Water conservancy project frequency converter state detection method and system

Abstract

The invention relates to the technical field of frequency converter anomaly detection, in particular to a hydraulic engineering frequency converter state detection method and system. The invention aims to improve the accuracy of abnormal data point detection in the frequency converter performance parameter time sequence data, and particularly focuses on the influence of environmental factors on performance parameters. Performance parameter time sequence data and environment parameter time sequence data are synchronously obtained, discrete wavelet transform is used for decomposing performance parameters in a multi-layer mode, and approximation coefficients are mainly analyzed to insight the operation state of the frequency converter. And after the suspected abnormal data segments are preliminarily screened, environment abnormal indexes are calculated in combination with environment parameters, and fluctuation degree values of the suspected abnormal data segments are analyzed. And based on the environmental anomaly index and the fluctuation degree value, adaptively adjusting the initial threshold value of each approximation coefficient to obtain an adaptive threshold value, and finally detecting abnormal data based on the adaptive threshold value. According to the method, abnormal data points caused by environmental factors can be effectively eliminated, and abnormal data points caused by abnormity of the frequency converter can be accurately identified.

Description

Technical Field

[0001] The present application relates to the technical field of inverter abnormality detection, and in particular to a method and system for detecting the state of an inverter in a water conservancy project. Background Art

[0002] The frequency converter is a kind of power control equipment that is frequently used in water conservancy projects and can play different functions. In water conservancy projects, the frequency converter mainly adjusts the speed of equipment such as water pumps and fans, and combines sensor technology to achieve precise control of parameters such as flow, pressure, and liquid level. It can not only achieve energy saving and consumption reduction, but also protect equipment through precise control and soft start. Therefore, the stable operation of the frequency converter is of great significance to ensure the safety and efficiency of water conservancy project construction. In order to detect abnormal conditions of the frequency converter in time, it is necessary to monitor the operation process of the frequency converter, so as to detect the abnormal state of the frequency converter, remind the staff to repair it in time, and ensure the safety of construction.

[0003] In the prior art, when performing abnormality detection on the working state of the inverter, wavelet transform is usually performed on the detection data of relevant parameters, and a fixed threshold is set to identify abnormal points caused by the abnormal state of the inverter. However, due to the particularity of the water conservancy project environment, environmental factors such as temperature, humidity, dust, etc. during the operation of the inverter will cause abnormal fluctuations in the performance parameters of the inverter. Therefore, if a fixed threshold is used to detect abnormal points, the abnormal points obtained will not be accurate enough, which may easily lead to misjudgment. Summary of the invention

[0004] In order to solve the technical problem that the detection data during the operation of the frequency converter in the water conservancy project is affected by environmental factors, resulting in misjudgment of abnormal point detection using a fixed threshold, the purpose of this application is to provide a water conservancy project frequency converter state detection method and system, the technical solution adopted is as follows: In a first aspect, the present application proposes a method for detecting the state of a water conservancy project inverter, the method comprising the following steps: Obtaining the time series data of the performance parameters of the inverter within a preset time period and the time series data of the environmental parameters when the inverter is running; Based on discrete wavelet transform, the performance parameter time series data is multi-layered decomposed to obtain approximate coefficients; based on the difference between the values ​​of each approximate coefficient and the corresponding initial thresholds, the suspected abnormal data segments are determined in the performance parameter time series data; the data other than the suspected abnormal data segments are used as pending data; in the environmental parameter time series data, the environmental abnormality index is obtained according to the change of the value corresponding to the moment of each suspected abnormal data segment; Select any suspected abnormal data segment as the target data segment and determine the corresponding comparison data segment; determine the fluctuation degree value of the target data segment according to the difference between the fluctuation of the values ​​in the target data segment and all the comparison data segments, and the difference between the changes in the values ​​of the approximate coefficients; Based on the fluctuation value of the target data segment and the corresponding environmental anomaly index, the initial threshold of each approximate coefficient is adjusted to obtain the adaptive threshold of the target data segment under each approximate coefficient; under each approximate coefficient, based on the difference between the numerical value corresponding to all suspected abnormal data segments and the adaptive threshold, as well as the difference between the numerical value of the pending data and the corresponding initial threshold, anomaly detection is performed to obtain abnormal data points.

[0005] Furthermore, the determining of suspected abnormal data segments in the performance parameter time series data based on the differences between the values ​​of the respective approximation coefficients and the respective corresponding initial thresholds includes: For the approximate coefficient generated by any layer of decomposition, the absolute value of each value in the approximate coefficient is compared with the corresponding initial threshold, and the value whose absolute value is greater than the corresponding initial threshold is set to 0 to obtain the adjusted approximate coefficient; Reconstruct all adjusted approximate coefficients to obtain first reconstructed time series data, and in the performance parameter time series data, subtract each value from the value at the corresponding moment in the first reconstructed time series data, and take the data point whose value is not 0 as the target point; The target points with continuous time series are grouped into suspected abnormal data segments, and all the suspected abnormal data segments in the performance parameter time series data are obtained.

[0006] Furthermore, the method for obtaining the environmental abnormality index includes: In the environmental parameter time series data, the corresponding environmental parameter data segment is determined according to the time of each suspected abnormal data segment; For any environmental parameter data segment, a first-order difference sequence of values ​​in the environmental parameter data segment is calculated, and the values ​​in the first-order difference sequence are averaged to obtain a first environmental parameter; Determine a second environmental parameter of the environmental parameter data segment based on a difference between a numerical mean value in the environmental parameter data segment and a numerical mean value of the environmental parameter time series data; Determine a third environmental parameter of the environmental parameter data segment based on a difference between a numerical mean and a numerical maximum in the environmental parameter data segment; The first environmental parameter, the second environmental parameter and the third environmental parameter are integrated to obtain an environmental anomaly index of the environmental parameter data segment, and the first environmental parameter, the second environmental parameter and the third environmental parameter are all positively correlated with the environmental anomaly index.

[0007] Furthermore, the method for obtaining the fluctuation degree value includes: Determining data anomaly indicators based on fluctuations of values ​​in the target data segment and each comparison data segment respectively; According to the difference between the data anomaly index of each comparison data segment and the data anomaly index of the target data segment, a fluctuation factor between the target data segment and each comparison data segment is obtained; Determine an overall fluctuation index between the target data segment and each of the comparison data segments based on the numerical differences between the target data segment and each of the comparison data segments in each of the approximation coefficients; The volatility factors between the target data segment and all the comparison data segments and the overall volatility index are weighted and averaged to obtain the volatility value of the target data segment.

[0008] Furthermore, the method for obtaining the data anomaly indicator includes: Select one of the target data segment and the corresponding comparison data segment as the data segment to be tested; Based on the values ​​in the data segment to be tested, determine the maximum variation range of the values ​​in the data segment to be tested; calculate the first-order difference sequence of the values ​​in the data segment to be tested, and average the absolute values ​​of the values ​​in the first-order difference sequence to obtain a first abnormal factor; use the mean of the values ​​in the data segment to be tested as the mean characteristic value, use the difference between each value in the data segment to be tested and the mean characteristic value as the difference factor, and use the mean of all the difference factors as the second abnormal factor; Based on the maximum variation range value, the first abnormal factor and the second abnormal factor of the data segment to be tested, a data abnormality index of the data segment to be tested is determined, and the maximum variation range value, the first abnormal factor and the second abnormal factor all change in the same direction as the data abnormality index.

[0009] Furthermore, the method for obtaining the adaptive threshold includes: Determine the environmental impact index based on the fluctuation degree value of the target data segment and the environmental anomaly index; When the environmental impact index is greater than or equal to the preset impact threshold, the sum of the environmental impact index and the preset constant is used as the adjustment weight; when the environmental impact index is less than the preset impact threshold, the environmental impact index is used as the adjustment weight; The initial thresholds of the respective approximation coefficients are weightedly adjusted based on the adjustment weights to obtain adaptive thresholds of the target data segment under the respective approximation coefficients.

[0010] Furthermore, the method for obtaining the environmental impact index includes: The value obtained by normalizing the product of the fluctuation degree value of the target data segment and the corresponding environmental anomaly index is used as the environmental impact index.

[0011] Furthermore, the method for obtaining abnormal data points includes: Under each approximate coefficient, the absolute value of each numerical value corresponding to each suspected abnormal data segment is compared with the adaptive threshold, and the numerical value whose absolute value is greater than the adaptive threshold is set to 0 to obtain the adjusted approximate coefficient data segment; Under each approximate coefficient, the absolute value of the value of the pending data is compared with the corresponding initial threshold, and the value whose absolute value is greater than the initial threshold is set to 0 to obtain the adjusted pending data; Based on the adjusted approximate coefficient data segments under each approximate coefficient and the adjusted pending data, each target approximate coefficient is obtained; all target approximate coefficients are reconstructed to obtain second reconstructed time series data, and in the performance parameter time series data, each numerical value is subtracted from the numerical value at the corresponding moment in the second reconstructed time series data, and data points with values ​​not equal to 0 are regarded as abnormal data points.

[0012] Furthermore, the method for obtaining the initial threshold includes: Calculate the standard deviation of all values ​​of each approximation coefficient; The mean of all values ​​of each approximation coefficient plus three times the corresponding standard deviation is used as the initial threshold of each approximation coefficient.

[0013] In the second aspect, the present application proposes a water conservancy project inverter status detection system, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it can implement any step of the water conservancy project inverter status detection method.

[0014] This application has the following beneficial effects: The purpose of this application is to analyze the influence of environmental factors on the performance parameters of the frequency converter, so as to set an adaptive threshold for the detection of abnormal data points, and avoid the problem of low accuracy of abnormal data point detection caused by misjudgment. Therefore, the performance parameter time series data of the frequency converter is first obtained, and the environmental parameter time series data of the same period is obtained at the same time. Then, the performance parameter time series data is multi-layered decomposed by discrete wavelet transform, and the information of different frequency components can be extracted. Since the performance parameter changes caused by environmental changes are usually time-persistent, they often appear abnormal within a period of time, so the approximate coefficient is mainly analyzed in this application to facilitate a deeper understanding of the operating state of the frequency converter. By comparing the difference between the value of the approximate coefficient and the initial threshold, the suspected abnormal data segment can be preliminarily screened out, laying the foundation for subsequent detailed abnormal detection. Then, combined with the environmental parameter time series data, the environmental abnormality index is calculated, and it can be judged whether the abnormal data segment is related to the environmental factors in the subsequent process; at the same time, by analyzing the difference between the fluctuation of the suspected abnormal data segment and the comparison data segment, and the difference in the change of the approximate coefficient value, the fluctuation degree value of the suspected abnormal data segment can be determined. Finally, based on the environmental anomaly index and fluctuation value corresponding to the suspected abnormal data segment, the initial threshold of each approximate coefficient is adaptively adjusted to obtain the adaptive threshold of the suspected abnormal data segment under each approximate coefficient. At this time, the adaptive threshold analyzes the relationship between environmental factors and the fluctuation of performance parameters, so it can more effectively eliminate the misjudged abnormal data points in subsequent abnormality detection, that is, it effectively eliminates the situation where the performance parameters change due to changes in environmental factors. Therefore, anomaly detection based on the adaptive threshold can obtain more accurate detection results. At this time, the abnormal data point is the abnormal data point caused by the abnormality of the inverter itself. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0016] Figure 1 A method flow chart of a method for detecting the state of a frequency converter of a water conservancy project provided by an embodiment of the present application; Figure 2 A schematic diagram of a 5-layer decomposition of discrete wavelet decomposition provided by an embodiment of the present application; Figure 3 A flow chart of a method for obtaining a fluctuation degree value provided by an embodiment of the present application; Figure 4A flow chart of a method for obtaining an adaptive threshold value provided by an embodiment of the present application; Figure 5 A system block diagram of a water conservancy project inverter state detection system provided by an embodiment of the present application; Figure 6 A schematic diagram of the system structure of a water conservancy project inverter state detection system provided by one embodiment of the present application; Figure 7 A schematic diagram of a computer-readable storage medium provided for one embodiment of the present application. DETAILED DESCRIPTION

[0017] In order to further explain the technical means and effects adopted by the present application to achieve the predetermined invention purpose, the following is a detailed description of a method and system for detecting the state of a frequency converter of a water conservancy project proposed in accordance with the present application, its specific implementation method, structure, features and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0018] 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 application belongs.

[0019] The following is a detailed description of a method and system for detecting the state of a frequency converter for a water conservancy project provided by the present application in conjunction with the accompanying drawings.

[0020] See also Figure 1 , which shows a method flow chart of a method for detecting the state of a frequency converter of a water conservancy project provided by an embodiment of the present application, the method comprising the following steps: Step S1: Acquire the time series data of the performance parameters of the inverter within a preset time period and the time series data of the environmental parameters when the inverter is running.

[0021] Voltage is the most basic performance parameter in the operation of the frequency converter. Monitoring it can timely discover the problem of abnormal fluctuations in the power grid voltage. Abnormal fluctuations in voltage will lead to increased equipment loss. Long-term use under unstable voltage conditions will affect the working state of the frequency converter and affect the life of the equipment. One of the main functions of the frequency converter is to change the frequency of the output power supply to meet the needs of different motor controls. Therefore, the abnormal fluctuation of the output frequency can also reflect the operating state of the frequency converter. Of course, the performance parameters of the frequency converter also include speed, current, etc. In one embodiment of the present application, voltage is selected as the target performance parameter, and based on the voltage sensor device deployed on the frequency converter, the voltage timing data of the frequency converter within a preset time period is obtained as the performance parameter timing data of the frequency converter. It should be noted that in other embodiments of the present application, the output frequency can also be selected as the target performance parameter, and the frequency timing data can be obtained as the performance parameter timing data of the frequency converter, which is not limited or elaborated here.

[0022] When the inverter itself is abnormal, such as: power supply voltage fluctuations, load instantaneous changes, etc., its performance parameters will fluctuate abnormally. However, due to the special environment in which the inverter is located in the water conservancy project, the temperature, humidity and dust concentration in the environment cannot be effectively controlled at a more suitable level, so the temperature, humidity and dust concentration of its working environment will change, and the change of environmental factors will also affect the working performance of the inverter, which will also cause performance parameters to fluctuate. Therefore, in the process of using a fixed threshold to detect abnormal data points, it is impossible to effectively distinguish between the two, which will cause misjudgment and fail to effectively identify whether the inverter is actually abnormal. Therefore, in the embodiment of the present application, by analyzing the impact of environmental factors on performance parameters, setting an adaptive threshold, distinguishing different situations, and improving the recognition accuracy of abnormal data points. Since the inverter is a high-power electronic component inside, it is very susceptible to the influence of the operating temperature. Therefore, in the embodiment of the present application, temperature is selected as the target environmental parameter, and the temperature time series data within a preset period is obtained based on the temperature sensor deployed around the inverter as the environmental parameter time series data. It should be noted that in other embodiments of the present application, humidity or dust concentration may also be selected as the target environmental parameter, and the corresponding environmental parameter time series data may be obtained, which is not limited or elaborated here.

[0023] At this point, the performance parameter time series data and environmental parameter time series data corresponding to the inverter can be obtained. It should be noted that the preset period is set to one hour here, and the sampling interval and period of the performance parameter time series data and the environmental parameter time series data should be consistent. The sampling interval is set to one minute. The specific period and sampling interval can be adjusted according to the implementation scenario, and no limitation or elaboration is made here.

[0024] Step S2: Perform multi-layer decomposition on the performance parameter time series data based on discrete wavelet transform to obtain approximate coefficients; determine suspected abnormal data segments in the performance parameter time series data based on the differences between the values ​​of each approximate coefficient and their corresponding initial thresholds; use data other than the suspected abnormal data segments as pending data; in the environmental parameter time series data, obtain the environmental anomaly index based on the changes in the values ​​corresponding to the moments of each suspected abnormal data segment.

[0025] Since the discrete wavelet transform can decompose the time series data into frequency bands of different scales, it can effectively capture multi-scale features and can be used to identify abnormal data. Abnormal data usually has obvious numerical deviation characteristics, so the performance parameter time series data can be multi-layered decomposition based on the discrete wavelet transform, and an initial threshold is set to preliminarily determine the suspected abnormal data segment in the performance parameter time series data. In view of the purpose of the embodiment of the present application is to analyze the impact of environmental factors on the performance parameters of the inverter, the change of environmental parameters corresponding to each suspected abnormal data segment can be analyzed in the environmental parameter time series data, so as to obtain the environmental anomaly index within the time period, and provide a reference for analyzing the impact of environmental factors in the subsequent process.

[0026] Discrete wavelet transform is a method for analyzing signals or data in the time and frequency domain. It can provide the time-frequency localization characteristics of the signal and effectively extract information from the signal. By performing discrete wavelet transform on the performance parameter time series data, the approximate coefficients (approximate components) and detail coefficients (detail components) of each layer can be obtained. Please refer to Figure 2 , which shows a schematic diagram of a 5-layer decomposition of discrete wavelet decomposition provided in one embodiment of the present application.

[0027] Among them, the approximate coefficient reflects the overall smoothness or approximate representation of the performance parameter time series data, and the detail coefficient reflects the local details of the performance parameter time series data, which can reveal the mutation in the signal, etc. It should be noted that in an embodiment of the present application, the number of decomposition layers is set to 5, and the specific number of decomposition layers can be adjusted according to the implementation scenario, which is not limited here. The discrete wavelet transform method is a technical means well known to those skilled in the art, which is not repeated here. The wavelet basis used can be the Haar wavelet, and the specific type of the wavelet basis can also be adjusted according to the implementation scenario, which is not limited here.

[0028] In view of the fact that the impact of environmental factors on performance parameters is time-persistent, it is necessary to analyze the fluctuation of performance parameters in a continuous time period. Therefore, the suspected abnormal part of the performance parameter time series data is determined based on the approximate coefficient. For each approximate coefficient, by comparing the value therein with the corresponding initial threshold, it is determined which values ​​in the approximate coefficient are abnormal, so as to facilitate the determination of suspected abnormal data segments in the performance parameter time series data. At this time, the suspected abnormal data segment may be caused by the change of environmental factors, or it may be caused by the abnormality of the inverter itself.

[0029] Preferably, in one embodiment of the present application, the method for obtaining the suspected abnormal data segment includes: First, for the approximate coefficient generated by any layer of decomposition, the absolute value of each value in the approximate coefficient is compared with the corresponding initial threshold, and the value whose absolute value is greater than the corresponding initial threshold is set to 0 to obtain the adjusted approximate coefficient.

[0030] Then all the adjusted approximate coefficients are reconstructed to obtain the first reconstructed time series data. In order to identify the data points that are different from the first reconstructed time series data and the performance parameter time series data, each value in the performance parameter time series data can be subtracted from the value at the corresponding moment in the first reconstructed time series data, and the data point with a value that is not 0 is taken as the target point, which is the preliminary abnormal data point.

[0031] Since continuous abnormal data points are more likely to be caused by environmental factors and are also targets that should be analyzed in depth, the target points that are continuous in time series are grouped into suspected abnormal data segments. Based on the above process, all suspected abnormal data segments in the performance parameter time series data can be obtained.

[0032] In other embodiments of the present application, the method for determining the target point may also be: in the performance parameter time series data, the difference between each value and the value at the corresponding moment in the first reconstructed time series data is calculated, and the data point whose difference value is greater than a preset deviation threshold is used as the target point, wherein the preset deviation threshold is set to 3, and the specific value can be adjusted according to the implementation scenario and is not limited here.

[0033] It should be noted that the initial threshold corresponding to each approximate coefficient is set as follows: the standard deviation of all values ​​in the approximate coefficient is calculated, and then the mean of all values ​​in the approximate coefficient plus three times the corresponding standard deviation. The setting of the initial threshold can also be adjusted based on experience or implementation scenarios, and is not limited here.

[0034] At this point, all suspected abnormal data segments in the performance parameter time series data can be obtained, and the data except the suspected abnormal data segments are used as pending data, which includes normal data points and abnormal data points caused by the abnormality of the inverter itself.

[0035] For the suspected abnormal data segment, since it may be affected by environmental factors, it is necessary to analyze the environmental parameter time series data to determine the changes in the environmental parameters in the time period corresponding to the suspected abnormal data segment and obtain the environmental anomaly index.

[0036] Preferably, in one embodiment of the present application, the method for obtaining an environmental anomaly indicator includes: In the environmental parameter time series data, the corresponding environmental parameter data segment is determined according to the moment of each suspected abnormal data segment. Through this process, the environmental parameters corresponding to the suspected abnormal data segment can be accurately obtained. For example, the suspected abnormal data segment corresponds to moment 3-moment 7, then in the environmental parameter time series data, the environmental parameter data segment corresponding to the suspected abnormal data segment is also moment 3-moment 7.

[0037] During the operation of the inverter, the temperature will gradually increase, and in a high temperature environment, the parameters of the motor windings will change, thereby affecting the calculation accuracy of the vector transformation, affecting the maximum output current, etc., and also affecting the stability of the inverter performance parameters. Therefore, in one embodiment of the present application, the increase in data values ​​in the environmental parameter data segment is mainly analyzed.

[0038] The increase of the values ​​in the environmental parameter data segment can be characterized based on the difference between the environmental parameters at adjacent moments. Therefore, for any environmental parameter data segment, the first-order difference sequence of the values ​​in the environmental parameter data segment is calculated, and the values ​​in the first-order difference sequence are averaged to obtain the first environmental parameter. The formula model of the first environmental parameter includes: in, Indicates The first environmental parameter of an environmental parameter data segment; Indicates The total number of values ​​in the first-order difference sequence corresponding to each environmental parameter data segment; Indicates The first order difference sequence corresponding to the environmental parameter data segment A numerical value.

[0039] In the formula model of the first environmental parameter, It represents the difference between the value of the latter moment and the value of the previous moment in the environmental parameter data segment. If the environmental parameter increases with time, then The value of will be a positive value, and the greater the degree of increase in the environmental parameter, the greater the value. Therefore, when the mean of all values ​​in the first-order difference sequence, that is, the larger the first environmental parameter, the greater the degree of increase in the environmental parameter at this time.

[0040] Similarly, the difference between the overall change trend of the environmental parameter data segment and the overall change trend of the environmental parameter time series data can also be analyzed at the overall level, that is, based on the difference between the numerical mean of the environmental parameter data segment and the numerical mean of the environmental parameter time series data, the second environmental parameter of the environmental parameter data segment can be determined. The formula model of the second environmental parameter includes: in, Indicates A second environmental parameter of an environmental parameter data segment; Indicates The numerical mean of the environmental parameter data segments; Represents the numerical mean of the environmental parameter time series data; Indicates the preset first parameter.

[0041] In the formula model of the second environmental parameter, by comparing the numerical mean of the environmental parameter data segment with the numerical mean of the environmental parameter time series data, the larger the ratio is, the higher the value in the environmental parameter data segment is compared with the entire environmental parameter time series data. It can also be said that the greater the degree of increase in the environmental parameter, the higher the value of the preset first parameter. The function of is to prevent the denominator from being 0. Here, the value can be 0.001. The specific value can be adjusted according to the implementation scenario and is not limited here.

[0042] In order to capture the difference between the overall change of the environmental parameter data segment and those environmental parameters that represent extreme or abnormal conditions, the third environmental parameter of the environmental parameter data segment can be determined based on the difference between the numerical mean and the numerical maximum in the environmental parameter data segment. The formula model of the third environmental parameter includes: in, Indicates The third environmental parameter of the environmental parameter data segment; Indicates The numerical mean of the environmental parameter data segments; Indicates the maximum value of the environmental parameter time series data; Indicates the preset second parameter.

[0043] In the formula model of the third environmental parameter, the average level of the values ​​in the environmental parameter data segment is used as the numerator, and the maximum value in the entire environmental parameter time series data is used as the denominator. At this time, the denominator represents the maximum value of the environmental parameter. If the numerator is larger, the third environmental parameter is larger, which means that the environmental parameter in the environmental parameter data segment is at a higher level, which means that the greater the degree of increase in the environmental parameter, the preset second parameter The function of is to prevent the denominator from being 0. Here, the value can be 0.001. The specific value can be adjusted according to the implementation scenario and is not limited here.

[0044] In other embodiments of the present application, the difference between the maximum value of the environmental parameter time series data and the numerical mean value of the environmental parameter data segment can also be negatively correlated and normalized to obtain the value as the third environmental parameter. The specific formula model is as follows: in, Indicates The third environmental parameter of the environmental parameter data segment; Indicates The numerical mean of the environmental parameter data segments; Indicates the maximum value of the environmental parameter time series data; Expressed as a natural constant An exponential function with base .

[0045] In this formula model, when the difference between the maximum value of the environmental parameter time series data and the mean value of the environmental parameter data segment The smaller the time is, the closer the average level of the environmental parameter data segment is to the maximum value of the entire environmental parameter time series data, which indicates that the environmental parameter in the environmental parameter data segment is at a higher level, indicating that the degree of increase in the environmental parameter is greater. Therefore, the difference is negatively correlated and normalized to achieve logical relationship correction, thereby obtaining the third environmental parameter.

[0046] Finally, in order to provide a comprehensive, multi-angle environmental anomaly index, the first environmental parameter, the second environmental parameter and the third environmental parameter can be integrated to obtain the environmental anomaly index of the environmental parameter data segment, and the first environmental parameter, the second environmental parameter and the third environmental parameter are all positively correlated with the environmental anomaly index. The formula model of the environmental anomaly index includes: in, Indicates Environmental anomaly indicators for each environmental parameter data segment; Indicates The first environmental parameter of an environmental parameter data segment; Indicates A second environmental parameter of an environmental parameter data segment; Indicates The third environmental parameter of the environmental parameter data segment; Represents the normalization function.

[0047] In the formula model of the environmental anomaly index, the larger the first environmental parameter, the larger the second environmental parameter, and the larger the third environmental parameter, the greater the environmental parameters are, which means that the environmental parameters are on an upward trend, which will have a greater impact on the performance parameters of the inverter. Therefore, the product of the three indicators is normalized and the value is used as the environmental anomaly index of the environmental parameter data segment.

[0048] In other embodiments of the present application, the normalized value of the sum of the first environmental parameter, the second environmental parameter, and the third environmental parameter may be used as the environmental anomaly indicator of the environmental parameter data segment.

[0049] Step S3: Select any suspected abnormal data segment as the target data segment and determine the corresponding comparison data segment; determine the fluctuation degree value of the target data segment based on the difference in the fluctuation of the values ​​in the target data segment and all the comparison data segments, and the difference in the change between the values ​​of the approximate coefficients.

[0050] Before determining the influence of environmental factors on the suspected abnormal data segment in the inverter performance parameters, the stability of the suspected abnormal data segment itself should be analyzed first. In order to more accurately measure the stability of the suspected abnormal data segment, in the embodiment of the present application, the suspected abnormal data segment is compared and analyzed with the comparison data segment. And because the fluctuation of the data will also be reflected in the approximate coefficients of each layer, in order to obtain a more accurate fluctuation degree value, the analysis is mainly carried out from two aspects, one is the difference in the fluctuation of the values ​​in the suspected abnormal data segment and the comparison data segment, and the other is the difference between the values ​​of the approximate coefficients, to determine the fluctuation degree value of the suspected abnormal data segment.

[0051] The method for determining the comparison data segment is as follows: for any suspected abnormal data segment, a preset number of data points before the timing of the suspected abnormal data segment constitute a comparison data segment, and similarly, a preset number of data points after the timing of the suspected abnormal data segment constitute a comparison data segment. If the suspected abnormal data segment is at the starting position of the performance parameter timing data, then only the preset number of data segments after the timing of the suspected abnormal data segment are obtained to form a comparison data segment. If the suspected abnormal data segment is at the end of the performance parameter timing data, then only the preset number of data segments before the timing of the suspected abnormal data segment are obtained to form a comparison data segment. The preset number is set to the segment length of the suspected abnormal data segment, and the specific value can be adjusted according to the implementation scenario, which is not limited here.

[0052] For the convenience of explanation and illustration, a suspected abnormal data segment is selected as the target data segment. Preferably, in one embodiment of the present application, the method for obtaining the fluctuation degree value includes: See also Figure 3, which shows a flow chart of a method for obtaining a fluctuation degree value provided by an embodiment of the present application, the method comprising the following steps: Step S301: Calculate data anomaly indicators based on fluctuations of values ​​in the target data segment and each corresponding comparison data segment.

[0053] Among the target data segment and the corresponding comparison data segment, one is selected as the data segment to be tested.

[0054] In view that the fluctuation of numerical values ​​can be characterized by the fluctuation range of numerical values, the changing trend of numerical values ​​and the difference between numerical values, the maximum variation range value of the numerical values ​​in the data segment to be tested is determined based on the numerical values ​​in the data segment to be tested. The maximum variation range value helps to understand the fluctuation amplitude of the numerical values ​​in the data segment to be tested.

[0055] And the first-order difference sequence of the values ​​in the data segment to be tested is calculated, and the absolute values ​​of the values ​​in the first-order difference sequence are averaged to obtain the first abnormal factor. The averaging of the first-order difference sequence can reflect the changes between the values ​​in the data segment to be tested, which helps to capture the slight changes in the values ​​and thus characterize the fluctuation of the data segment to be tested.

[0056] At the same time, the mean of the values ​​in the data segment to be tested is taken as the mean eigenvalue, the difference between each value in the data segment to be tested and the mean eigenvalue is taken as the difference factor, and the mean of all difference factors is taken as the second abnormal factor. The second abnormal factor reflects the degree of discreteness of the values ​​in the data segment to be tested relative to the mean, which helps to identify the discrete characteristics of the values ​​in the data segment to be tested.

[0057] Finally, based on the maximum variation range value, the first abnormal factor and the second abnormal factor of the data segment to be tested, the data abnormality index of the data segment to be tested is determined, and the maximum variation range value, the first abnormal factor and the second abnormal factor all change in the same direction as the data abnormality index. The formula model of the data abnormality index can be, for example: in, Indicates the data segment to be tested Data anomaly indicators; Indicates the data segment to be tested The range between the elements; Indicates the data segment to be tested The mean of the absolute values ​​of the elements in the corresponding first-order difference sequence; Indicates the data segment to be tested The total number of values ​​in ; Indicates the data segment to be tested Middle numerical values; Indicates the data segment to be tested The numerical mean of , also known as the mean eigenvalue.

[0058] In the formula model of data anomaly index, the range of the values ​​in the data segment to be tested characterizes the instability of the values ​​in the data segment to be tested to a certain extent. Then, the absolute values ​​of all the values ​​in the first-order difference sequence are calculated and averaged to obtain the first anomaly factor , which is used to illustrate the degree of fluctuation between data values ​​in the data segment to be tested. Then the mean of all values ​​in the data segment to be tested is calculated as the mean eigenvalue This value represents the average level of the values ​​in the data segment to be tested, and calculates the difference between each value in the data segment to be tested and the mean characteristic value, and serves as the difference factor The difference factor reflects the deviation between each value in the data segment to be tested and the overall average level. The larger the value, the greater the deviation, that is, the stronger the fluctuation. The second abnormal factor The larger the value is, the higher the abnormality of the data segment to be tested is.

[0059] In other embodiments of the present application, the sum of the maximum variation range value, the first abnormal factor and the second abnormal factor may also be used as the data abnormality indicator of the data segment to be tested.

[0060] Step S302: Obtain the fluctuation degree value of the target data segment according to the difference in data anomaly index between the target data segment and each corresponding comparison data segment, and the change difference between the numerical values ​​of the approximate coefficients.

[0061] According to the difference between the data anomaly index of each comparison data segment and the data anomaly index of the target data segment, the fluctuation factor between the target data segment and each comparison data segment is obtained. The fluctuation factor can be used to characterize the abnormal fluctuation characteristics between the target data segment and its corresponding comparison data segment.

[0062] Then, based on the overall numerical difference between the target data segment and each comparison data segment in each approximate coefficient, the overall volatility index between the target data segment and each comparison data segment is determined. Since the approximate coefficient reflects the characteristics of the data at different scales, the overall volatility index can comprehensively evaluate the volatility difference between the target data segment and the comparison data segment at different levels, thereby more accurately determining the volatility of the target data segment. The formula model of the overall volatility index includes: in, Indicates the target data segment The corresponding The overall volatility index between the comparison data segments; Represents the total number of approximate coefficients, that is, the number of decomposition levels; Indicated in Among the approximate coefficients, the target data segment The corresponding numerical mean; Indicated in Among the approximate coefficients, the target data segment No. The numerical mean corresponding to the comparison data segment.

[0063] In the formula model of the overall volatility index, since the target data segment is a screened-out suspected abnormal data segment, if there is a relatively obvious difference in the values ​​of the target data segment and its corresponding comparison data segment at different decomposition levels, then it can be considered that the volatility of the target data segment is relatively large, that is, the instability is relatively high.

[0064] Finally, the fluctuation factors between the target data segment and all the comparison data segments and the overall fluctuation index are weighted and averaged to obtain the fluctuation degree value of the target data segment. The formula model of the fluctuation degree value can be, for example, as follows: in, Indicates the target data segment The volatility value of Indicates the target data segment The number of comparison data segments; Indicates the target data segment The corresponding The overall volatility index between the comparison data segments; Indicates the target data segment Data anomaly indicators; Indicates the target data segment The corresponding The data anomaly indicators of the comparison data segments.

[0065] In the formula model of the fluctuation degree value, the larger the overall fluctuation index between the target data segment and the comparison data segment, the greater the fluctuation degree of the target data segment and the higher the instability of the data.

[0066] Step S4: Based on the fluctuation value of the target data segment and the corresponding environmental anomaly index, the initial threshold of each approximation coefficient is adjusted to obtain the adaptive threshold of the target data segment under each approximation coefficient; under each approximation coefficient, based on the difference between the numerical value corresponding to all suspected abnormal data segments and the adaptive threshold, as well as the difference between the numerical value of the pending data and the corresponding initial threshold, anomaly detection is performed to obtain abnormal data points.

[0067] In step S2, the environmental anomaly index is analyzed and obtained, and in step S3, the fluctuation degree value of the suspected abnormal data segment is analyzed and obtained. Therefore, the influence of environmental factors on the suspected abnormal data segment can be determined based on these two indicators, so as to adjust the initial threshold of each approximate coefficient to obtain an adaptive threshold, which is used to eliminate the influence of environmental factors, so as to facilitate the use of the adaptive threshold in the subsequent process to more accurately identify the abnormal data points caused by the inverter abnormality.

[0068] Preferably, in one embodiment of the present application, the method for obtaining the adaptive threshold includes: See also Figure 4 , which shows a flow chart of a method for obtaining an adaptive threshold provided by an embodiment of the present application, the method comprising the following steps: Step S401: Determine the environmental impact index based on the fluctuation degree value of the target data segment and the corresponding environmental anomaly index.

[0069] The value after normalizing the product of the fluctuation degree value of the target data segment and the environmental anomaly index is used as the environmental impact index. The formula model of the environmental impact index can be specifically, for example: in, Indicates the target data segment Environmental impact indicators; Indicates the target data segment The environmental anomaly index of the corresponding environmental parameter data segment; Indicates the target data segment The volatility value of Represents the normalization function.

[0070] In the formula model of environmental impact index, since environmental factors will affect the performance parameters of the inverter, causing abnormal fluctuations, and the inverter's own abnormalities will also cause abnormal fluctuations in performance parameters, in order to reduce misjudgment, that is, to make the abnormal data points finally obtained be caused by the inverter abnormality rather than the environmental factors, the environmental impact index is calculated based on the fluctuation degree value of the target data segment and the environmental abnormality index. Because the larger the environmental abnormality index, the more environmental factors are changing during the period. If the corresponding instability of the target data segment during the period is greater, then it means that the instability of the performance parameters during the period is caused by the change of environmental factors; conversely, if the environmental abnormality index is smaller and the fluctuation degree value is larger, then it means that the instability of the performance parameters during the period is caused by the abnormality of the inverter, and the possibility of being affected by environmental factors is smaller. Based on the above logic, the environmental anomaly index is multiplied by the fluctuation degree value and the product is normalized to obtain the environmental impact index of the target data segment. The larger the value, the more it indicates that the instability of the target data segment is affected by environmental factors. In the subsequent process, we should focus on eliminating this part of the impact to obtain more accurate abnormal data points.

[0071] Step S402: adjusting the initial thresholds of the respective approximation coefficients based on the environmental impact index of the target data segment to obtain adaptive thresholds of the target data segment under the respective approximation coefficients.

[0072] When the environmental impact index is greater than or equal to the preset impact threshold, the sum of the environmental impact index and the preset constant is used as the adjustment weight; when the environmental impact index is less than the preset impact threshold, the environmental impact index is used as the adjustment weight; Based on the adjustment weight, the initial threshold of each approximate coefficient is weightedly adjusted to obtain the adaptive threshold of the target data segment under each approximate coefficient. The formula model of the adaptive threshold includes: in, Indicates the target data segment In the The corresponding adaptive thresholds in the approximation coefficients; Indicates The initial threshold corresponding to the approximation coefficient; Indicates the target data segment Environmental impact indicators; Indicates a preset constant; Indicates the preset impact threshold.

[0073] In the formula model of the adaptive threshold, based on the analysis in step S401, it can be known that the larger the environmental impact index is, the more it indicates that the fluctuation of the performance parameters of the inverter during this period is caused by changes in environmental factors. The method for determining the suspected abnormal data segment in step S2 is to set the values ​​greater than the initial threshold in each approximate coefficient to 0, and then reconstruct and compare them with the performance parameter time series data. Therefore, in order to eliminate the misjudgment caused by the influence of environmental factors, the initial threshold corresponding to the period with a large environmental impact index should be increased. On the contrary, when the environmental impact index is small, it indicates that the fluctuation of the performance parameters at this time is caused by the abnormality of the inverter itself. In order to more accurately identify these abnormal data points, the initial threshold corresponding to the period can be reduced. Therefore, based on the aforementioned logic, the formula of the above adaptive threshold is constructed.

[0074] It should be noted that the value of the preset impact threshold is set to 0.5, and the specific value can be adjusted according to the implementation scenario and is not limited here; in order to prevent the adjustment degree from being too large when the initial threshold is increased, in one embodiment of the present application, the preset constant Set to 1. The specific value can be adjusted according to the implementation scenario and is not limited here.

[0075] In other embodiments of the present application, for suspected abnormal data segments whose environmental impact indicators are less than a preset environmental threshold, the initial thresholds of such suspected abnormal data segments under different approximation coefficients may be directly used as adaptive thresholds. The reason is that, under the initial thresholds, the data points in such suspected abnormal data segments can already be detected, so they do not need to be adjusted.

[0076] Based on the above process, the adaptive threshold of each suspected abnormal data segment under different approximate coefficients can be obtained. Then, under each approximate coefficient, outlier point detection can be performed based on the difference between the numerical values ​​corresponding to all suspected abnormal data segments and their respective adaptive thresholds, as well as the difference between the numerical values ​​of the pending data and the corresponding initial thresholds, thereby obtaining abnormal data points.

[0077] Preferably, in one embodiment of the present application, the method for obtaining abnormal data points includes: Under each approximate coefficient, the absolute value of each numerical value corresponding to each suspected abnormal data segment is compared with the adaptive threshold, and the numerical value whose absolute value is greater than the adaptive threshold is set to 0 to obtain the adjusted approximate coefficient data segment.

[0078] Under each approximate coefficient, the absolute value of the numerical value of the pending data is compared with the corresponding initial threshold, and the numerical value whose absolute value is greater than the initial threshold is set to 0 to obtain the adjusted pending data.

[0079] Based on the adjusted approximate coefficient data segments and the adjusted pending data under each approximate coefficient, each target approximate coefficient is obtained; at this time, each target approximate coefficient is composed of the adjusted approximate coefficient data segments and the adjusted pending data, so all the target approximate coefficients are reconstructed to obtain the second reconstructed time series data, and in the performance parameter time series data, each numerical value is subtracted from the numerical value at the corresponding moment in the second reconstructed time series data, and the data points with values ​​not equal to 0 are regarded as abnormal data points.

[0080] In other embodiments of the present application, the method for determining abnormal data points may also be: in the performance parameter time series data, after subtracting each numerical value from the numerical value at the corresponding moment in the second reconstructed time series data, the data point with a value greater than a preset deviation threshold is taken as an abnormal data point, wherein the preset deviation threshold is set to 3, and the specific numerical value can be adjusted according to the implementation scenario and is not limited here.

[0081] It should be noted that the methods of obtaining approximate coefficients based on discrete wavelet transform and reconstructing target approximate coefficients are both well-known technical means to those skilled in the art, and will not be elaborated here; at the same time, in order to facilitate calculation, all index data involved in the calculation in the embodiment of the present application are pre-processed to eliminate the dimension effect. The specific means of eliminating the dimension effect are well-known technical means to those skilled in the art, and will not be limited here.

[0082] In summary, the purpose of the embodiment of the present application is to analyze the influence of environmental factors on the performance parameters of the frequency converter, so as to set an adaptive threshold for the detection of abnormal data points, and avoid the problem of low accuracy of abnormal data point detection caused by misjudgment. Therefore, the performance parameter time series data of the frequency converter is first obtained, and the environmental parameter time series data of the same period is obtained at the same time. Then, the performance parameter time series data is multi-layered decomposed by discrete wavelet transform, and the information of different frequency components can be extracted. Since the performance changes caused by environmental changes usually have time persistence, they often appear abnormal within a period of time, so the approximate coefficient is mainly analyzed in this application, so as to understand the operating state of the frequency converter more deeply. By comparing the difference between the numerical value of the approximate coefficient and the initial threshold, the suspected abnormal data segment can be preliminarily screened out, laying the foundation for subsequent detailed abnormal detection. Then, in combination with the environmental parameter time series data, the environmental abnormality index is calculated, and it can be judged whether the abnormal data segment is related to the environmental factors in the subsequent process; at the same time, by analyzing the difference between the fluctuation of the suspected abnormal data segment and the comparison data segment, and the difference in the change of the approximate coefficient, the fluctuation degree value of the suspected abnormal data segment can be determined. Finally, based on the environmental anomaly index and fluctuation value corresponding to the suspected abnormal data segment, the initial threshold of each approximate coefficient is adaptively adjusted to obtain the adaptive threshold of the suspected abnormal data segment under each approximate coefficient. At this time, the adaptive threshold analyzes the relationship between environmental factors and the fluctuation of performance parameters, so it can more effectively eliminate the misjudged abnormal data points, that is, it effectively eliminates the situation where the performance parameters change due to changes in environmental factors. Therefore, anomaly detection based on the adaptive threshold can obtain more accurate detection results. At this time, the abnormal data point is the abnormal data point caused by the abnormality of the inverter itself.

[0083] This embodiment also provides a water conservancy project inverter status detection system, which includes a processor, a memory and a computer program, wherein the memory is used to store the corresponding computer program, and the processor is used to run the corresponding computer program. When the computer program runs on the processor, it can implement any one of the steps of a water conservancy project inverter status detection method.

[0084] See also Figure 5 , which shows a system block diagram of a water conservancy project inverter state detection system provided by an embodiment of the present application, including: a data acquisition module 501: used to implement step S1 in the above method, an environmental impact factor analysis module 502: used to implement step S2 in the above method, an inverter performance parameter fluctuation degree analysis module 503: used to implement step S3 in the above method, and an abnormal data point detection module 504: used to implement step S4 in the above method.

[0085] See also Figure 6, which shows a system structure diagram of a water conservancy project inverter status detection system provided by an embodiment of the present application, including a processor 600, a memory 601, a bus 602 and a communication interface 603, wherein the processor 600, the communication interface 603 and the memory 601 are connected via the bus 602; wherein the memory 601 may include a high-speed random access memory, the bus 602 may be an ISA bus, a PCI bus or an EISA bus, etc., and the processor 600 may be an integrated circuit chip with signal processing capabilities.

[0086] The present application also provides a computer-readable storage medium corresponding to the method provided in the above-mentioned embodiment. Figure 7 , the storage medium shown is a CD, on which a computer program (ie, a program product) is stored. When the computer program is run by a processor, it will execute the method provided by any of the aforementioned embodiments.

[0087] It should be noted that examples of the computer-readable storage medium may also include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), read-only memory (ROM), and other optical and magnetic storage media, which are not listed here one by one.

[0088] It should be noted that the above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

[0089] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A method for detecting the state of a frequency converter in a water conservancy project, characterized in that: The method comprises: Obtaining the time series data of the performance parameters of the inverter within a preset period and the time series data of the environmental parameters when the inverter is running; Based on discrete wavelet transform, the performance parameter time series data is multi-layered decomposed to obtain approximate coefficients; based on the difference between the values ​​of each approximate coefficient and the corresponding initial thresholds, the suspected abnormal data segments are determined in the performance parameter time series data; the data other than the suspected abnormal data segments are used as pending data; in the environmental parameter time series data, the environmental abnormality index is obtained according to the change of the value corresponding to the moment of each suspected abnormal data segment; Select any suspected abnormal data segment as the target data segment and determine the corresponding comparison data segment; determine the fluctuation degree value of the target data segment according to the difference between the fluctuation of the values ​​in the target data segment and all the comparison data segments, and the change difference between the values ​​of the approximate coefficients; Based on the fluctuation value of the target data segment and the corresponding environmental anomaly index, the initial threshold of each approximate coefficient is adjusted to obtain the adaptive threshold of the target data segment under each approximate coefficient; under each approximate coefficient, based on the difference between the numerical value corresponding to all suspected abnormal data segments and the adaptive threshold, as well as the difference between the numerical value of the pending data and the corresponding initial threshold, anomaly detection is performed to obtain abnormal data points.

2. A method for detecting the state of a frequency converter of a water conservancy project according to claim 1, characterized in that: The method of determining a suspected abnormal data segment in the performance parameter time series data based on the difference between the value of each approximation coefficient and the initial threshold value corresponding to each coefficient comprises: For the approximate coefficient generated by any layer of decomposition, the absolute value of each value in the approximate coefficient is compared with the corresponding initial threshold, and the value whose absolute value is greater than the corresponding initial threshold is set to 0 to obtain the adjusted approximate coefficient; Reconstruct all adjusted approximate coefficients to obtain first reconstructed time series data, and in the performance parameter time series data, subtract each value from the value at the corresponding moment in the first reconstructed time series data, and take the data point whose value is not 0 as the target point; The target points with continuous time series are grouped into suspected abnormal data segments, and all the suspected abnormal data segments in the performance parameter time series data are obtained.

3. A method for detecting the state of a frequency converter of a water conservancy project according to claim 1, characterized in that: The method for obtaining the environmental abnormality indicator includes: In the environmental parameter time series data, the corresponding environmental parameter data segment is determined according to the time of each suspected abnormal data segment; For any environmental parameter data segment, a first-order difference sequence of values ​​in the environmental parameter data segment is calculated, and the values ​​in the first-order difference sequence are averaged to obtain a first environmental parameter; Determine a second environmental parameter of the environmental parameter data segment based on a difference between a numerical mean value in the environmental parameter data segment and a numerical mean value of the environmental parameter time series data; Determine a third environmental parameter of the environmental parameter data segment based on a difference between a numerical mean and a numerical maximum in the environmental parameter data segment; The first environmental parameter, the second environmental parameter and the third environmental parameter are integrated to obtain an environmental anomaly index of the environmental parameter data segment, and the first environmental parameter, the second environmental parameter and the third environmental parameter are all positively correlated with the environmental anomaly index.

4. A method for detecting the state of a frequency converter of a water conservancy project according to claim 1, characterized in that: The method for obtaining the fluctuation degree value includes: Determining data anomaly indicators based on fluctuations of values ​​in the target data segment and each comparison data segment respectively; According to the difference between the data anomaly index of each comparison data segment and the data anomaly index of the target data segment, a fluctuation factor between the target data segment and each comparison data segment is obtained; Determine an overall fluctuation index between the target data segment and each of the comparison data segments based on the numerical differences between the target data segment and each of the comparison data segments in each of the approximation coefficients; The volatility factors between the target data segment and all the comparison data segments and the overall volatility index are weighted and averaged to obtain the volatility value of the target data segment.

5. A method for detecting the state of a frequency converter of a water conservancy project according to claim 4, characterized in that: The method for obtaining the data anomaly indicator includes: Select one of the target data segment and the corresponding comparison data segment as the data segment to be tested; Based on the values ​​in the data segment to be tested, determine the maximum variation range of the values ​​in the data segment to be tested; calculate the first-order difference sequence of the values ​​in the data segment to be tested, and average the absolute values ​​of the values ​​in the first-order difference sequence to obtain a first abnormal factor; use the mean of the values ​​in the data segment to be tested as the mean characteristic value, use the difference between each value in the data segment to be tested and the mean characteristic value as the difference factor, and use the mean of all the difference factors as the second abnormal factor; Based on the maximum variation range value, the first abnormal factor and the second abnormal factor of the data segment to be tested, a data abnormality index of the data segment to be tested is determined, and the maximum variation range value, the first abnormal factor and the second abnormal factor all change in the same direction as the data abnormality index.

6. A method for detecting the state of a frequency converter of a water conservancy project according to claim 1, characterized in that: The method for obtaining the adaptive threshold comprises: Determine the environmental impact index based on the fluctuation degree value of the target data segment and the environmental anomaly index; When the environmental impact index is greater than or equal to the preset impact threshold, the sum of the environmental impact index and the preset constant is used as the adjustment weight; when the environmental impact index is less than the preset impact threshold, the environmental impact index is used as the adjustment weight; The initial thresholds of the respective approximation coefficients are weightedly adjusted based on the adjustment weights to obtain adaptive thresholds of the target data segment under the respective approximation coefficients.

7. A method for detecting the state of a frequency converter of a water conservancy project according to claim 6, characterized in that: The method for obtaining the environmental impact index includes: The value obtained by normalizing the product of the fluctuation degree value of the target data segment and the corresponding environmental anomaly index is used as the environmental impact index.

8. A method for detecting the state of a frequency converter in a water conservancy project according to claim 1, characterized in that: The method for obtaining abnormal data points includes: Under each approximate coefficient, the absolute value of each numerical value corresponding to each suspected abnormal data segment is compared with the adaptive threshold, and the numerical value whose absolute value is greater than the adaptive threshold is set to 0 to obtain the adjusted approximate coefficient data segment; Under each approximate coefficient, the absolute value of the value of the pending data is compared with the corresponding initial threshold, and the value whose absolute value is greater than the initial threshold is set to 0 to obtain the adjusted pending data; Based on the adjusted approximate coefficient data segments under each approximate coefficient and the adjusted pending data, each target approximate coefficient is obtained; all target approximate coefficients are reconstructed to obtain second reconstructed time series data, and in the performance parameter time series data, each numerical value is subtracted from the numerical value at the corresponding moment in the second reconstructed time series data, and data points with values ​​not equal to 0 are regarded as abnormal data points.

9. A method for detecting the state of a frequency converter of a water conservancy project according to claim 1, characterized in that: The method for obtaining the initial threshold value includes: Calculate the standard deviation of all values ​​of each approximation coefficient; The mean of all values ​​of each approximation coefficient plus three times the corresponding standard deviation is used as the initial threshold of each approximation coefficient.

10. A water conservancy project inverter state detection system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of a method for detecting the state of a frequency converter for a water conservancy project as described in any one of claims 1 to 9 are implemented.

Images