An equipment operation anomaly detection method and system in water conservancy and hydropower projects

Abstract

The present invention relates to the technical field of wind turbine blade operation detection technology, and specifically to a method and system for detecting equipment operation anomalies in water conservancy and hydropower projects, comprising: obtaining a vibration data sequence of wind turbine blades in a water conservancy and hydropower project to obtain a number of initial vibration bands and their characteristic angles; clustering the initial vibration bands according to the characteristic angles to obtain a number of suspected resonance segments; obtaining the resonance probability of each suspected resonance segment based on the characteristic angle of each initial vibration band within each suspected resonance segment, combined with the change in vibration data within each suspected resonance segment; obtaining a correction value for each vibration data within each suspected resonance segment, and then obtaining a vibration anomaly index for performing anomaly detection. The present invention aims to solve the problem of mistaking normal vibration of the blades under the influence of resonance as abnormal vibration during abnormal detection of wind turbine vibration based on the amplitude of the wind turbine blades.

Description

Technical Field

[0001] The present invention relates to the technical field of wind turbine blade operation detection, and in particular to a method and system for detecting abnormal operation of equipment in water conservancy and hydropower projects. Background Art

[0002] As a key component of clean energy, wind turbines are widely used in water conservancy and hydropower projects to achieve diversified energy utilization. However, wind turbines operate in complex environments, often outdoors, at high altitudes, and in adverse weather conditions, making them susceptible to natural factors such as wind, sand, rain, snow, and lightning. Furthermore, due to the long-term, high-intensity operation of wind turbine blades in complex environments, they often suffer varying degrees of damage, which in turn causes blade vibration. When the damage is severe, the vibration intensity increases, resulting in unnecessary power waste and reduced wind turbine efficiency.

[0003] Currently, vibration sensors placed on target wind turbine blades acquire time-domain vibration data, and by analyzing the amplitude of the vibration data, determine whether the wind turbine blades are experiencing abnormal vibration. However, due to the complex offshore environment in which ocean-type cooling fans are located, the frequency of the complex and ever-changing sea breeze may be the same as or close to the natural frequency of the wind turbine blades at certain times, causing a resonance effect in the wind turbine blades, resulting in a short-term increase in the vibration of the wind turbine blades. As a result, when using existing methods to detect abnormalities in target wind turbine blades, the vibration data caused by the resonance effect is confused with the abnormal vibration data caused by wind turbine blade damage, resulting in inaccurate abnormality detection results. Summary of the Invention

[0004] The present invention provides a method and system for detecting abnormal operation of equipment in a water conservancy and hydropower project, so as to solve the existing problems.

[0005] The present invention provides a method and system for detecting abnormal operation of equipment in a water conservancy and hydropower project using the following technical solutions:

[0006] The present invention proposes a method for detecting abnormal operation of equipment in a water conservancy and hydropower project, the method comprising the following steps:

[0007] Acquire a vibration data sequence of a wind turbine blade in a water conservancy and hydropower project, wherein the vibration data sequence includes a plurality of vibration data corresponding to a collection time and amplitude;

[0008] Based on the change of vibration data in the vibration data sequence, a number of initial vibration bands are obtained, and then a characteristic angle of each initial vibration band is obtained; based on the characteristic angle of each initial vibration band, a normal reference cluster is obtained, and then a number of suspected resonance segments are obtained; the suspected resonance segments are composed of the number of initial resonance bands; based on the difference between the characteristic angle of each initial vibration band in each suspected resonance segment and the characteristic angle of the cluster center point of the normal reference cluster, the vibration outlier of each suspected resonance segment is obtained;

[0009] According to the change of vibration data in each suspected resonance segment, the correlation coefficient of each suspected resonance segment is obtained; according to the correlation coefficient and vibration outlier degree of each suspected resonance segment, the resonance probability of each suspected resonance segment is obtained;

[0010] According to the resonance probability of each suspected resonance segment and the amplitude of each vibration data in each suspected resonance segment, a correction value of each vibration data in each suspected resonance segment is obtained, thereby obtaining a vibration data correction sequence;

[0011] According to the change of the correction value of the vibration data in the vibration data correction sequence, the vibration abnormality index of the wind turbine blade to be detected is obtained, and abnormality detection is performed.

[0012] Furthermore, the method of obtaining a plurality of initial vibration bands based on the change of the vibration data in the vibration data sequence and then obtaining the characteristic angle of each initial vibration band includes the following specific methods:

[0013] Obtaining a minimum value in the vibration data sequence, taking the vibration data between two adjacent minimum values and the vibration data with the shortest acquisition time between the two adjacent minimum values as data within an initial vibration band, and obtaining a plurality of initial vibration bands;

[0014] The first The vibration data in the initial vibration band are mapped to a two-dimensional coordinate system with the horizontal axis being the acquisition time and the vertical axis being the amplitude, and a number of data points are obtained; the vector pointing from the data point with the largest amplitude to the data point with the smallest acquisition time is recorded as the first The first vector of the initial vibration band is the vector that points the data point with the smallest amplitude to the data point with the largest acquisition time, which is recorded as the first vector. The second vector of the initial vibration band; The angle between the first vector and the second vector of the initial vibration band is recorded as The characteristic angle of the initial vibration band.

[0015] Furthermore, the method of obtaining a normal reference cluster based on the characteristic angle of each initial vibration band and then obtaining a number of suspected resonance segments includes the following specific methods:

[0016] The absolute value of the difference between the characteristic angles of any two initial vibration bands is used as the distance metric, and the DBSCAN clustering algorithm is used to set the cluster radius to , set the minimum number of sample points to , clustering all initial vibration bands to obtain several clusters; among them, is the preset cluster radius, is the preset minimum number of sample points;

[0017] The cluster containing the most initial vibration bands is recorded as the normal reference cluster, and the initial vibration bands that are not in the normal reference cluster are recorded as suspected vibration bands; in the vibration data sequence, adjacent suspected vibration bands are merged to obtain several merged vibration bands, and the merged vibration bands are recorded as suspected resonance segments.

[0018] Furthermore, the vibration outlier degree of each suspected resonance segment is obtained according to the difference between the characteristic angle of each initial vibration band in each suspected resonance segment and the characteristic angle of the cluster center point of the normal reference cluster, including the specific method of:

[0019]

[0020] Where, Indicates the The vibration outlier of the suspected resonance segment, represents the number of initial vibration bands contained in the normal reference cluster, Indicates the The number of initial vibration bands contained in the suspected resonance segment, Indicates the The first The absolute value of the difference between the characteristic angles of the suspected initial vibration band and the cluster center point of the normal reference cluster.

[0021] Furthermore, the correlation coefficient of each suspected resonance segment is obtained according to the change of the vibration data in each suspected resonance segment, including the specific method of:

[0022] The first Among the suspected resonance segments The minimum absolute value of the difference between the acquisition time of the vibration data with the largest amplitude and the acquisition time of the vibration data with the largest amplitude is recorded as Among the suspected resonance segments The time characteristic data of the vibration data; The time characteristic data of each vibration data in the suspected resonance segment is obtained A time characteristic data sequence of a suspected resonance segment;

[0023] According to The amplitude of each vibration data in the suspected resonance segment is obtained The amplitude sequence of the suspected resonance segment; get the The time characteristic data series of the suspected resonance segment and the The Pearson correlation coefficient of the amplitude sequence of the suspected resonance segment is denoted as The correlation coefficient of the suspected resonance segment.

[0024] Furthermore, the resonance probability of each suspected resonance segment is obtained according to the correlation coefficient and vibration outlier of each suspected resonance segment, including the specific method of:

[0025]

[0026] Where, Indicates the The resonance probability of a suspected resonance segment, Indicates the The vibration outlier of the suspected resonance segment, Indicates the The correlation coefficient of the suspected resonance segment, represents the normalization function, is an exponential function with a natural constant as its base.

[0027] Furthermore, the correction value of each vibration data in each suspected resonance segment is obtained according to the resonance probability of each suspected resonance segment and the amplitude of each vibration data in each suspected resonance segment, including the specific method of:

[0028]

[0029] Where, Indicates the The first Correction value of vibration data, Indicates the The first The amplitude of the vibration data, Indicates the The resonance probability of a suspected resonance segment, represents the mean value of the amplitude of all vibration data in all initial vibration bands in the normal reference cluster, represents the absolute value function.

[0030] Furthermore, the method of obtaining the vibration data correction sequence includes the following specific steps:

[0031] The amplitude of any vibration data in any initial vibration band in the normal reference cluster is recorded as the correction value of the vibration data; the correction value of each vibration data in each suspected resonance segment is combined to obtain a vibration data correction sequence.

[0032] Furthermore, the method of obtaining the vibration anomaly index of the wind turbine blade to be detected based on the change of the correction value of the vibration data in the vibration data correction sequence and performing anomaly detection includes the following specific methods:

[0033]

[0034] Where, Indicates the abnormal vibration index of the wind turbine blade to be detected. Indicates the number of extreme value points contained in the vibration data correction sequence, Indicates the first The absolute value of the difference between the correction value of the extreme point and the mean of the correction values of all vibration data, express function;

[0035] Preset indicator thresholds If the vibration abnormality index of the wind turbine blade to be detected is greater than the index threshold , the vibration of the wind turbine blade to be detected is abnormal.

[0036] The present invention also proposes a system for detecting abnormal operation of equipment in water conservancy and hydropower projects, comprising a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, the steps of the above method are implemented.

[0037] The beneficial effects of the technical solution of the present invention are as follows: in the process of detecting abnormal vibration of wind turbine blade operation, several initial vibration bands are obtained from a vibration data sequence; based on the characteristic that a normal vibration data sequence collected by a vibration sensor on the wind turbine blade is composed of several continuous and similar vibration bands, suspected resonance segments are obtained from the initial vibration bands; based on the characteristic that when the wind turbine blade is affected by resonance, the amplitude of the collected vibration data rapidly increases and then rapidly decreases, the resonance probability of each suspected resonance segment is calculated, so that the resonance probability of the suspected resonance segment affected by resonance is greater than the resonance probability of the suspected resonance segment affected by wind turbine blade wear; and then, when obtaining a correction value of vibration data in each suspected resonance segment based on the resonance probability of each suspected resonance segment, the correction value of vibration data in the suspected resonance segment affected by resonance is relatively close to the amplitude of normal vibration data not affected by resonance, and the correction value of vibration data in the suspected resonance segment affected by wind turbine blade wear is significantly different from the amplitude of normal vibration data not affected by resonance; so that the vibration abnormality index of the wind turbine blade to be detected obtained based on the vibration data correction sequence is more in line with the actual situation, thereby improving the accuracy of the abnormality detection result. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0039] Figure 1 The present invention is a flowchart of the steps of a method for detecting abnormal operation of equipment in a water conservancy and hydropower project. DETAILED DESCRIPTION

[0040] To further illustrate the technical means and effectiveness of the present invention in achieving its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effectiveness of a method for detecting abnormal operation of equipment in a water conservancy and hydropower project according to the present invention. In the following description, references to different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0041] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0042] The following describes in detail a method and system for detecting abnormal operation of equipment in a water conservancy and hydropower project provided by the present invention with reference to the accompanying drawings.

[0043] See also Figure 1 , which shows a flowchart of a method for detecting abnormal operation of equipment in a water conservancy and hydropower project provided by one embodiment of the present invention, the method comprising the following steps:

[0044] Step S001: Acquire a vibration data sequence of a wind turbine blade in a water conservancy and hydropower project.

[0045] This embodiment performs abnormality detection on the operating vibration of the wind turbine blade, thereby obtaining a vibration data sequence of the wind turbine blade to be detected.

[0046] Specifically, a vibration data sequence is obtained by installing a vibration sensor on the wind turbine blade to be tested. The vibration data sequence contains a number of vibration data, each of which corresponds to a collection time and amplitude. The collection frequency of the vibration sensor is seconds, the collection time is Hours, the collection frequency preset in this embodiment , preset collection time , is described using this as an example, and other values can be set in other implementations.

[0047] Step S002: Based on the changes in vibration data in the vibration data sequence, several initial vibration bands are obtained, and then the characteristic angle of each initial vibration band is obtained; based on the characteristic angle of each initial vibration band, a normal reference cluster is obtained, and then several suspected resonance segments are obtained; based on the difference between the characteristic angle of each initial vibration band in each suspected resonance segment and the characteristic angle of the cluster center point of the normal reference cluster, the vibration outlier of each suspected resonance segment is obtained.

[0048] It should be noted that the amplitude of normal vibration data affected by wind turbine resonance differs significantly from that of normal vibration data not affected by wind turbine resonance, which can lead to inaccurate detection of wind turbine blade vibration anomalies. Therefore, before detecting anomalies in wind turbine blade vibration, the amplitude of normal vibration data affected by wind turbine resonance is corrected.

[0049] It should be further noted that when a wind turbine blade is affected by resonance, the amplitude of the collected vibration data initially increases rapidly. Then, as the sea breeze frequency changes, the amplitude of the collected vibration data decreases rapidly. Therefore, the two endpoints of the vibration data segment under the influence of resonance are considered to be two local minima. Thus, the local minima in the vibration data sequence are obtained, resulting in several initial vibration bands.

[0050] It should be further explained that when wind turbine blades are affected by resonance, the amplitude of the collected vibration data first increases rapidly, then decreases rapidly as the sea breeze frequency changes. Therefore, the characteristic angle of each initial vibration band is obtained based on the time difference and amplitude difference between the vibration data with the largest amplitude and the two endpoints within each initial diagnostic band. When wind turbine blades are not affected by resonance, the collected time-series normal vibration data is continuous and similar vibration bands; this makes the characteristic angles of normal initial vibration bands unaffected by resonance relatively similar. Therefore, the initial vibration bands are clustered based on the characteristic angle of each initial vibration band to obtain a normal reference cluster, and then the characteristic angles of several suspected resonance segments and normal vibration bands are obtained. Because the normal reference cluster contains a large number of initial vibration bands unaffected by resonance and abnormal vibration data, the characteristic angle of the cluster center of the normal reference cluster is considered the characteristic angle of the normal vibration band.

[0051] It should be further explained that because several initial vibration bands are derived based on the distribution of minimum values in the vibration data sequence, a single initial vibration band may contain both normal vibration data affected by resonance and normal vibration data not affected by resonance. Consequently, each suspected resonance band may also contain normal vibration data not affected by resonance. Therefore, the probability of each suspected resonance band containing normal vibration data not affected by resonance is calculated.

[0052] It should be further explained that when calculating the likelihood of each suspected resonant segment containing normal vibration data unaffected by resonance, since each suspected resonant segment is formed by merging a number of initial vibration bands, the likelihood of each suspected resonant segment containing normal vibration data unaffected by resonance is calculated based on the initial vibration bands contained in the suspected resonant segment. Since the difference between an initial vibration band and a normal vibration band in a suspected resonant segment reflects the number of normal vibration data unaffected by vibration contained in the initial vibration band, the degree of outliers of each suspected resonant segment, i.e., the likelihood of each suspected resonant segment containing normal vibration data unaffected by resonance, is obtained based on the difference between the characteristic angle of each initial vibration band and the characteristic angle of the normal vibration band in each suspected resonant segment.

[0053] Specifically, the minimum value in the vibration data sequence is obtained, and the vibration data between two adjacent minimum values and the vibration data with the shortest acquisition time between two adjacent minimum values are taken as data within an initial vibration band to obtain several initial vibration bands. Among them, the vibration data corresponding to the minimum value with the longest acquisition time belongs to its adjacent initial vibration band. Among them, obtaining the minimum value in a vibration data sequence is an existing well-known technology and will not be described in detail in this embodiment. The vibration data in the initial vibration band are mapped to a two-dimensional coordinate system with the horizontal axis being the acquisition time and the vertical axis being the amplitude, and a number of data points are obtained. The vector pointing from the data point with the largest amplitude to the data point with the smallest acquisition time is recorded as The first vector of the initial vibration band is the vector that points the data point with the smallest amplitude to the data point with the largest acquisition time, which is recorded as the first vector. The second vector of the initial vibration band. The angle between the first vector and the second vector of the initial vibration band is recorded as The characteristic angle of the initial vibration band.

[0054] Furthermore, the absolute value of the difference between the characteristic angles of any two initial vibration bands is used as the distance metric, and the DBSCAN clustering algorithm is used to set the cluster radius to , set the minimum number of sample points to , cluster all initial vibration bands to obtain several clusters. Among them, the cluster radius preset in this embodiment is , the minimum number of sample points , is described as an example, and other implementations may be set to other values. The Chinese name of the DBSCAN clustering algorithm is density-based clustering algorithm, which is a well-known technology and will not be described in detail in this embodiment.

[0055] Furthermore, the cluster containing the most initial vibration bands is recorded as the normal reference cluster, and the initial vibration bands not in the normal reference cluster are recorded as suspected vibration bands. In the vibration data sequence, adjacent suspected vibration bands are merged to obtain several merged vibration bands, which are recorded as suspected resonance segments.

[0056] Furthermore, the specific calculation formula for obtaining the vibration outlier degree of each suspected resonance segment is as follows:

[0057]

[0058] Where, Indicates the The vibration outlier of the suspected resonance segment, represents the number of initial vibration bands contained in the normal reference cluster, Indicates the The number of initial vibration bands contained in the suspected resonance segment, Indicates the The first The absolute value of the difference between the characteristic angles of the suspected initial vibration band and the cluster center point of the normal reference cluster.

[0059] What needs to be explained is that The larger the value, the The greater the difference between the initial vibration band in the suspected resonance segment and the initial vibration band in the normal reference cluster, the greater the difference between the initial vibration band in the suspected resonance segment and the normal reference cluster. The smaller the probability that a suspected resonance segment contains normal vibration data that is not affected by resonance; The larger the value of , the more the cluster center of the normal reference cluster can reflect the characteristics of the normal initial vibration band, that is, The greater the credibility.

[0060] At this point, the vibration outlier of each suspected resonance segment is obtained.

[0061] Step S003: obtaining a correlation coefficient of each suspected resonance segment according to the change of vibration data in each suspected resonance segment; obtaining a resonance probability of each suspected resonance segment according to the correlation coefficient and vibration outlier degree of each suspected resonance segment.

[0062] It should be noted that due to the significant difference between the vibration bands collected from normal wind turbine blades not affected by resonance and the vibration bands collected from abnormal wind turbine blades, the suspected vibration bands may be the initial vibration bands affected by wind turbine blade resonance or the initial vibration bands affected by wind turbine blade wear. Furthermore, since each suspected resonance segment is composed of several suspected vibration bands, the probability that the suspected vibration band contained in each suspected resonance segment is the initial vibration band affected by wind turbine blade resonance, i.e., the resonance probability of each suspected resonance segment, is calculated.

[0063] It should be further explained that when calculating the resonance probability of each suspected resonance segment, the amplitude of the collected vibration data affected by resonance first increases rapidly and then decreases rapidly. This means that the amplitude of the vibration data within the normal vibration band affected by resonance is inversely proportional to the difference in acquisition time between that vibration data and the vibration data with the maximum amplitude within that normal vibration band. Therefore, the resonance probability of each suspected resonance segment is determined based on the changes in the vibration data within that segment.

[0064] Specifically, the Among the suspected resonance segments The minimum absolute value of the difference between the acquisition time of the vibration data with the largest amplitude and the acquisition time of the vibration data with the largest amplitude is recorded as Among the suspected resonance segments The time characteristic data of the vibration data. The time characteristic data of each vibration data in the suspected resonance segment is obtained The time characteristic data sequence of the suspected resonance segment. The amplitude of each vibration data in the suspected resonance segment is obtained The amplitude sequence of the suspected resonance segment. The time characteristic data series of the suspected resonance segment and the The Pearson correlation coefficient of the amplitude sequence of the suspected resonance segment is denoted as The Pearson correlation coefficient of two sequences is obtained as a known technique and will not be described in this embodiment.

[0065] Furthermore, the specific calculation formula for obtaining the resonance probability of each suspected resonance segment according to the correlation coefficient of each suspected resonance segment is as follows:

[0066]

[0067] Where, Indicates the The resonance probability of a suspected resonance segment, Indicates the The vibration outlier of the suspected resonance segment, Indicates the The correlation coefficient of the suspected resonance segment, Represents a linear normalization function, and the normalization object is each suspected resonance segment ; is an exponential function with a natural constant as the base. The model shows an inverse proportional relationship. As the input of the model, the implementer can set the inverse proportional function according to the actual situation.

[0068] What needs to be explained is that The smaller the value, the The more the suspected resonance segment is consistent with the change characteristics of the vibration data under the influence of resonance, the more The greater the resonance probability of a suspected resonance segment; The larger the value, the The smaller the probability that the first suspected resonance segment contains a normal vibration band that is not affected by the resonance, the smaller the probability that the first suspected resonance segment contains a normal vibration band that is not affected by the resonance. The greater the resonance probability of the suspected resonance segment.

[0069] At this point, the resonance probability of each suspected resonance segment is obtained.

[0070] Step S004: obtaining a correction value of each vibration data in each suspected resonance segment according to the resonance probability of each suspected resonance segment and the amplitude of each vibration data in each suspected resonance segment, thereby obtaining a vibration data correction sequence.

[0071] It should be noted that because the amplitude of normal vibration data affected by resonance is greater than the amplitude of normal vibration data not affected by resonance, the amplitude of the normal vibration data affected by resonance is corrected to reduce the difference between the amplitudes of the normal vibration data affected by resonance and the amplitudes of the normal vibration data not affected by resonance. The correction value for each vibration data within each suspected resonance segment is obtained based on the resonance probability of each suspected resonance segment, the amplitude of each vibration data within each suspected resonance segment, and the amplitude of all normal vibration data.

[0072] Specifically, according to The amplitude of each vibration data in the suspected resonance segment is The resonance probability of the suspected resonance segment is obtained The first The specific calculation formula for the correction value of vibration data is as follows:

[0073]

[0074] Where, Indicates the The first Correction value of vibration data, Indicates the The first The amplitude of the vibration data, Indicates the The resonance probability of a suspected resonance segment, represents the mean value of the amplitude of all vibration data in all initial vibration bands in the normal reference cluster, represents the absolute value function.

[0075] What needs to be explained is that The larger the value, the The first The greater the difference between the amplitude of the vibration data and the normal vibration data not under the influence of resonance, that is, The first The amplitude of each vibration data needs to be adjusted more; The larger the value, the The greater the possibility that the vibration data in the suspected resonance segment is normal vibration data affected by resonance, the greater the possibility that the vibration data in the suspected resonance segment is normal vibration data affected by resonance. The data within the suspected resonance segment requires greater adjustment.

[0076] The amplitude of any vibration data in any initial vibration band in the normal reference cluster is recorded as the correction value of the vibration data; the correction value of each vibration data in each suspected resonance segment is combined to obtain a vibration data correction sequence.

[0077] At this point, the vibration data correction sequence is obtained.

[0078] Step S005: obtaining a vibration anomaly index of the wind turbine blade to be detected according to the change of the correction value of the vibration data in the vibration data correction sequence, and performing anomaly detection.

[0079] It should be noted that, since the amplitude of normal vibration data not affected by resonance changes slightly, the vibration abnormality index of the wind turbine blade to be detected is obtained according to the change of the vibration data correction value in the vibration data correction sequence.

[0080] Specifically, the extreme value points in the vibration data correction sequence are obtained according to Newton's method. The Newton's method is a well-known technique and will not be described in detail in this embodiment. The specific calculation formula for obtaining the vibration anomaly index of the wind turbine blade to be inspected based on the amplitude of each extreme value point in the vibration data correction sequence is as follows:

[0081]

[0082] Where, Indicates the abnormal vibration index of the wind turbine blade to be detected. Indicates the number of extreme value points contained in the vibration data correction sequence, Indicates the first The absolute value of the difference between the correction value of the extreme point and the mean of the correction values of all vibration data, express Function, this embodiment is used for normalization processing.

[0083] What needs to be explained is that The larger the value of , the stronger the fluctuation of the vibration data correction value in the vibration data correction sequence of the wind turbine blade to be detected, that is, the greater the vibration abnormality index of the wind turbine blade to be detected.

[0084] Furthermore, preset indicator thresholds If the vibration abnormality index of the wind turbine blade to be detected is greater than the index threshold , it is considered that the vibration of the wind turbine blade to be detected is abnormal, and the wind turbine needs to be stopped and the wind turbine blade to be detected needs to be repaired. , is described using this as an example, and other implementations may set it to other values.

[0085] At this point, this embodiment is completed.

[0086] Another embodiment of the present invention provides a system for detecting abnormal operation of equipment in water conservancy and hydropower projects. The system includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, steps S001 to S005 of the above method are implemented.

[0087] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for detecting abnormal operation of equipment in a water conservancy and hydropower project, characterized in that: The method comprises the following steps: Acquire a vibration data sequence of a wind turbine blade in a water conservancy and hydropower project, wherein the vibration data sequence includes a plurality of vibration data corresponding to a collection time and amplitude; Based on the change of vibration data in the vibration data sequence, a number of initial vibration bands are obtained, and then a characteristic angle of each initial vibration band is obtained; based on the characteristic angle of each initial vibration band, a normal reference cluster is obtained, and then a number of suspected resonance segments are obtained; the suspected resonance segments are composed of the number of initial resonance bands; based on the difference between the characteristic angle of each initial vibration band in each suspected resonance segment and the characteristic angle of the cluster center point of the normal reference cluster, the vibration outlier of each suspected resonance segment is obtained; According to the change of vibration data in each suspected resonance segment, the correlation coefficient of each suspected resonance segment is obtained; according to the correlation coefficient and vibration outlier degree of each suspected resonance segment, the resonance probability of each suspected resonance segment is obtained; According to the resonance probability of each suspected resonance segment and the amplitude of each vibration data in each suspected resonance segment, a correction value of each vibration data in each suspected resonance segment is obtained, thereby obtaining a vibration data correction sequence; According to the change of the correction value of the vibration data in the vibration data correction sequence, the vibration abnormality index of the wind turbine blade to be detected is obtained, and abnormality detection is performed.

2. The method for detecting abnormal operation of equipment in a water conservancy and hydropower project according to claim 1, characterized in that: The method of obtaining a plurality of initial vibration bands based on the change of the vibration data in the vibration data sequence and then obtaining the characteristic angle of each initial vibration band includes the following specific methods: Obtaining a minimum value in the vibration data sequence, taking the vibration data between two adjacent minimum values and the vibration data with the shortest acquisition time between the two adjacent minimum values as data within an initial vibration band, and obtaining a plurality of initial vibration bands; The first The vibration data in the initial vibration band are mapped to a two-dimensional coordinate system with the horizontal axis being the acquisition time and the vertical axis being the amplitude, and a number of data points are obtained; the vector pointing from the data point with the largest amplitude to the data point with the smallest acquisition time is recorded as the first The first vector of the initial vibration band is the vector that points the data point with the smallest amplitude to the data point with the largest acquisition time, which is recorded as the first vector. The second vector of the initial vibration band; The angle between the first vector and the second vector of the initial vibration band is recorded as The characteristic angle of the initial vibration band.

3. The method for detecting abnormal operation of equipment in a water conservancy and hydropower project according to claim 1, characterized in that: The method of obtaining a normal reference cluster based on the characteristic angle of each initial vibration band and then obtaining a number of suspected resonance segments includes the following specific steps: The absolute value of the difference between the characteristic angles of any two initial vibration bands is used as the distance metric, and the DBSCAN clustering algorithm is used to set the cluster radius to , set the minimum number of sample points to , clustering all initial vibration bands to obtain several clusters; among them, is the preset cluster radius, is the preset minimum number of sample points; The cluster containing the most initial vibration bands is recorded as the normal reference cluster, and the initial vibration bands that are not in the normal reference cluster are recorded as suspected vibration bands; in the vibration data sequence, adjacent suspected vibration bands are merged to obtain several merged vibration bands, and the merged vibration bands are recorded as suspected resonance segments.

4. The method for detecting abnormal operation of equipment in a water conservancy and hydropower project according to claim 1, characterized in that: The vibration outlier degree of each suspected resonance segment is obtained based on the difference between the characteristic angle of each initial vibration band in each suspected resonance segment and the characteristic angle of the cluster center point of the normal reference cluster, including the specific method of: Where, Indicates the The vibration outlier of the suspected resonance segment, represents the number of initial vibration bands contained in the normal reference cluster, Indicates the The number of initial vibration bands contained in the suspected resonance segment, Indicates the The first The absolute value of the difference between the characteristic angles of the suspected initial vibration band and the cluster center point of the normal reference cluster.

5. The method for detecting abnormal operation of equipment in a water conservancy and hydropower project according to claim 1, characterized in that: The method of obtaining the correlation coefficient of each suspected resonance segment according to the change of the vibration data in each suspected resonance segment includes the following specific methods: The first Among the suspected resonance segments The minimum absolute value of the difference between the acquisition time of the vibration data with the largest amplitude and the acquisition time of the vibration data with the largest amplitude is recorded as Among the suspected resonance segments Time characteristic data of vibration data; According to The time characteristic data of each vibration data in the suspected resonance segment is obtained A time characteristic data sequence of a suspected resonance segment; According to The amplitude of each vibration data in the suspected resonance segment is obtained The amplitude sequence of the suspected resonance segment; get the The time characteristic data series of the suspected resonance segment and the The Pearson correlation coefficient of the amplitude sequence of the suspected resonance segment is denoted as The correlation coefficient of the suspected resonance segment.

6. The method for detecting abnormal operation of equipment in a water conservancy and hydropower project according to claim 1, characterized in that: The specific method for obtaining the resonance probability of each suspected resonance segment based on the correlation coefficient and vibration outlier of each suspected resonance segment is as follows: Where, Indicates the The resonance probability of a suspected resonance segment, Indicates the The vibration outlier of the suspected resonance segment, Indicates the The correlation coefficient of the suspected resonance segment, represents the normalization function, is an exponential function with a natural constant as its base.

7. The method for detecting abnormal operation of equipment in a water conservancy and hydropower project according to claim 1, characterized in that: The method of obtaining the correction value of each vibration data in each suspected resonance segment according to the resonance probability of each suspected resonance segment and the amplitude of each vibration data in each suspected resonance segment includes the following specific methods: Where, Indicates the The first Correction value of vibration data, Indicates the The first The amplitude of the vibration data, Indicates the The resonance probability of a suspected resonance segment, represents the mean value of the amplitude of all vibration data in all initial vibration bands in the normal reference cluster, represents the absolute value function.

8. The method for detecting abnormal operation of equipment in a water conservancy and hydropower project according to claim 1, characterized in that: The specific method of obtaining the vibration data correction sequence includes: The amplitude of any vibration data in any initial vibration band in the normal reference cluster is recorded as the correction value of the vibration data; the correction value of each vibration data in each suspected resonance segment is combined to obtain a vibration data correction sequence.

9. The method for detecting abnormal operation of equipment in a water conservancy and hydropower project according to claim 1, characterized in that: The method of obtaining the vibration abnormality index of the wind turbine blade to be detected based on the change of the correction value of the vibration data in the vibration data correction sequence and performing abnormality detection includes the following specific methods: Where, Indicates the abnormal vibration index of the wind turbine blade to be detected. Indicates the number of extreme value points contained in the vibration data correction sequence, Indicates the first The absolute value of the difference between the correction value of the extreme point and the mean of the correction values of all vibration data, express function; Preset indicator thresholds If the vibration abnormality index of the wind turbine blade to be detected is greater than the index threshold , the vibration of the wind turbine blade to be detected is abnormal.

10. A system for detecting abnormal operation of equipment in a water conservancy and hydropower project, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the method for detecting abnormal operation of equipment in a water conservancy and hydropower project as described in any one of claims 1 to 9 are implemented.

Images