Power equipment fault prediction method based on vibration analysis

Abstract

The invention discloses a power equipment fault prediction method based on vibration analysis. The method comprises the following steps: collecting vibration signal data of power equipment; preprocessing the vibration signal data and converting the vibration signal data into vibration analysis data; segmenting the vibration analysis data into a plurality of sub-segments according to a time sequence, the sub-segments including continuous vibration analysis data points, and extracting feature data from the vibration analysis data in the sub-segments; according to the feature data in the sub-segments, calculating dynamic adjustment coefficients of the sub-segments, performing dynamic smoothing processing on the vibration analysis data in the sub-segments by using the dynamic adjustment coefficients of the sub-segments, and calculating a deviation value between a predicted value and an actual value of the vibration analysis data in each sub-segment; and presetting a deviation value threshold value, comparing the deviation value with the deviation value threshold value, and judging whether the vibration analysis data in the sub-segment is abnormal or not. According to the invention, the early warning of the power equipment fault is realized, the operation safety and reliability of the power system are improved, the accuracy is higher, and powerful support is provided for the monitoring and management of the power equipment.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis and monitoring, and in particular to a method for predicting faults of electric power equipment based on vibration analysis. Background Art

[0002] With the continuous expansion of the scale and increase in complexity of the power system, the stable operation of power equipment is crucial to ensure the quality and safety of power supply. However, various faults may occur in power equipment during long-term operation, which may not only cause power supply interruption, but also pose a serious threat to the security of the power grid. In order to timely discover and prevent potential faults and improve the operational reliability of power equipment, it is particularly important to develop efficient and accurate fault prediction technology. Therefore, the present invention aims to provide a power equipment fault prediction method based on vibration analysis to achieve effective detection and early warning of early faults of power equipment.

[0003] Traditional power equipment fault detection methods mainly rely on regular manual inspections and simple online monitoring methods, such as temperature monitoring and oil quality analysis. Although these methods can reflect the operating status of the equipment to a certain extent, due to the lack of in-depth analysis of the internal status of the equipment, it is often difficult to detect early faults in a timely manner. In addition, manual inspections are time-consuming, labor-intensive, costly, and have certain subjectivity and limitations; simple online monitoring methods can usually only reflect the state changes on the surface of the equipment and cannot fully evaluate the health of the equipment.

[0004] In recent years, with the development of sensor technology and data analysis technology, fault diagnosis technology based on vibration analysis has gradually become a research hotspot. In the prior art, there are already some methods that use vibration signals for fault prediction. These methods usually collect vibration signals generated during equipment operation and analyze and process them to identify abnormal conditions of the equipment. However, most of the existing vibration analysis methods focus on the extraction of static features, and pay insufficient attention to the dynamic changes of vibration signals, resulting in poor accuracy and real-time performance of fault prediction. In addition, when processing vibration signals, the prior art often ignores the importance of different frequency components, which makes the fault prediction model insufficiently robust.

[0005] In view of the problems existing in the prior art, the present invention proposes a new method for predicting faults of power equipment based on vibration analysis. The method not only considers the static characteristics of the vibration signal, but also introduces a dynamic adjustment coefficient to perform dynamic smoothing on the vibration signal, thereby improving the accuracy and stability of fault prediction. Summary of the invention

[0006] A method for predicting faults of electric power equipment based on vibration analysis, characterized in that it comprises the following steps:

[0007] Collect vibration signal data of power equipment;

[0008] Preprocess the vibration signal data, and transform the vibration signal data into vibration analysis data after filtering, noise reduction and standardization;

[0009] The vibration analysis data is divided into multiple sub-segments according to the time series, and the sub-segments contain continuous vibration analysis data points, and feature data is extracted from the vibration analysis data in the sub-segments;

[0010] According to the characteristic data in the sub-segment, the dynamic adjustment coefficient of the sub-segment is calculated, and the vibration analysis data in the sub-segment is dynamically smoothed by using the dynamic adjustment coefficient of the sub-segment, and the deviation value between the predicted value and the actual value of the vibration analysis data in each sub-segment is calculated;

[0011] A deviation value threshold is preset, and the deviation value is compared with the deviation value threshold to determine whether the vibration analysis data in the sub-segment is abnormal.

[0012] Preferably, the preprocessing of the vibration signal data is performed by filtering, noise reduction, and standardization to transform the vibration analysis data;

[0013] The filtering removes high-frequency noise and low-frequency drift in the vibration signal data through a bandpass filter, and retains the vibration signal frequency band;

[0014] The noise reduction is performed by using a wavelet transform method to reduce the noise of the filtered vibration signal data, further reducing the influence of random noise;

[0015] The standardization process normalizes the vibration signal data after noise reduction so that the mean value of the data is 0 and the standard deviation is 1, thereby generating vibration analysis data.

[0016] Preferably, the vibration analysis data is divided into a plurality of sub-segments according to a time series, comprising the following steps:

[0017] Determine the sub-segment length L and step length S, where L is the number of vibration analysis data points contained in each sub-segment, and S is the number of interval data points between adjacent sub-segments;

[0018] Starting from the starting point of the vibration analysis data, sub-segments of length L are selected in sequence according to the step length S to form multiple sub-segments; the starting point position P of the i-th sub-segment i Calculated by the following formula:

[0019] P i =(i-1)×S

[0020] Where i is the sequence number of the sub-segment;

[0021] The end point position Q of the i-th sub-segment i Calculated by the following formula:

[0022] Q i =P i +L-1

[0023] Among them, Q i is the position of the last data point of the i-th subsegment.

[0024] Preferably, the feature data extracted from the vibration analysis data in the sub-segment includes frequency domain features and time domain features;

[0025] The frequency domain feature is obtained by performing fast Fourier transform on the vibration analysis data in the sub-segment to obtain frequency domain data X(f):

[0026]

[0027] Where x(t) is the vibration analysis data of the tth data point in the subsegment, F is the frequency, and j is the imaginary unit;

[0028] Calculate the frequency domain energy E of the vibration analysis data within the sub-segment i :

[0029]

[0030] Calculate the main frequency f0 of the vibration analysis data in the sub-segment through frequency domain data:

[0031]

[0032] The main frequency F0 marks the characteristic vibration frequency of the vibration analysis signal.

[0033] Preferably, the time domain feature comprises the following steps:

[0034] Calculate the mean μ of the vibration analysis data within the sub-segment i :

[0035]

[0036] Where x(t) is the vibration analysis data of the tth data point in the sub-segment, L is the length of the sub-segment, P i and Q i are the starting and ending points of the i-th sub-segment respectively;

[0037] The mean μ of the vibration analysis data i Calculate the standard deviation σ of the vibration analysis data within the sub-segment i ;

[0038]

[0039] The peak PP i Calculate through the vibration data in the sub-segment, the formula is as follows:

[0040]

[0041] The time domain characteristic data is the mean value μ of the vibration analysis data. i , standard deviation σ i and peak PP i , achieving the key eigenvalues ​​in the time domain features.

[0042] Preferably, the dynamic adjustment coefficient of the sub-segment comprises the following steps:

[0043] The mean μ of the vibration analysis data within the sub-segment i , standard deviation σ i and peak PP i Calculate the time domain feature comprehensive index T of the sub-segment i ;

[0044] T i =w1·μ i +w2·σ i +w3·PP i

[0045] Among them, w1, w2, and w3 are preset weight coefficients respectively;

[0046] The main frequency f0 and frequency domain energy E of the vibration analysis data in the sub-segment i Calculate the frequency domain feature comprehensive index F within the sub-segment i ;

[0047] F i =w4·E i +w5·f0

[0048] Among them, w4 and w5 are preset weight coefficients respectively;

[0049] According to the time domain characteristics of the sub-segment comprehensive index T i And the frequency domain feature comprehensive index F i Calculate the comprehensive characteristic index C of the sub-segment i ;

[0050] C i =α·T i +(1-α)·F i

[0051] Among them, α is the comprehensive weight of the preset time domain features and frequency domain features;

[0052] According to the comprehensive characteristic index C of the sub-segment i Calculate the dynamic adjustment coefficient β of the sub-segment i ;

[0053]

[0054] Among them, θ is the preset threshold.

[0055] Preferably, the dynamic smoothing process is performed on the vibration analysis data in the sub-segment by using the dynamic adjustment coefficient of the sub-segment, and the dynamic smoothing process is performed on the vibration analysis data in each sub-segment i to calculate the smoothing value S of each data point t. i (t), the dynamic smoothing processing formula is:

[0056] S i (t) = β i ·x(t)+(1-β i )·S i-1 (t)

[0057] Among them, S i-1 (t) is the average value of the previous sub-segment at time t.

[0058] Preferably, the deviation value is calculated by using the predicted value and actual value of the vibration analysis data in the sub-segment, and the predicted value of the vibration analysis data in the sub-segment for:

[0059]

[0060] The deviation value D i (t) is calculated by the difference between the predicted value and the actual value, and the formula is:

[0061]

[0062] Among them, D i (t) is the deviation value of the i-th sub-segment at time t.

[0063] Preferably, the deviation value threshold comprises the following steps:

[0064] Calculate the average deviation value of all sub-segments

[0065]

[0066] Where N is the total number of sub-segments;

[0067] Calculate the standard deviation σ of the deviation values ​​of all sub-segments D ;

[0068]

[0069] By the mean deviation and the standard deviation σ of the deviation value D Preset deviation threshold Δ:

[0070]

[0071] Wherein, k is a preset multiple coefficient.

[0072] Preferably, the vibration analysis data in the sub-segment is obtained by using the deviation value D i (t) comparing with a preset deviation value threshold Δ to determine whether an abnormality exists;

[0073] If D i (t)>Δ, it is determined that the vibration data in the sub-segment is abnormal and the power equipment is faulty;

[0074] If D i (t)≤Δ, it is determined that the vibration data in the sub-segment is normal and the power equipment is not faulty.

[0075] Compared with the prior art, the technical solution of this application has the following technical effects:

[0076] The present invention solves the noise interference problem through the preprocessing technology of vibration signals and obtains more accurate vibration analysis data. The high-frequency noise and low-frequency drift in the vibration signal are removed by a bandpass filter, and the main frequency band of the vibration signal is retained; the filtered vibration signal is then denoised by the wavelet transform method to further reduce the influence of random noise, and the denoised vibration signal data is normalized so that the mean of the data is 0 and the standard deviation is 1, and vibration analysis data is generated. This series of preprocessing steps effectively eliminates noise interference and improves the quality of vibration signals, thereby providing a more accurate data basis for subsequent feature extraction and fault prediction.

[0077] The present invention solves the problem of data processing efficiency and improves the real-time performance of fault prediction by using the time series segmentation technology of vibration analysis data. The vibration analysis data is segmented into multiple sub-segments according to the time series, each sub-segment contains continuous vibration analysis data points, and the appropriate sub-segment length and step length are determined. In this way, a large amount of vibration data can be processed in segments, which reduces the amount of data processed in a single time and improves the efficiency of data processing. At the same time, the division of sub-segments also helps to capture the local characteristics of the vibration signal, improving the real-time performance and accuracy of fault prediction.

[0078] The present invention solves the adaptability problem of dynamic changes of vibration signals and improves the accuracy of fault prediction by calculating and applying the dynamic adjustment coefficient. The comprehensive characteristic index of the sub-segment is obtained by calculating the comprehensive index of the time domain characteristics and the comprehensive index of the frequency domain characteristics of the sub-segment, and the dynamic adjustment coefficient of the sub-segment is calculated based on the comprehensive characteristic index. The dynamic smoothing of the vibration analysis data in the sub-segment using the dynamic adjustment coefficient can better adapt to the dynamic changes of the vibration signal and reduce the misjudgment caused by the signal mutation. This technical solution significantly improves the accuracy and robustness of fault prediction.

[0079] The present invention solves the problem of fault detection sensitivity and achieves early fault warning by comparing the deviation value with the preset threshold value. By calculating the deviation value between the predicted value and the actual value of the vibration analysis data in each sub-segment and comparing it with the preset deviation value threshold, it is determined whether the vibration data in the sub-segment is abnormal. This method can detect small vibration changes in time, thereby achieving early warning of power equipment failures. Compared with traditional fault detection methods, the technical solution of the present invention has higher sensitivity and reliability, can effectively prevent potential fault risks, and ensure the safe and stable operation of the power system.

[0080] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application so that it can be implemented in accordance with the contents of the specification, and to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the following is a detailed description of the preferred embodiments of the present application in conjunction with the accompanying drawings as follows.

[0081] Based on the detailed description of the specific embodiments of the present application in combination with the accompanying drawings below, those skilled in the art will become more aware of the above and other objects, advantages and features of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] In order to more clearly illustrate the technical solutions 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 some embodiments of the present application. For ordinary technicians in this field, other drawings can also be obtained based on these drawings without creative work. In all drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, each element or part is not necessarily drawn according to the actual scale.

[0083] Figure 1 It is the system flow chart. Specific embodiments

[0084] To make the purpose, technical scheme and advantages of the embodiment of the present application clearer, the technical scheme in the embodiment of the present application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiment of the present application. Obviously, the described embodiment is a part of the embodiment of the present application, rather than all of the embodiments. In the following description, specific details such as specific configuration and components are provided only to help fully understand the embodiments of the present application. Therefore, it should be clear to those skilled in the art that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present application. In addition, for clarity and brevity, the description of known functions and structures is omitted in the embodiment.

[0085] It should be understood that the references to "one embodiment" or "this embodiment" throughout the specification mean that the specific features, structures, or characteristics associated with the embodiment are included in at least one embodiment of the present application. Therefore, the references to "one embodiment" or "this embodiment" appearing throughout the specification do not necessarily refer to the same embodiment. In addition, these specific features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0086] In addition, the present application may repeat reference numerals and / or letters in different examples. This repetition is for the purpose of simplicity and clarity, and does not in itself indicate the relationship between the various embodiments and / or settings discussed.

[0087] The term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, B exists alone, and A and B exist at the same time. The term " / and" in this article describes another type of association object relationship, indicating that there can be two relationships. For example, A / and B can mean: A exists alone, and A and B exist alone. In addition, the character " / " in this article generally indicates that the previous and next associated objects are in an "or" relationship.

[0088] The term "at least one" in this article is merely a description of the association relationship of associated objects, indicating that there may be three relationships. For example, at least one of A and B can mean: A exists alone, A and B exist at the same time, and B exists alone.

[0089] It should also be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusions.

[0090] Example 1

[0091] This embodiment mainly describes a method for predicting power equipment faults based on vibration analysis. Figure 1 As shown, the following steps are included:

[0092] S1: Collect vibration signal data of power equipment; pre-process the vibration signal data, and transform the vibration signal data into vibration analysis data after filtering, noise reduction and standardization;

[0093] Vibration sensors are used to continuously monitor the vibration signal data of power equipment in real time. Through advanced data acquisition system integration, they are precisely configured to set the data acquisition interval per second. The high frequency of data acquisition ensures the integrity and real-time nature of the collected vibration signal data, and can effectively capture the slight changes in the vibration of power equipment, providing reliable basic data for subsequent fault prediction.

[0094] Preprocess the vibration signal data

[0095] In order to improve the accuracy and reliability of subsequent analysis, the collected vibration signal data is preprocessed.

[0096] The preprocessing steps include filtering, noise reduction and standardization, as follows:

[0097] Filtering: Remove high-frequency noise and low-frequency drift in the vibration signal data, retain the main frequency band of the vibration signal, and reduce the interference of irrelevant signals. Use a bandpass filter to filter the vibration signal data. The bandpass filter can effectively remove high-frequency noise and low-frequency drift, and retain the vibration signal in the intermediate frequency range.

[0098] The filter processing selects the cutoff frequency according to the operating frequency range of the power equipment. If the operating frequency range of the power equipment is 50Hz to 1kHz, select a bandpass filter with a passband range of 50Hz to 1kHz. Apply the selected bandpass filter to the original vibration signal data to remove high-frequency noise and low-frequency drift and retain the main vibration signal.

[0099] Noise reduction: Reduce the impact of random noise, improve the signal-to-noise ratio, and use the wavelet transform method to perform noise reduction on the filtered vibration signal data. Wavelet transform is a multi-resolution analysis method that can effectively separate signals and noise.

[0100] Use wavelet basis function: According to the characteristics of vibration signal, select appropriate wavelet basis function, such as Db4, Sym8; perform wavelet decomposition: perform multi-scale wavelet decomposition on the filtered vibration signal data to obtain wavelet coefficients at different scales; threshold processing: perform threshold processing on wavelet coefficients to remove noise components in wavelet coefficients; commonly used threshold processing methods include hard threshold method and soft threshold method. Reconstruct signal: Reconstruct the vibration signal according to the processed wavelet coefficients to obtain the vibration signal data after noise reduction.

[0101] The standardization process normalizes the vibration signal data after noise reduction so that the mean value of the data is 0 and the standard deviation is 1, thereby generating vibration analysis data.

[0102] The vibration analysis data generated by S1 provides high-quality basic data for subsequent feature extraction and fault prediction, ensuring the accuracy and reliability of fault prediction.

[0103] S2: Divide the vibration analysis data into multiple sub-segments according to the time series, each sub-segment contains continuous vibration analysis data points, and extract feature data from the vibration analysis data in the sub-segment;

[0104] In the power equipment fault prediction method based on vibration analysis, a key step is to divide the vibration analysis data into multiple sub-segments according to the time series. This process is to better capture the characteristics of the vibration signal changing over time, so as to more accurately identify potential failure modes. Specifically, this step includes the following aspects:

[0105] Determine the sub-segment length and step size: First, you need to determine the number of vibration analysis data points contained in each sub-segment (i.e., the sub-segment length) and the number of data points between adjacent sub-segments (i.e., the step size). For example, if 1000 vibration data points are collected per second, the sub-segment length can be selected as 1000 data points (equivalent to 1 second of data), and the step size can be set to 500 data points, which means that each time half the length of the sub-segment is moved to form the next sub-segment. This choice depends on the requirements for time resolution in the actual application and the availability of computing resources.

[0106] Determination of sub-segments: After determining the sub-segment length and step length, you can start from the starting point of the vibration analysis data and select data points of corresponding lengths in sequence according to the set step length to form multiple sub-segments. For example, if we have a vibration signal containing 10,000 data points, if the above sub-segment length and step length are used, 10 sub-segments can be formed, each containing 1,000 data points.

[0107] Sub-segment position calculation: Starting from the starting point of the vibration analysis data, sub-segments of length L are selected in sequence according to the step length S to form multiple sub-segments;

[0108] The starting point position P of the i-th sub-segment i Calculated by the following formula:

[0109] P i =(i-1)×S

[0110] Where i is the sequence number of the sub-segment;

[0111] The end point position Q of the i-th sub-segment i Calculated by the following formula:

[0112] Q i =P i +L-1

[0113] Among them, Qi is the position of the last data point of the i-th subsegment;

[0114] After the vibration analysis data is divided into multiple sub-segments, feature extraction is performed on the data in each sub-segment. This step aims to extract key information that can reflect the status of the equipment from a large amount of raw data. Feature extraction usually includes two parts: time domain features and frequency domain features:

[0115] Time domain features: The main focus is on the statistical characteristics of the vibration signal on the time axis, such as the mean, standard deviation, and peak value. These features can directly reflect the intensity and fluctuation of the vibration signal and are an important basis for evaluating the health of the equipment. For example, in a certain sub-segment, if the peak value of the vibration signal is found to be abnormally high, it may mean that the equipment has been subjected to a large shock or vibration during this period of time.

[0116] Frequency domain features: The time domain signal is converted to the frequency domain through the fast Fourier transform (FFT), and then the frequency domain energy and main frequency characteristics are calculated. Frequency domain features can reveal the frequency composition of the vibration signal and are particularly useful for diagnosing specific faults in rotating machinery. For example, motor bearing faults often generate abnormally high energy at specific frequencies, and these problems can be located by analyzing frequency domain features.

[0117] For time domain features, the following steps are included:

[0118] Calculate the mean μ of the vibration analysis data within the sub-segment i :

[0119]

[0120] Where x(t) is the vibration analysis data of the tth data point in the sub-segment, L is the length of the sub-segment, P i and Q i are the starting and ending points of the i-th sub-segment respectively;

[0121] The mean μ of the vibration analysis data i Calculate the standard deviation σ of the vibration analysis data within the sub-segment i ;

[0122]

[0123] The peak PP i Calculate through the vibration data in the sub-segment, the formula is as follows:

[0124]

[0125] The time domain characteristic data is the mean value μ of the vibration analysis data. i , standard deviation σ i and peak PPi , achieving the key eigenvalues ​​in the time domain features.

[0126] For the frequency domain feature, the frequency domain feature is obtained by performing fast Fourier transform on the vibration analysis data in the sub-segment to obtain the frequency domain data X(f):

[0127]

[0128] Where x(t) is the vibration analysis data of the tth data point in the subsegment, f is the frequency, and j is the imaginary unit;

[0129] Calculate the frequency domain energy E of the vibration analysis data within the sub-segment i :

[0130]

[0131] Calculate the main frequency f0 of the vibration analysis data in the sub-segment through frequency domain data:

[0132]

[0133] The main frequency f0 marks the characteristic vibration frequency of the vibration analysis signal.

[0134] Through detailed analysis of vibration data, signs of equipment performance degradation can be detected earlier, and timely measures can be taken to avoid failures, thereby significantly reducing maintenance costs and downtime. By adopting a combination of multiple features, a high fault recognition rate can be maintained even when the signal is interfered with to a certain extent.

[0135] S3: Calculate the dynamic adjustment coefficient of the sub-segment according to the characteristic data in the sub-segment, use the dynamic adjustment coefficient of the sub-segment to dynamically smooth the vibration analysis data in the sub-segment, and calculate the deviation between the predicted value and the actual value of the vibration analysis data in each sub-segment;

[0136] The dynamic adjustment coefficient of the sub-segment is calculated by the mean value μ of the vibration analysis data in the sub-segment. i , standard deviation σ i and peak PP i Calculate the time domain feature comprehensive index T of the sub-segment i ;

[0137] T i =w1·μ i +w2·σ i +w3·PP i

[0138] Among them, w1, w2, w3 are preset weight coefficients respectively, and w1+w2+w3=1.

[0139] The main frequency f0 and frequency domain energy E of the vibration analysis data in the sub-segment i Calculate the frequency domain feature comprehensive index F within the sub-segment i ;

[0140] F i =w4·E i +w5·f0

[0141] Among them, w4 and w5 are preset weight coefficients respectively; among them, w4+w5=1.

[0142] According to the time domain characteristics of the sub-segment comprehensive index T i And the frequency domain feature comprehensive index F i Calculate the comprehensive characteristic index C of the sub-segment i ;

[0143] C i =α·T i +(1-α)·F i

[0144] Among them, α is the comprehensive weight of the preset time domain features and frequency domain features;

[0145] According to the comprehensive characteristic index C of the sub-segment i Calculate the dynamic adjustment coefficient β of the sub-segment i ;

[0146]

[0147] Among them, θ is the preset threshold.

[0148] The vibration analysis data in the sub-segment is dynamically smoothed by using the dynamic adjustment coefficient of the sub-segment. The vibration analysis data in each sub-segment i is dynamically smoothed to calculate the smoothing value S of each data point t. i (t), the dynamic smoothing processing formula is:

[0149] S i (t) = β i ·x(t)+(1-β i )·S i-1 (t)

[0150] Among them, S i-1 (t) is the average value of the previous sub-segment at time t.

[0151] The deviation value is used to calculate the predicted value and actual value of the vibration analysis data in the sub-segment. for:

[0152]

[0153] The deviation value D i (t) is calculated by the difference between the predicted value and the actual value, and the formula is:

[0154]

[0155] Among them, D i (t) is the deviation value of the i-th sub-segment at time t.

[0156] Dynamic smoothing can effectively reduce random noise in the data, making the signal smoother and easier to identify real fault characteristics. The smoothed data is more stable, which helps to improve the accuracy of the prediction model and enable more accurate monitoring of potential faults in power equipment.

[0157] S4: preset a deviation value threshold, compare the deviation value with the deviation value threshold, and determine whether the vibration analysis data in the sub-segment is abnormal;

[0158] Presetting the deviation value threshold and comparing the deviation value with the deviation value threshold is the key to judging whether the vibration analysis data in the sub-segment is abnormal; the deviation value threshold includes the following steps:

[0159] Calculate the average deviation value of all sub-segments

[0160]

[0161] Where N is the total number of sub-segments;

[0162] Calculate the standard deviation σ of the deviation values ​​of all sub-segments D ;

[0163]

[0164] By the mean deviation and the standard deviation σ of the deviation value D Preset deviation threshold Δ:

[0165]

[0166] Wherein, k is a preset multiple coefficient.

[0167] The vibration analysis data in the sub-segment is calculated by the deviation value D i (t) comparing with a preset deviation value threshold Δ to determine whether an abnormality exists;

[0168] If D i (t)>Δ, it is determined that the vibration data in the sub-segment is abnormal and the power equipment is faulty;

[0169] If D i(t)≤Δ, it is determined that the vibration data in the sub-segment is normal and the power equipment is not faulty.

[0170] At the same time, any abnormality in the vibration analysis data within any sub-segment is judged as a machine failure, further ensuring the safety of the power equipment.

[0171] Through the above implementation steps, this embodiment can more accurately determine whether there is a potential fault in the power equipment, thereby achieving early warning of power equipment failure and improving the operational safety and reliability of the power system. Compared with traditional methods, the technical solution of the present invention has higher accuracy and real-time performance, providing strong support for the health management of power equipment.

[0172] The above are only preferred embodiments of the present invention, which do not limit the scope of protection of the present invention. For those skilled in the art, the present invention may have various modifications and changes. Any changes, modifications, replacements, integrations and parameter changes to these embodiments within the spirit and principles of the present invention through conventional substitutions or without departing from the principles and spirit of the present invention fall within the scope of protection of the present invention.

Claims

1. A method for predicting faults of power equipment based on vibration analysis, characterized in that: The following steps are involved: Collect vibration signal data of power equipment; Preprocess the vibration signal data, and transform the vibration signal data into vibration analysis data after filtering, noise reduction and standardization; The vibration analysis data is divided into multiple sub-segments according to the time series, and the sub-segments contain continuous vibration analysis data points, and feature data is extracted from the vibration analysis data in the sub-segments; According to the characteristic data in the sub-segment, the dynamic adjustment coefficient of the sub-segment is calculated, and the vibration analysis data in the sub-segment is dynamically smoothed by using the dynamic adjustment coefficient of the sub-segment, and the deviation value between the predicted value and the actual value of the vibration analysis data in each sub-segment is calculated; A deviation value threshold is preset, and the deviation value is compared with the deviation value threshold to determine whether the vibration analysis data in the sub-segment is abnormal.

2. A method for predicting faults of electric power equipment based on vibration analysis according to claim 1, characterized in that: The preprocessing of the vibration signal data is performed by filtering, noise reduction and standardization to transform it into vibration analysis data; The filtering removes high-frequency noise and low-frequency drift in the vibration signal data through a bandpass filter, and retains the vibration signal frequency band; The noise reduction is performed by using a wavelet transform method to reduce the noise of the filtered vibration signal data, further reducing the influence of random noise; The standardization process normalizes the vibration signal data after noise reduction so that the mean value of the data is 0 and the standard deviation is 1, thereby generating vibration analysis data.

3. A method for predicting faults of electric power equipment based on vibration analysis according to claim 2, characterized in that: The vibration analysis data is divided into a plurality of sub-segments according to a time series, comprising the following steps: Determine the sub-segment length L and step length S, where L is the number of vibration analysis data points contained in each sub-segment, and S is the number of interval data points between adjacent sub-segments; Starting from the starting point of the vibration analysis data, sub-segments of length L are selected in sequence according to the step length S to form multiple sub-segments; The starting point position P of the i-th sub-segment i Calculated by the following formula: P i =(i-1)×S Where i is the sequence number of the sub-segment; The end point position Q of the i-th sub-segment i Calculated by the following formula: Q i =P i +L-1 Among them, Q i is the position of the last data point of the i-th subsegment.

4. A method for predicting faults of electric power equipment based on vibration analysis according to claim 3, characterized in that: Extracting feature data from the vibration analysis data in the sub-segment includes frequency domain features and time domain features; The frequency domain feature is obtained by performing fast Fourier transform on the vibration analysis data in the sub-segment to obtain frequency domain data X(f): Where x(t) is the vibration analysis data of the tth data point in the subsegment, f is the frequency, and j is the imaginary unit; Calculate the frequency domain energy E of the vibration analysis data within the sub-segment i : Calculate the main frequency F0 of the vibration analysis data in the sub-segment through frequency domain data: The main frequency f0 marks the characteristic vibration frequency of the vibration analysis signal.

5. The method for predicting faults of electric power equipment based on vibration analysis according to claim 4, characterized in that: The time domain characterization comprises the following steps: Calculate the mean μ of the vibration analysis data within the sub-segment i : Where x(t) is the vibration analysis data of the tth data point in the sub-segment, L is the length of the sub-segment, P i and Q i are the starting and ending points of the i-th sub-segment respectively; The mean μ of the vibration analysis data i Calculate the standard deviation σ of the vibration analysis data within the sub-segment i ; The peak PP i Calculate through the vibration data in the sub-segment, the formula is as follows: The time domain characteristic data is the mean value μ of the vibration analysis data. i , standard deviation σ i and peak PP i , achieving the key eigenvalues ​​in the time domain features.

6. The method for predicting faults of electric power equipment based on vibration analysis according to claim 1, characterized in that: The dynamic adjustment coefficient of the sub-segment comprises the following steps: The mean μ of the vibration analysis data within the sub-segment i , standard deviation σ i and peak PP i Calculate the time domain feature comprehensive index T of the sub-segment i ; T i =w1·μ i +w2·s i +w3·PP i Among them, w1, w2, and w3 are preset weight coefficients respectively; The main frequency f0 and frequency domain energy E of the vibration analysis data in the sub-segment i Calculate the frequency domain feature comprehensive index F within the sub-segment i ; F i =w4·E i +w5·f0 Among them, w4 and w5 are preset weight coefficients respectively; According to the time domain characteristics of the sub-segment comprehensive index T i And the frequency domain feature comprehensive index F i Calculate the comprehensive characteristic index C of the sub-segment i ; C i =α·T i +(1-α)·F i Among them, α is the comprehensive weight of the preset time domain features and frequency domain features; According to the comprehensive characteristic index C of the sub-segment i Calculate the dynamic adjustment coefficient β of the sub-segment i ; Among them, θ is the preset threshold.

7. A method for predicting faults of electric power equipment based on vibration analysis according to claim 6, characterized in that: The dynamic smoothing process is performed on the vibration analysis data in the sub-segment by using the dynamic adjustment coefficient of the sub-segment. The dynamic smoothing process is performed on the vibration analysis data in each sub-segment i, and the smoothing value S of each data point t is calculated. i (t), the dynamic smoothing processing formula is: S i (t)=β i ·x(t)+(1-β i )·S i-1 (t) Among them, S i-1 (t) is the average value of the previous sub-segment at time t.

8. The method for predicting faults of electric power equipment based on vibration analysis according to claim 1, characterized in that: The deviation value is calculated by using the predicted value and actual value of the vibration analysis data in the sub-segment. for: The deviation value D i (t) is calculated by the difference between the predicted value and the actual value, and the formula is: Among them, D i (t) is the deviation value of the i-th sub-segment at time t.

9. The method for predicting faults of electric power equipment based on vibration analysis according to claim 1, characterized in that: The deviation value threshold comprises the following steps: Calculate the average deviation value of all sub-segments Where D is the total number of sub-segments; Calculate the standard deviation σ of the deviation values ​​of all sub-segments D ; By the mean deviation and the standard deviation σ of the deviation value D Preset deviation threshold Δ: Wherein, k is a preset multiple coefficient.

10. A method for predicting faults of electric power equipment based on vibration analysis according to claim 8 or 9, characterized in that: The vibration analysis data in the sub-segment is determined by the deviation value D i (t) comparing with a preset deviation value threshold Δ to determine whether an abnormality exists; If D i (t)>Δ, it is determined that the vibration data in the sub-segment is abnormal and the power equipment is faulty; If D i (t)≤Δ, it is determined that the vibration data in the sub-segment is normal and the power equipment is not faulty.

Images