical index exceeds the abnormal alarm threshold, perform an abnormal alarm for the vibration of the generator bearing to remind relevant personnel that the vibration of the generator bearing has exceeded the safe range and there may be faults or potential hazards, which need to be dealt with in time.

[0059] In the technical solution of the embodiment of the present invention, the current vibration data of the generator bearing is processed by frequency bands to obtain the current vibration data of different frequency bands, and the root mean square observation value of the current vibration data of different frequency bands is determined; different fitting methods are used in different frequency bands to determine the root mean square prediction value of the current vibration data of different frequency bands; based on the current residual matrix of the root mean square observation value and the root mean square prediction value of the current vibration data of different frequency bands, and multiple historical residual matrices, the current statistical index is determined; according to the current statistical index and the statistical index threshold, the vibration abnormality monitoring of the generator bearing is carried out. This solution can monitor the vibration data of different frequency bands at the same time, determine the statistical index according to the residual fusion of the root mean square observation value and the root mean square prediction value of the vibration data of each frequency band, and carry out vibration abnormality monitoring based on the determined statistical index. Compared with the technical solution that does not divide frequency bands and directly monitors vibration abnormalities based on the root mean square value, it can effectively reduce the false alarm rate and improve the accuracy of vibration abnormality monitoring of the generator bearing.

[0060] Embodiment 2

[0061] Figure 2 It is a flowchart of a method for monitoring a generator bearing provided according to Embodiment 2 of the present invention. This embodiment is a further refinement of the above Embodiment 1 in terms of processing the current vibration data by frequency bands to obtain the current vibration data of different frequency bands; and a further refinement of using different fitting methods in different frequency bands to fit and obtain the root mean square prediction value of the current vibration data of different frequency bands based on the root mean square observation values of the historical vibration data of different frequency bands at multiple historical operating times; and a further refinement of determining the current statistical index based on the current residual matrix formed by the root mean square observation value and the root mean square prediction value, and the historical residual matrices at multiple historical operating times.

[0062] As Figure 2 shown, the method includes:

[0063] S110. Obtain the current vibration data of the generator bearing at the current operating time.

[0064] S121. Through the time series decomposition algorithm, process the current vibration data by frequency bands to obtain the current vibration data of different frequency bands, and the different frequency bands at least include a low frequency band, a middle frequency band, and a high frequency band.

[0065] In this step, the current vibration data X t ={X t1 , X t2 ,..., X tn} can be processed by frequency bands through the Seasonal and Trend decomposition using Loess (STL) algorithm to obtain the current vibration data of the low frequency band as St = {S t1 , S t2 ,..., S tn}, the current sub - band vibration data in the middle frequency band is T t = {T t1 , T t2 ,..., T tn} and the current sub - band vibration data in the high frequency band is R t = {R t1 , R t2 ,..., R tn}. Among them, the time - series decomposition algorithm can be an algorithm for processing the current vibration data in sub - bands.

[0066] S122. Determine the root - mean - square observation value of the current sub - band vibration data.

[0067] In this step, the root - mean - square observation value of the current sub - band vibration data in the low - frequency band is The root - mean - square observation value of the current sub - band vibration data in the middle - frequency band is The root - mean - square observation value of the current sub - band vibration data in the high - frequency band is

[0068] S131. Through the support vector regression algorithm, using the root - mean - square observation values of the historical sub - band vibration data in the low - frequency band at multiple historical operating times as input, fit to obtain the root - mean - square prediction value of the current sub - band vibration data in the low - frequency band.

[0069] Among them, the root - mean - square observation values of the historical sub - band vibration data in the low - frequency band at multiple historical operating times can be expressed as RMSS[[ID=4l]] t-τ ,..., RMSS t-τ-g , where t - τ,..., t - τ - g are different multiple historical operating times, τ can be 1, and g can be understood as the time - series step size.

[0070] In this step, through the support vector regression SVR algorithm, using RMSS t-τ ,..., RMSS t-τ-g as input, fit to obtain the root - mean - square prediction value of the current sub - band vibration data in the low - frequency band, which can be expressed as ).

[0071] S132. Through the ensemble learning algorithm, using the root - mean - square observation values of the historical sub - band vibration data in the middle - frequency band at multiple historical operating times as input, fit to obtain the root - mean - square prediction value of the current sub - band vibration data in the middle - frequency band.

[0072] Among them, the root - mean - square observation values of the historical sub - band vibration data in the middle - frequency band at multiple historical operating times can be expressed as RMSTt-τ ,...,RMST t-τ-g , t - τ,..., t - τ - g are different historical operating moments, where τ can be 1 and g can be understood as the time series step size.

[0073] In one embodiment, the first layer of the ensemble learning algorithm consists of a support vector regression algorithm, a K - Nearest Neighbors (KNN) algorithm, and a random forest (RF) algorithm, and the second layer of the ensemble learning algorithm consists of a gradient boosting decision tree (GBDT) algorithm. The output of the first layer is used as the input of the second layer.

[0074] Among them, the K - Nearest Neighbors (KNN) algorithm can be understood as a classification and regression algorithm for nearest neighbor search and belongs to supervised learning. The random forest (RF) algorithm can be understood as an ensemble learning algorithm based on decision trees. It constructs multiple decision trees and combines the results of these decision trees to perform classification or regression. The gradient boosting decision tree (GBDT) algorithm can be understood as an iterative decision tree ensemble learning algorithm based on the gradient boosting framework.

[0075] In this step, using RMST t-τ ,...,RMST t-τ-g as the input of each algorithm (SVR, KNN, RF) included in the first layer of the ensemble learning algorithm, and using the output of the first layer as the input of the second - layer GBDT algorithm, the root - mean - square prediction value of the current sub - band vibration data in the middle frequency band is obtained by fitting, which can be expressed as:

[0076]

[0077] To improve the generalization of the ensemble learning algorithm, 5% cross - validation (5th - fold) is performed in the ensemble learning. Through the cross - validation, the outputs of the three algorithms in the first layer can be closer to the expected output.

[0078] S133. Using a probabilistic neural network, with the root - mean - square observed values of the historical sub - band vibration data in the high - frequency band at multiple historical operating moments as the input, the root - mean - square prediction value of the current sub - band vibration data in the high - frequency band is obtained by fitting.

[0079] Among them, the root - mean - square observed values of the historical sub - band vibration data in the high - frequency band at multiple historical operating moments can be expressed as RMSR t-τ ,...,RMSR t-τ-g , t - τ,..., t - τ - g are different historical operating moments, where τ can be 1 and g can be understood as the time series step size.

[0080] In this step, through the probabilistic neural network, using RMSR t-τ ,..., RMSR t-τ-g as the input, the root mean square prediction value of the current sub-band vibration data in the high frequency band is obtained by fitting where F represents the activation function, b i represents the bias of the neuron in the probabilistic neural network, l i represents the weight of the neuron in the probabilistic neural network, and F~N(0,1) indicates that the result after the function F transformation satisfies the standard normal distribution.

[0081] S141. Determine the mean and standard deviation corresponding to the historical residual matrices at multiple historical operation times.

[0082] where the historical residual matrices at multiple historical operation times can be expressed as P j , j can be from 1 to M, representing M different historical residual matrices, and there is no limit on M. The determination method of P j is basically the same as that of the current residual matrix P t , and will not be elaborated here.

[0083] In this step, the mean corresponding to the historical residual matrices at multiple historical operation times can be expressed as The standard deviation can be expressed as Σ is a matrix of all 1s.

[0084] S142. Based on the current residual matrix formed by the root mean square observed value and the root mean square prediction value, and combining the mean and the standard deviation, determine the current statistical index.

[0085] In this step, based on the current residual matrix P t , the above mean and the standard deviation S p , construct the Hotelling’s T 2 statistic as the current statistical index.

[0086] In one embodiment, based on the current residual matrix formed by the root mean square observed value and the root mean square prediction value, and combining the mean and the standard deviation, determining the current statistical index includes: determining the difference between the current residual matrix and the mean; multiplying the transpose of the difference, the inverse of the standard deviation, and the difference in sequence to obtain the current statistical index. That is, the current statistical index H can be determined by the following formula:

[0087] S150. Based on the current statistical index and the statistical index threshold, perform vibration anomaly monitoring on the generator bearing.

[0088] The technical solution of the embodiment of the present invention decomposes the current vibration data into current sub-band vibration data in the low-frequency band, medium-frequency band, and high-frequency band through the STL algorithm, and calculates the root mean square observation value of each current sub-band vibration data; uses the SVR algorithm for fitting in the low-frequency band, uses the ensemble learning algorithm for fitting in the medium-frequency band, and uses the probabilistic neural network algorithm for fitting in the high-frequency band to obtain the root mean square prediction value of each current sub-band vibration data; calculates the residual between the root mean square prediction value and the root mean square observation value to form a third-order residual matrix, and uses Hotelling’s T 2 index for fusion and threshold determination, monitors multiple frequency bands simultaneously, reduces the false alarm rate, and improves the early warning lead time.

[0089] Embodiment III

[0090] Figure 3 is a schematic structural diagram of a generator bearing monitoring device provided according to Embodiment III of the present invention. This embodiment is applicable to the situation of vibration anomaly monitoring of generator bearings, such as Figure 3 shown. The specific structure of the device includes:

[0091] An acquisition module 31, configured to acquire the current vibration data of the generator bearing at the current operating moment;

[0092] A processing module 32, configured to perform sub-band processing on the current vibration data to obtain current sub-band vibration data in different frequency bands, and determine the root mean square observation value of the current sub-band vibration data;

[0093] A fitting module 33, configured to adopt different fitting methods in different frequency bands, and based on the root mean square observation values of the historical sub-band vibration data at multiple historical operating moments, fit to obtain the root mean square prediction value of the current sub-band vibration data;

[0094] A determination module 34, configured to determine the current statistical index based on the current residual matrix formed by the root mean square observation value and the root mean square prediction value, and the historical residual matrices at multiple historical operating moments;

[0095] A monitoring module 35, configured to perform vibration anomaly monitoring of the generator bearing according to the current statistical index and the statistical index threshold.

[0096] The generator bearing monitoring device provided in this embodiment obtains the current vibration data of the generator bearing at the current operating moment through an acquisition module; processes the current vibration data through a processing module to obtain current sub-band vibration data in different frequency bands, and determines the root mean square observation value of the current sub-band vibration data; uses different fitting methods in different frequency bands through a fitting module, and fits the root mean square prediction value of the current sub-band vibration data based on the root mean square observation values of the historical sub-band vibration data at multiple historical operating moments; determines the current statistical index through a determination module based on the current residual matrix formed by the root mean square observation value and the root mean square prediction value, and the historical residual matrices at multiple historical operating moments; and monitors the vibration abnormality of the generator bearing through a monitoring module according to the current statistical index and the statistical index threshold. This solution can simultaneously monitor the vibration data in different frequency bands, determine the statistical index based on the residual fusion of the root mean square observation value and the root mean square prediction value of the vibration data in each frequency band, and perform vibration abnormality monitoring based on the determined statistical index. Compared with the technical solution that does not divide the frequency band and directly performs vibration abnormality monitoring based on the root mean square value, it can effectively reduce the false alarm rate and improve the accuracy of vibration abnormality monitoring of the generator bearing.

[0097] Further, the processing module 32 is specifically configured to:

[0098] Perform sub-band processing on the current vibration data through a time series decomposition algorithm to obtain current sub-band vibration data in different frequency bands, and the different frequency bands at least include a low frequency band, a middle frequency band, and a high frequency band.

[0099] Further, the fitting module 33 is specifically configured to:

[0100] Use a support vector regression algorithm, take the root mean square observation values of the historical sub-band vibration data in the low frequency band at multiple historical operating moments as inputs, and fit the root mean square prediction value of the current sub-band vibration data in the low frequency band;

[0101] Use an ensemble learning algorithm, take the root mean square observation values of the historical sub-band vibration data in the middle frequency band at multiple historical operating moments as inputs, and fit the root mean square prediction value of the current sub-band vibration data in the middle frequency band;

[0102] Use a probabilistic neural network, take the root mean square observation values of the historical sub-band vibration data in the high frequency band at multiple historical operating moments as inputs, and fit the root mean square prediction value of the current sub-band vibration data in the high frequency band.

[0103] Further, the first layer of the ensemble learning algorithm consists of a support vector regression algorithm, a K-nearest neighbor algorithm, and a random forest algorithm, and the second layer of the ensemble learning algorithm consists of a gradient boosting decision tree algorithm, and the output of the first layer is used as the input of the second layer.

[0104] Further, the determination module 34 is specifically configured to:

[0105] Determine the mean and standard deviation corresponding to the historical residual matrices at multiple historical operation times;

[0106] Based on the current residual matrix formed by the root mean square observation value and the root mean square prediction value, and in combination with the mean and the standard deviation, determine the current statistical index.

[0107] Further, the determination module 34 is specifically configured to:

[0108] Determine the difference between the current residual matrix and the mean;

[0109] Multiply the transpose of the difference, the inverse of the standard deviation, and the difference in sequence to obtain the current statistical index.

[0110] Further, the statistical index threshold includes an abnormal warning threshold and an abnormal alarm threshold. The abnormal warning threshold is the first multiple of the maximum value among the historical statistical indexes at multiple historical operation times, and the abnormal alarm threshold is the second multiple of the maximum value. The first multiple is less than the second multiple.

[0111] The generator bearing monitoring device provided by the embodiments of the present invention can execute the generator bearing monitoring method provided by any embodiment of the present invention, and has corresponding functional modules and beneficial effects for executing the method.

[0112] Embodiment 4

[0113] Figure 4 It is a schematic structural diagram of an electronic device for implementing the embodiments of the present invention. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are only examples and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0114] Such as Figure 4As shown, the electronic device 10 includes at least one processor 11 and a memory communicatively connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc. Among them, the memory stores a computer program executable by the at least one processor. The processor 11 can execute various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 into the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0115] Multiple components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disc, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0116] The processor 11 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the processor 11 include but are not limited to a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the generator bearing monitoring method.

[0117] In some embodiments, the generator bearing monitoring method can be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as the storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed onto the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the generator bearing monitoring method described above can be executed. Alternatively, in other embodiments, the processor 11 can be configured to execute the generator bearing monitoring method by any other appropriate means (e.g., by means of firmware).

[0118] The various embodiments of the systems and techniques described above in this document can be implemented in digital electronic circuitry, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems-on-chip (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: being implemented in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be a special-purpose or general-purpose programmable processor that receives data and instructions from a storage system, at least one input device, and at least one output device, and transmits the data and instructions to the storage system, the at least one input device, and the at least one output device.

[0119] The computer programs for implementing the methods of the present invention can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general purpose computer, a special purpose computer, or other programmable data processing apparatus, such that the computer programs, when executed by the processor, cause the functions / operations specified in the flowchart and / or block diagram to be implemented. The computer programs can be executed entirely on the machine, partly on the machine, as a stand-alone software package partly on the machine and partly on a remote machine or entirely on the remote machine or server.

[0120] In the context of the present invention, a computer-readable storage medium can be a tangible medium that can contain or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. The computer-readable storage medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, the computer-readable storage medium can be a machine-readable signal medium. More specific examples of the machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0121] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the electronic device. Other kinds of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0122] The systems and techniques described herein can be implemented in a computing system including backend components (e.g., as a data server), or a computing system including middleware components (e.g., an application server), or a computing system including frontend components (e.g., a user computer having a graphical user interface or a web browser through which the user can interact with an implementation of the systems and techniques described herein), or a computing system including any combination of such backend components, middleware components, or frontend components. The components of the system can be interconnected by digital data communication in any form or medium (e.g., a communication network). Examples of communication networks include: local area network (LAN), wide area network (WAN), blockchain network, and the Internet.

[0123] A computing system can include a client and a server. The client and the server are generally remote from each other and typically interact through a communication network. The client - server relationship is created by computer programs running on the respective computers and having a client - server relationship with each other. The server can be a cloud server, also known as a cloud computing server or a cloud host, which is a host product in the cloud computing service system, solving the defects of difficult management and weak business scalability existing in traditional physical hosts and VPS services.

[0124] It should be understood that various forms of the processes shown above can be used, reordering, adding, or deleting steps. For example, the steps recited in the present invention can be executed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution of the present invention can be achieved, and no limitation is made herein.

[0125] The above - mentioned specific embodiments do not constitute a limitation on the protection scope of the present invention. Those skilled in the art should understand that various modifications, combinations, sub - combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for monitoring a generator bearing, characterized in that, Including: Obtain the current vibration data of the generator bearing at the current operating moment; Perform frequency band division processing on the current vibration data to obtain the current frequency band division vibration data in different frequency bands, and determine the root mean square observation value of the current frequency band division vibration data; Adopt different fitting methods in different frequency bands, and based on the root mean square observation values of the historical frequency band division vibration data at multiple historical operating moments, fit to obtain the root mean square prediction value of the current frequency band division vibration data; Determine the current statistical index based on the current residual matrix formed by the root mean square observation value and the root mean square prediction value, and the historical residual matrices at multiple historical operating moments; Perform vibration anomaly monitoring of the generator bearing according to the current statistical index and the statistical index threshold.

2. The method according to claim 1, wherein Perform frequency band division processing on the current vibration data to obtain the current frequency band division vibration data in different frequency bands, including: Perform frequency band division processing on the current vibration data through a time series decomposition algorithm to obtain the current frequency band division vibration data in different frequency bands, and the different frequency bands at least include a low frequency band, a middle frequency band, and a high frequency band.

3. The method according to claim 1, characterized in that Adopt different fitting methods in different frequency bands, and based on the root mean square observation values of the historical frequency band division vibration data at multiple historical operating moments, fit to obtain the root mean square prediction value of the current frequency band division vibration data, including: Through the support vector regression algorithm, use the root mean square observation values of the historical frequency band division vibration data in the low frequency band at multiple historical operating moments as input, and fit to obtain the root mean square prediction value of the current frequency band division vibration data in the low frequency band; Through the ensemble learning algorithm, use the root mean square observation values of the historical frequency band division vibration data in the middle frequency band at multiple historical operating moments as input, and fit to obtain the root mean square prediction value of the current frequency band division vibration data in the middle frequency band; 4. The method according to claim 3, characterized in that Through the probabilistic neural network, use the root mean square observation values of the historical frequency band division vibration data in the high frequency band at multiple historical operating moments as input, and fit to obtain the root mean square prediction value of the current frequency band division vibration data in the high frequency band.

5. The method according to claim 1, wherein The first layer of the ensemble learning algorithm consists of a support vector regression algorithm, a K-nearest neighbor algorithm, and a random forest algorithm, and the second layer of the ensemble learning algorithm consists of a gradient boosting decision tree algorithm. The output of the first layer is used as the input of the second layer. Determine the current statistical index based on the current residual matrix formed by the root mean square observation value and the root mean square prediction value, and the historical residual matrices at multiple historical operating moments, including: Determine the mean and standard deviation corresponding to the historical residual matrices at multiple historical operating moments; 6. The method according to claim 5, characterized in that, According to the current residual matrix formed by the root mean square observation value and the root mean square prediction value, combined with the mean and the standard deviation, determine the current statistical index. According to the current residual matrix formed by the root mean square observation value and the root mean square prediction value, combined with the mean and the standard deviation, determine the current statistical index, including: Determine the difference between the current residual matrix and the mean; Multiply the transpose of the difference, the inverse of the standard deviation, and the difference in sequence to obtain the current statistical index.

7. The method according to claim 1, wherein The statistical index threshold includes an abnormal warning threshold and an abnormal alarm threshold. The abnormal warning threshold is the first multiple of the maximum value among the historical statistical indexes at multiple historical operation times, and the abnormal alarm threshold is the second multiple of the maximum value, where the first multiple is less than the second multiple.

8. A generator bearing monitoring device, characterized in that, Comprising: an acquisition module configured to acquire current vibration data of a generator bearing at a current operation time; a processing module configured to perform frequency band division processing on the current vibration data to obtain current frequency band division vibration data in different frequency bands and determine the root mean square observation value of the current frequency band division vibration data; a fitting module configured to adopt different fitting methods in different frequency bands and fit to obtain the root mean square prediction value of the current frequency band division vibration data based on the root mean square observation values of the historical frequency band division vibration data at multiple historical operation times; a determination module configured to determine a current statistical index based on a current residual matrix formed by the root mean square observation value and the root mean square prediction value, and historical residual matrices at multiple historical operation times; a monitoring module configured to perform vibration abnormality monitoring on the generator bearing according to the current statistical index and the statistical index threshold.

9. An electronic device, characterized in that, Comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the method according to any one of claims 1-7.

10. 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 according to any one of claims 1-7.

Images