Dedusting fan fault prediction method and device based on hybrid rice algorithm

Abstract

The invention relates to the technical field of fault detection, and discloses a dust removal fan fault prediction method and device based on a hybrid rice algorithm, and the method comprises the steps: carrying out the feature extraction of historical vibration data, and obtaining feature data; establishing a vibration curve prediction model based on the feature data, and obtaining a predicted vibration curve through the vibration curve prediction model and the first vibration data; calculating an optimal weight parameter based on a hybrid rice algorithm and obtaining a fan fault prediction model; and inputting the predicted vibration curve into a fan fault prediction model to obtain a fault type, and determining a corresponding maintenance suggestion. The vibration curve prediction model and the fan fault prediction model are established through the historical vibration data and the hybrid rice algorithm, the predicted vibration curve is obtained through the first vibration data and the vibration curve prediction model, and the predicted vibration curve is input to the fan fault prediction model to obtain the fault type and the corresponding maintenance suggestion. Therefore, the fault reason of the dust removal fan can be judged more accurately.

Description

Technical Field

[0001] The present application relates to the technical field of fault detection, and in particular to a method and device for predicting dust removal fan faults based on a hybrid rice algorithm. Background Art

[0002] By analyzing historical data, it can be found that when the vibration curve of the dust removal fan changes significantly, the dust removal fan will have various possible fault conditions after running for a period of time. At the same time, the failure of the dust removal fan will not only cause serious safety hazards to the dust removal fan itself and the staff operating the dust removal fan, but may also lead to a series of problems such as production suspension. Therefore, by modeling the historical data collected by the sensor, the vibration curve of the dust removal fan in the future time period can be predicted, which can help the staff to determine the cause of the dust removal fan failure in time according to the predicted vibration curve, so as to take maintenance measures in advance to prevent safety accidents.

[0003] However, the existing technology only judges the cause of the dust removal fan failure based on basic models such as SVM and CNN and the predicted vibration curve, and does not optimize the model parameters, resulting in a low accuracy rate in judging the cause of the dust removal fan failure. Summary of the invention

[0004] In order to solve the problem of low accuracy in judging the cause of the dust removal fan failure mentioned above, the embodiment of the present application provides a dust removal fan failure prediction method and device based on a hybrid rice algorithm, and its technical solution is as follows:

[0005] In a first aspect, an embodiment of the present application provides a dust removal fan fault prediction method based on a hybrid rice algorithm, comprising:

[0006] Perform feature extraction on the acquired historical vibration data to obtain feature data;

[0007] Establishing a vibration curve prediction model based on the characteristic data, and obtaining a predicted vibration curve through the vibration curve prediction model and the first vibration data corresponding to the first time period; wherein the output data of the vibration curve prediction model is the predicted vibration curve;

[0008] The optimal weight parameters are calculated based on the hybrid rice algorithm, and the wind turbine fault prediction model is obtained through the optimal weight parameters and historical vibration data; wherein the input data of the wind turbine fault prediction model includes the predicted vibration curve;

[0009] The predicted vibration curve is input into the fan fault prediction model to obtain the fault type, and the corresponding maintenance suggestions are determined according to the fault type.

[0010] In an optional solution of the first aspect, feature extraction is performed on the acquired historical vibration data to obtain feature data, including:

[0011] Determine the maximum amplitude and the minimum amplitude based on historical vibration data;

[0012] Calculate the peak-to-peak amplitude based on the maximum amplitude and the minimum amplitude;

[0013] Obtain the highest frequency value in the historical vibration data, and determine the sampling frequency value according to the highest frequency value; wherein the sampling frequency value is greater than or equal to twice the highest frequency value;

[0014] Acquire amplitude sampling points according to the sampling frequency value and the preset sampling duration, and calculate the amplitude effective value according to all amplitude sampling points and the preset first formula;

[0015] The amplitude data is obtained according to the maximum amplitude, the minimum amplitude, the peak-to-peak amplitude and the effective amplitude;

[0016] The amplitude data is used as feature data.

[0017] In another optional solution of the first aspect, feature extraction is performed on the acquired historical vibration data to obtain feature data, further comprising:

[0018] The historical vibration data is transformed and processed by a preset algorithm to obtain frequency domain data;

[0019] Obtaining a main frequency value according to the frequency domain data and a preset first function;

[0020] Identify the harmonic frequency according to the frequency domain data and a preset second function; wherein the value of the harmonic frequency is an integer multiple of the main frequency value;

[0021] According to the main frequency value and the recognition result, the frequency data is obtained;

[0022] The frequency data is used as the feature data.

[0023] In another optional solution of the first aspect, obtaining a predicted vibration curve by using a vibration curve prediction model and first vibration data corresponding to a first time period includes:

[0024] Inputting first vibration data corresponding to a first time period into a vibration curve prediction model to obtain second vibration data corresponding to a second time period; wherein the first time period is before the second time period;

[0025] A predicted vibration curve is drawn according to the second vibration data and a preset third function.

[0026] In another optional solution of the first aspect, the optimal weight parameters are calculated based on the hybrid rice algorithm, and the wind turbine fault prediction model is obtained by using the optimal weight parameters and historical vibration data, including:

[0027] Generate an initial gene sequence corresponding to each set of initial weight parameters; wherein each set of initial weight parameters includes at least two initial weight parameters, each initial gene sequence includes at least two initial genes corresponding to the initial weight parameters, and the number of initial weight parameters is the same as the number of initial genes;

[0028] Generate at least two initial models corresponding to each set of initial weight parameters through a preset algorithm, and evaluate the classification accuracy of each initial model;

[0029] The fitness function value of each initial gene sequence is determined according to all classification correct rates, and the rice seeds corresponding to each initial gene sequence are grouped according to the fitness function value to obtain a maintenance group, a recovery group and a sterile group;

[0030] The optimal weight parameter is determined by hybridizing the rice seeds in the maintenance group with the rice seeds in the sterile group and selfing the rice seeds in the recovery group;

[0031] According to the initial model, optimal weight parameters and historical vibration data, a fan fault prediction model is obtained.

[0032] In another alternative of the first aspect, the optimal weight parameters are determined by hybridizing the rice seeds in the maintenance group with the rice seeds in the sterile group, and selfing the rice seeds in the recovery group, comprising:

[0033] Performing a hybridization operation on the rice seeds in the maintenance group and the rice seeds in the sterile group according to a preset second formula to obtain a first gene sequence;

[0034] Performing a self-pollination operation on rice seeds in the recovery group according to a preset third formula to obtain a second gene sequence;

[0035] When it is detected that the first fitness function value corresponding to the first gene sequence is greater than or equal to the second fitness function value corresponding to the second gene sequence, the first gene sequence is used as the optimal gene sequence;

[0036] When it is detected that the first fitness function value is less than the second fitness function value, the second gene sequence is used as the optimal gene sequence;

[0037] The optimal weight parameters are determined based on the optimal gene sequence.

[0038] In another optional solution of the first aspect, the predicted vibration curve is input into the fan fault prediction model to obtain the fault type, and a corresponding maintenance suggestion is determined according to the fault type, including:

[0039] Input the predicted vibration curve and the corresponding environmental data into the fan fault prediction model to obtain the fault type;

[0040] When it is identified that a fault type exists in the preset database, a maintenance suggestion corresponding to the fault type is determined in the preset database; wherein the preset database includes at least one fault type and a maintenance suggestion corresponding to each fault type;

[0041] When it is identified that the fault type does not exist in the preset database, the fault type and the first vibration data are sent to the administrator, so that the administrator can respond to the dust removal fan fault corresponding to the fault type in a timely manner.

[0042] In a second aspect, an embodiment of the present application provides a dust removal fan fault prediction device based on a hybrid rice algorithm, comprising:

[0043] A first processing module is used to extract features from the acquired historical vibration data to obtain feature data;

[0044] A second processing module is used to establish a vibration curve prediction model based on the characteristic data, and obtain a predicted vibration curve through the vibration curve prediction model and the first vibration data corresponding to the first time period; wherein the output data of the vibration curve prediction model is the predicted vibration curve;

[0045] The third processing module is used to calculate the optimal weight parameter based on the hybrid rice algorithm, and obtain the fan fault prediction model through the optimal weight parameter and the historical vibration data; wherein the input data of the fan fault prediction model includes the predicted vibration curve;

[0046] The fourth processing module is used to input the predicted vibration curve into the fan fault prediction model to obtain the fault type, and determine the corresponding maintenance suggestion according to the fault type.

[0047] In a third aspect, an embodiment of the present application further provides a dust removal fan fault prediction device based on a hybrid rice algorithm, comprising a processor and a memory;

[0048] The processor is connected to the memory;

[0049] A memory for storing executable program codes;

[0050] The processor runs the program corresponding to the executable program code by reading the executable program code stored in the memory, so as to implement the dust removal fan fault prediction method based on the hybrid rice algorithm provided by the first aspect of the embodiment of the present application or any implementation method of the first aspect.

[0051] In a fourth aspect, an embodiment of the present application provides a computer storage medium, which stores a computer program. The computer program includes program instructions. When the program instructions are executed by a processor, the dust removal fan fault prediction method based on the hybrid rice algorithm provided by the first aspect of the embodiment of the present application or any implementation method of the first aspect can be implemented.

[0052] The beneficial effects brought by the technical solutions provided by some embodiments of this specification include at least:

[0053] In the dust removal fan fault prediction process based on the hybrid rice algorithm, a vibration curve prediction model is established through the collected historical vibration data, and the fan fault prediction model is obtained by optimizing the parameters through the hybrid rice algorithm. Then, the collected first vibration data is input into the vibration curve prediction model to obtain the predicted vibration curve, and the predicted vibration curve is input into the fan fault prediction model to obtain the fault type, thereby determining the corresponding maintenance suggestions, and achieving more accurate judgment of the cause of the dust removal fan failure. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments will be 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 paying creative work.

[0055] Figure 1 An overall flow chart of a dust removal fan fault prediction method based on a hybrid rice algorithm provided in an embodiment of the present application;

[0056] Figure 2 A schematic diagram of the structure of a dust removal fan fault prediction device based on a hybrid rice algorithm provided in an embodiment of the present application;

[0057] Figure 3 A schematic structural diagram of another dust removal fan fault prediction device based on a hybrid rice algorithm provided in an embodiment of the present application. DETAILED DESCRIPTION

[0058] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0059] In the following introduction, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance. The following introduction provides multiple embodiments of the present application, and different embodiments can be replaced or combined, so the present application can also be considered to include all possible combinations of the same and / or different embodiments recorded. Therefore, if one embodiment includes features A, B, and C, and another embodiment includes features B and D, then the present application should also be considered to include embodiments containing one or more of all other possible combinations of A, B, C, and D, although the embodiment may not be clearly recorded in the following text.

[0060] The following description provides examples and does not limit the scope, applicability or examples set forth in the claims. Changes may be made to the functions and arrangements of the elements described without departing from the scope of the present application. Various processes or components may be appropriately omitted, substituted or added to each example. For example, the described method may be performed in an order different from the order described, and various steps may be added, omitted or combined. In addition, features described in some examples may be combined in other examples.

[0061] See also Figure 1 , Figure 1 An overall flow chart of a dust removal fan fault prediction method based on a hybrid rice algorithm provided in an embodiment of the present application is shown.

[0062] like Figure 1 As shown, the dust removal fan fault prediction method based on the hybrid rice algorithm may at least include the following steps:

[0063] Step 101: extract features from the acquired historical vibration data to obtain feature data.

[0064] In the embodiment of the present application, the dust removal fan fault prediction method based on the hybrid rice algorithm can be but is not limited to application on terminal devices, and the terminal devices can be but are not limited to desktop computers, laptop computers and tablet computers. In the dust removal fan fault prediction process based on the hybrid rice algorithm, a vibration curve prediction model is established through the collected historical vibration data, and the fan fault prediction model is obtained by optimizing the parameters through the hybrid rice algorithm, and then the collected first vibration data is input into the vibration curve prediction model to obtain the predicted vibration curve, and the predicted vibration curve is input into the fan fault prediction model to obtain the fault type, thereby determining the corresponding maintenance suggestions, and achieving a more accurate judgment of the cause of the dust removal fan failure.

[0065] Specifically, in the process of predicting dust removal fan faults based on the hybrid rice algorithm, the historical vibration data collected by the vibration sensor can be first obtained, and then the historical vibration data can be feature extracted to obtain feature data. Among them, the feature extraction method can be but is not limited to fast Fourier transform, autoencoder, convolutional neural network, recurrent neural network, wavelet transform, Hilbert-Huang transform and other methods.

[0066] It is understandable that historical vibration data can include but is not limited to vibration frequency, vibration amplitude, etc. collected by vibration sensors at fixed time intervals; characteristic data can include but is not limited to the mean vibration amplitude, peak vibration amplitude, frequency peak, power spectrum density, etc. that can be obtained based on historical vibration data.

[0067] As an option of the embodiment of the present application, feature extraction is performed on the acquired historical vibration data to obtain feature data, including:

[0068] Determine the maximum amplitude and the minimum amplitude based on historical vibration data;

[0069] Calculate the peak-to-peak amplitude based on the maximum amplitude and the minimum amplitude;

[0070] Obtain the highest frequency value in the historical vibration data, and determine the sampling frequency value according to the highest frequency value; wherein the sampling frequency value is greater than or equal to twice the highest frequency value;

[0071] Acquire amplitude sampling points according to the sampling frequency value and the preset sampling duration, and calculate the amplitude effective value according to all amplitude sampling points and the preset first formula;

[0072] The amplitude data is obtained according to the maximum amplitude, the minimum amplitude, the peak-to-peak amplitude and the effective amplitude;

[0073] The amplitude data is used as feature data.

[0074] Specifically, in the process of extracting features from historical vibration data to obtain feature data, the feature data may be but is not limited to amplitude data, wherein the amplitude data may be but is not limited to including a maximum amplitude value, a minimum amplitude value, a peak-to-peak amplitude value, and an effective amplitude value.

[0075] It can be understood that the maximum amplitude and the minimum amplitude can be obtained from the historical vibration data respectively through the max() function and the min function in the NumPy library of the Python language, and then the absolute value of the difference between the maximum amplitude and the minimum amplitude can be calculated, that is, the peak-to-peak amplitude, which represents the maximum amplitude variation range of the vibration signal within a complete cycle.

[0076] Furthermore, according to the Nyquist-Shannon sampling theorem, in order to accurately reconstruct the original signal, the sampling frequency must be at least twice the highest frequency component of the signal. For example, if the highest frequency in the vibration signal is 1kHz, the sampling frequency should be at least 2kHz. Therefore, the highest frequency value in the historical vibration data can be obtained and the minimum value of the sampling frequency value can be calculated based on the highest frequency value, that is, twice the highest frequency value. Then, the specific value of the sampling frequency value can be set in combination with the minimum value and other factors.

[0077] Thereafter, the number of sampling points N can be calculated according to the set sampling frequency value f and the preset sampling duration T, that is, N=fT. For example, if the sampling duration is 10 seconds and the sampling frequency is 2kHz, the total number of sampling points is N=fT=2000×10=20000 sampling points.

[0078] After the number of sampling points is obtained, sampling can be performed according to the required number of sampling points according to the sampling frequency value to obtain amplitude sampling points, and then the amplitude effective value can be calculated by the following preset first formula:

[0079]

[0080] Where RMS is the effective value of the amplitude, in millimeters; N is the number of sampling points, without unit; x i is the value of the i-th sampling point, in millimeters.

[0081] It can be understood that the effective value of the amplitude is the result of taking the square root of the average value of the square of the signal, which represents the "average energy" of the signal. Suppose we have a series of vibration displacement data points x1, x2, x3, ..., x n (x i represents the displacement value of the ith data point), in millimeters (mm). These data points represent the vibration displacement within a certain period of time. If there are 5 data points, x1 = 0.2mm, x2 = 0.5mm, x3 = -0.3mm, x4 = 0.1mm, and x5 = -0.4mm, then to calculate the effective value of the amplitude of these 5 data points, you need to follow the following 3 steps:

[0082] 1. Calculate the square of each data point: x1 2 =(0.2) 2 =0.04mm 2 , x2 2 =(0.5) 2 =0.25mm 2 , x3 2 =(-0.3) 2 =0.09mm 2 , x4 2 =(0.1)2 =0.01mm 2 , x5 2 =(-0.4) 2 =0.16mm 2 ;

[0083] 2. Calculate the average of all square values: Average square value = (x1 2 +x2 2 +x3 2 +x4 2 +x5 2 ) / n=(0.04+0.25+0.09+0.01+0.16) / 5=0.11mm 2 ;

[0084] 3. Calculate the square root: effective value of amplitude = arithmetic square root of average square value ≈ 0.33mm.

[0085] Therefore, the effective value of the amplitude of the vibration data composed of these five data points is approximately 0.33 mm. It is worth noting that in actual applications, the data set may contain more data points, and can be calculated using, but not limited to, computer software such as MATLAB or LabVIEW.

[0086] Thereafter, the maximum amplitude value, the minimum amplitude value, the peak-to-peak amplitude value, and the effective amplitude value may be combined to obtain amplitude data, and the amplitude data may be used as feature data.

[0087] As another option of the embodiment of the present application, feature extraction is performed on the acquired historical vibration data to obtain feature data, which further includes:

[0088] The historical vibration data is transformed and processed by a preset algorithm to obtain frequency domain data;

[0089] Obtaining a main frequency value according to the frequency domain data and a preset first function;

[0090] Identify the harmonic frequency according to the frequency domain data and a preset second function; wherein the value of the harmonic frequency is an integer multiple of the main frequency value;

[0091] According to the main frequency value and the recognition result, the frequency data is obtained;

[0092] The frequency data is used as the feature data.

[0093] Specifically, in the process of extracting features from historical vibration data to obtain feature data, the feature data may also include frequency data, wherein the frequency data may include, but is not limited to, identification results of the main frequency value and the harmonic frequency.

[0094] It can be understood that the historical vibration data can be transformed by a preset algorithm (such as Fourier transform, fast Fourier transform, etc.), and then the frequency domain data corresponding to the historical vibration data can be obtained, so that the index corresponding to the maximum amplitude can be found in the frequency domain data through the preset first function, that is, the numpy.argmax() function, and the corresponding frequency value, that is, the main frequency value, can be obtained through the index.

[0095] Next, after obtaining the main frequency value, all peaks in the spectrum can be found through the scipy.signal.find_peaks() function, and then the preset second function, that is, the np.isclose() function, can be used to determine whether the value of each peak frequency is close to an integer multiple of the main frequency value, thereby identifying whether a harmonic frequency exists, where the value of the harmonic frequency is an integer multiple of the main frequency value. If a harmonic frequency exists, its frequency value or its multiple relationship with the main frequency value is recorded to form a recognition result, so that the recognition result can be combined with the main frequency value to obtain frequency data, and the frequency data is used as feature data.

[0096] Step 102: establishing a vibration curve prediction model based on the characteristic data, and obtaining a predicted vibration curve through the vibration curve prediction model and first vibration data corresponding to the first time period.

[0097] Specifically, after obtaining the feature data, the existing machine learning model (such as linear regression, support vector machine, decision tree, random forest, neural network, etc.) can be trained through the feature data to obtain a vibration curve prediction model for predicting future vibration trends, wherein the input data of the vibration curve prediction model is the vibration data within a certain time period, and the output data is the vibration curve corresponding to the vibration data in a certain time period in the future after the time period.

[0098] Furthermore, first vibration data corresponding to the first time period can be obtained and input as input data into a vibration curve prediction model, which will predict the vibration data in a future time period after the first time period, thereby drawing a predicted vibration curve corresponding to the predicted vibration data based on the predicted vibration data and a drawing function or drawing tool and outputting it, wherein the drawing function may be but is not limited to the seaborn.lineplot() function in the Seaborn library, and the drawing tool may be but is not limited to MATLAB software, Mathcad software, ProSig software, SigFit software, etc.

[0099] As another option of the embodiment of the present application, obtaining a predicted vibration curve through a vibration curve prediction model and first vibration data corresponding to a first time period includes:

[0100] Inputting first vibration data corresponding to a first time period into a vibration curve prediction model to obtain second vibration data corresponding to a second time period; wherein the first time period is before the second time period;

[0101] A predicted vibration curve is drawn according to the second vibration data and a preset third function.

[0102] Specifically, after establishing a vibration curve prediction model based on the characteristic data, the first vibration data corresponding to the first time period can be input into the vibration curve prediction model to obtain the second vibration data corresponding to the second time period calculated by the vibration curve prediction model. Then, the vibration curve prediction model can draw a predicted vibration curve corresponding to the second vibration data through a preset third function, namely, the matplotlib.pyplot.plot() function, wherein the first time period is before the second time period, and there is no overlapping part between the first time period and the second time period.

[0103] Step 103: Calculate the optimal weight parameters based on the hybrid rice algorithm, and obtain a wind turbine fault prediction model through the optimal weight parameters and historical vibration data.

[0104] Specifically, after obtaining the predicted vibration curve, the optimal weight parameters of the support vector machine model can be calculated through the hybrid rice algorithm, and then the model can be trained according to the optimal weight parameters and historical vibration data to obtain a fan fault prediction model for dust removal fan fault detection, wherein the input data of the fan fault prediction model includes the predicted vibration curve and other conditions (such as humidity, temperature, pressure value, etc.), and the output data is the fault type (such as misalignment fault, loose fault, bearing fault, etc.).

[0105] As another option of the embodiment of the present application, the optimal weight parameters are calculated based on the hybrid rice algorithm, and the wind turbine fault prediction model is obtained through the optimal weight parameters and historical vibration data, including:

[0106] Generate an initial gene sequence corresponding to each set of initial weight parameters; wherein each set of initial weight parameters includes at least two initial weight parameters, each initial gene sequence includes at least two initial genes corresponding to the initial weight parameters, and the number of initial weight parameters is the same as the number of initial genes;

[0107] Generate at least two initial models corresponding to each set of initial weight parameters through a preset algorithm, and evaluate the classification accuracy of each initial model;

[0108] The fitness function value of each initial gene sequence is determined according to all classification correct rates, and the rice seeds corresponding to each initial gene sequence are grouped according to the fitness function value to obtain a maintenance group, a recovery group and a sterile group;

[0109] The optimal weight parameter is determined by hybridizing the rice seeds in the maintenance group with the rice seeds in the sterile group and selfing the rice seeds in the recovery group;

[0110] According to the initial model, optimal weight parameters and historical vibration data, a fan fault prediction model is obtained.

[0111] Specifically, in the process of determining the fan fault prediction model through the hybrid rice algorithm, a basic model (such as a support vector machine model, also known as an SVM model) can be selected first, and the key parameters of the model can be regarded as "genes". Then, an initial "gene sequence" can be constructed. In this sequence, each "gene" represents a specific parameter value (such as gene 1 represents the regularization parameter C in the SVM model, and gene 2 represents the σ parameter of the Gaussian kernel function in the SVM model).

[0112] Next, a set of random parameter values ​​corresponding to the model can be generated and regarded as a "chromosome" or "gene sequence". The random generation method can be but is not limited to the following three steps:

[0113] 1. Determine the parameter range: set a reasonable value range for each parameter, such as a set of positive integers;

[0114] 2. Random initialization: Generate a random value for each parameter;

[0115] 3. Encoding parameters: Encode each parameter value into a "gene". For example, binary encoding is used for parameter value 6, and the gene is 110.

[0116] 4. Combined gene sequences: connect all genes into a complete sequence, such as connecting gene 0110 and gene 0010 into 01100010.

[0117] It can be understood that the parameter value here is the value of the initial weight parameter, each set of initial weight parameters includes at least two initial weight parameters, each initial gene sequence includes at least two initial genes corresponding to the initial weight parameters, and the number of initial weight parameters is the same as the number of initial genes.

[0118] It is worth noting that the support vector machine (SVM) includes the concepts of both algorithm and model. SVM is both an algorithm for building a model and a specific model type. Therefore, after generating the initial gene sequence, the convolutional neural network and the SVM algorithm can be used to train at least two SVM models, wherein each SVM model includes a set of initial weight parameters corresponding to the initial gene sequence, and each set of initial weight parameters is different from each other, thereby obtaining at least two corresponding initial models, wherein the input data of each initial model includes the predicted vibration curve and environmental factors, and the output data are all fault types.

[0119] Afterwards, the classification accuracy of each initial model can be evaluated, and the evaluation method can be but not limited to the k-fold cross validation method and the leave-one-out cross validation method. After obtaining the classification accuracy of each initial model, the classification accuracy can be used as the fitness function value of the initial gene sequence corresponding to each initial model, and then the fitness function values ​​can be arranged in order from large to small, and the rice seeds corresponding to each initial gene sequence can be grouped into a maintenance group, a recovery group, and a sterile group in turn. Among them, the grouping method can be but not limited to grouping according to a preset ratio, equal division, etc. For example, if there are 30 rice seeds, they can be divided into 6 maintenance groups, 9 recovery groups, and 15 sterile groups according to a preset ratio of 2:3:5; they can also be equally divided into 10 maintenance groups, 10 recovery groups, and 10 sterile groups. Among them, the fitness function value of the rice seeds in the maintenance group is greater than the fitness function value of the rice seeds in the recovery group, and the fitness function value of the rice seeds in the recovery group is greater than the fitness function value of the rice seeds in the sterile group.

[0120] It should be noted that the fitness function value is a quantitative indicator used to evaluate the quality of the solution in the optimization algorithm. In optimization algorithms such as genetic algorithms and particle swarm optimization algorithms, the fitness function is used to guide the search process to find the optimal solution or a solution close to the optimal solution. In the hybrid rice algorithm, each initial gene sequence corresponds to a rice seed, that is, a set of solutions. In this solution, the rice seed refers to the initial weight parameter, and the optimal rice seed refers to the optimal weight parameter, that is, the optimal solution.

[0121] Next, the formula for hybridization operation can be constructed by, but not limited to, single-point crossover, multi-point crossover, uniform crossover, arithmetic crossover, ordered crossover, mixed crossover and the like. At the same time, the formula for self-pollination operation can be constructed by, but not limited to, direct replication, gene mutation, gene recombination, partial replication, adaptive self-pollination and the like. Then, according to the two constructed formulas, the hybridization operation of rice seeds in the maintenance group and the rice seeds in the sterile group, as well as the self-pollination operation of rice seeds in the recovery group can be respectively realized.

[0122] After completing the hybridization operation and the self-pollination operation, the results of the two operations can be compared, and an optimal rice seed can be determined based on the comparison result, and then the optimal weight parameter can be obtained, so that an initial model with the optimal weight parameter can be obtained, and the model can be trained through historical vibration data to generate a fan fault prediction model for determining the fault type of the dust removal fan, wherein the optimal weight parameter includes at least two parameter values.

[0123] As another option of the embodiment of the present application, the optimal weight parameter is determined by hybridizing the rice seeds in the maintenance group with the rice seeds in the sterile group, and selfing the rice seeds in the recovery group, including:

[0124] Performing a hybridization operation on the rice seeds in the maintenance group and the rice seeds in the sterile group according to a preset second formula to obtain a first gene sequence;

[0125] Performing a self-pollination operation on rice seeds in the recovery group according to a preset third formula to obtain a second gene sequence;

[0126] When it is detected that the first fitness function value corresponding to the first gene sequence is greater than or equal to the second fitness function value corresponding to the second gene sequence, the first gene sequence is used as the optimal gene sequence;

[0127] When it is detected that the first fitness function value is less than the second fitness function value, the second gene sequence is used as the optimal gene sequence;

[0128] The optimal weight parameters are determined based on the optimal gene sequence.

[0129] Specifically, in the process of implementing the hybridization operation and the selfing operation on the rice seeds, the rice seeds in the maintenance group and the rice seeds in the sterile group can be hybridized by the following preset second formula to obtain the first gene sequence:

[0130]

[0131] Among them, x i,d t+1 represents the dth gene in the gene sequence of the i-th rice seed in the next generation of sterile line; x i,d t represents the dth gene in the gene sequence of the i-th rice seed in the contemporary sterile line; x k1,d t represents the dth gene in the gene sequence of the k1th rice seed selected from the maintenance line; x k2,d t represents the dth gene in the gene sequence of the k2th rice seed selected from the sterile line; α is a random number in the range of [-1,1], which is used to control the degree of hybridization.

[0132] During the hybridization operation, if the fitness function value of the gene sequence of the contemporary rice seeds is greater than that of the previous generation, the contemporary rice seeds will be used as the first rice seeds, otherwise the previous generation rice seeds will be used as the first rice seeds, until the maximum number of iterations is reached, and a first rice seed is obtained, and then the first gene sequence corresponding to the first rice seed can be determined, wherein the maximum number of iterations is a preset value, which can usually be determined based on the specific number of rice seeds involved in the hybridization operation.

[0133] Then, the rice seeds in the recovery group can be self-pollinated using the following preset third formula to obtain the second gene sequence:

[0134]

[0135] Among them, x i t+1 represents the new individual produced by self-pollination of the i-th restorer line; x i t represents the i-th individual in the current restoration line; x k t represents the kth individual randomly selected from the recovery line; r is a uniform random number in the range of [0,1].

[0136] During the selfing operation, if the fitness function value of the gene sequence of the selfing offspring is greater than that of the original individual, the original individual is replaced by the selfing offspring, and the number of selfings is reset to 0; if the fitness function value of the gene sequence of the selfing offspring is less than or equal to the original individual, the original individual is retained, and the number of selfings is increased by 1. If the original individual has not been replaced when the number of selfings reaches the maximum allowed number, a reset operation is performed according to the following formula, so that when the selfing operation fails to improve the fitness function value of the individual, a new individual is generated through the reset operation to avoid falling into a local optimal solution, where the maximum allowed number is a preset value, which can usually be determined according to the specific number of rice seeds involved in the selfing operation:

[0137]

[0138] In the above formula, x i d’ Rice seeds representing the restorer line x i d New rice seeds generated by the reset operation; x max is the upper limit of gene quantification; x min is the lower limit of gene quantization; s is a uniform random number in the range of [0,1].

[0139] It is understandable that in genetic algorithms or similar optimization algorithms, gene quantization refers to the process of converting the expression form of genes into numerical form so that the algorithm can operate. The upper and lower limits of gene quantization refer to the maximum and minimum values ​​of gene quantization, which are mainly used to limit the effective range of gene values. For example, if the gene is binary coded, its upper limit can be but is not limited to 2. n -1, and its lower limit can be but not limited to 0, where n represents the number of bits of the binary code of the gene (such as 4 bits for 0110). In practical applications, the upper and lower limits of the quantized genes can be set according to the specific problem. Once the upper and lower limits are determined, they can be used in the genetic algorithm to control the gene values ​​of individuals (i.e., solutions). When new individuals are generated, it can be ensured that the gene values ​​of these individuals are within this range to ensure the feasibility of the solution.

[0140] It should be noted that when the number of self-pollination reaches the maximum allowed number and the original individual has been replaced, the current individual is used as the second rice seed; when the number of self-pollination reaches the maximum allowed number and the original individual has not been replaced, the original individual is reset and a new individual is generated, and then the new individual can be used as the original individual for the next round of self-pollination operation, and the value of the maximum allowed number is adjusted to determine the second rice seed. Thereafter, the second gene sequence corresponding to the second rice seed can be obtained according to the second rice seed.

[0141] After obtaining the first gene sequence and the second gene sequence, when it is detected that the first fitness function value corresponding to the first gene sequence is greater than or equal to the second fitness function value corresponding to the second gene sequence, the first gene sequence is used as the optimal gene sequence; when it is detected that the first fitness function value is less than the second fitness function value, the second gene sequence is used as the optimal gene sequence. Subsequently, the optimal gene sequence can be disassembled to obtain at least two genes, and then the weight parameters corresponding to each gene can be determined and combined into the optimal weight parameters. It is worth noting that the optimal weight parameters include at least two parameter values.

[0142] Step 104: input the predicted vibration curve into the fan fault prediction model to obtain the fault type, and determine the corresponding maintenance suggestion according to the fault type.

[0143] Specifically, after obtaining the predicted vibration curve and the fan fault prediction model, the predicted vibration curve and other conditions (such as humidity, temperature, pressure values, etc.) can be input into the fan fault prediction model, and then the fan fault prediction model can be used to predict the possible fault types of the dust removal fan (such as misalignment fault, looseness fault, bearing fault, etc.). In this way, the maintenance suggestions corresponding to the fault types can be determined based on historical experience, expert knowledge, equipment debugging records, etc., and the equipment management personnel can be notified in time to repair the faulty parts, thereby improving the accuracy of the fault type judgment of the dust removal fan, so that maintenance personnel can locate the problem more quickly and take corresponding maintenance measures.

[0144] As another option of the embodiment of the present application, the predicted vibration curve is input into the fan fault prediction model to obtain the fault type, and the corresponding maintenance suggestion is determined according to the fault type, including:

[0145] Input the predicted vibration curve and the corresponding environmental data into the fan fault prediction model to obtain the fault type;

[0146] When it is identified that a fault type exists in the preset database, a maintenance suggestion corresponding to the fault type is determined in the preset database; wherein the preset database includes at least one fault type and a maintenance suggestion corresponding to each fault type;

[0147] When it is identified that the fault type does not exist in the preset database, the fault type and the first vibration data are sent to the administrator, so that the administrator can respond to the dust removal fan fault corresponding to the fault type in a timely manner.

[0148] Specifically, after obtaining the predicted vibration curve and the fan fault prediction model, the environmental data within the first time period can be obtained first, such as the temperature data within the first time period (intake temperature, exhaust temperature, motor temperature, ambient temperature, etc.), the humidity data within the first time period (relative humidity, absolute humidity, etc.), the air pressure data within the first time period (atmospheric pressure, intake pressure, exhaust pressure, etc.), the flow data within the first time period (intake flow, exhaust flow, etc.), the noise data within the first time period (noise level, noise frequency, etc.), the dust concentration within the first time period (intake dust concentration, exhaust dust concentration, etc.), the light intensity within the first time period, and the electromagnetic interference within the first time period.

[0149] After obtaining the environmental data in the first time period, the predicted vibration curve corresponding to the first time period and the environmental data can be input into the fan fault prediction model, which is analyzed and processed by the fan fault prediction model to predict the type of fault that may occur in the dust removal fan in the future time period, and then the preset database can be searched to see whether the fault type exists, wherein the preset database includes at least one fault type and maintenance suggestions corresponding to each fault type. If the fault type exists in the preset database, the maintenance suggestion corresponding to the fault type can be obtained, and the maintenance suggestion can be set by the management personnel based on historical experience or expert knowledge; if the fault type does not exist in the preset database, the fault type and the first vibration data corresponding to the first time period can be sent to the management personnel, so that the management personnel can respond to the dust removal fan fault corresponding to the fault type in a timely manner, such as communicating with other relevant staff to solve the problem, consulting the equipment manual, user guide and technical documents, using online forums, technical communities and support websites to seek solutions to similar problems, contacting the manufacturer's technical support or hiring external experts for consultation, etc.

[0150] It can be understood that when the frequency spectrum of the predicted vibration curve mainly shows that the amplitude of the first harmonic frequency accounts for a large amplitude component, it can be preliminarily determined that the fan is unbalanced. At this time, dynamic balancing correction is required, that is, the weight is increased or decreased on both sides of the blade to eliminate the unbalanced torque, thereby reducing the vibration level; when multiple frequencies appear in the frequency spectrum of the predicted vibration curve, especially when the amplitudes of the first harmonic frequency, the second harmonic frequency and the third harmonic frequency are large, it can be preliminarily determined that it is a misalignment fault. At this time, it is necessary to check the alignment of the coupling and make adjustments to ensure the concentricity of the two shafts; when multiple frequencies appear in the frequency spectrum of the predicted vibration curve, and the amplitudes from the first harmonic frequency to the tenth harmonic frequency are large, it can be preliminarily determined that the fault is caused by the loose structure of the machine foot. At this time, it is necessary to check the tightness of the machine foot and foundation, and reinforce the loose parts, and reinstall them if necessary; when the amplitude appears at the non-harmonic frequency in the frequency spectrum of the predicted vibration curve, it can be preliminarily determined that it is a bearing fault. At this time, the bearing needs to be replaced, and before replacement, it is also necessary to check whether the bearing seat is damaged and make necessary repairs.

[0151] See also Figure 2 , Figure 2 A structural schematic diagram of a dust removal fan fault prediction device based on a hybrid rice algorithm provided in an embodiment of the present application is shown.

[0152] like Figure 2 As shown, the dust removal fan fault prediction device based on the hybrid rice algorithm may include at least a first processing module 201, a second processing module 202, a third processing module 203 and a fourth processing module 204, wherein:

[0153] The first processing module 201 is used to extract features from the acquired historical vibration data to obtain feature data;

[0154] The second processing module 202 is used to establish a vibration curve prediction model based on the characteristic data, and obtain a predicted vibration curve through the vibration curve prediction model and the first vibration data corresponding to the first time period; wherein the output data of the vibration curve prediction model is the predicted vibration curve;

[0155] The third processing module 203 is used to calculate the optimal weight parameter based on the hybrid rice algorithm, and obtain the wind turbine failure prediction model through the optimal weight parameter and the historical vibration data; wherein the input data of the wind turbine failure prediction model includes the predicted vibration curve;

[0156] The fourth processing module 204 is used to input the predicted vibration curve into the fan fault prediction model to obtain the fault type, and determine the corresponding maintenance suggestion according to the fault type.

[0157] In some possible embodiments, feature extraction is performed on the acquired historical vibration data to obtain feature data, including:

[0158] The first processing module 201 is specifically used for:

[0159] Determine the maximum amplitude and the minimum amplitude based on historical vibration data;

[0160] Calculate the peak-to-peak amplitude based on the maximum amplitude and the minimum amplitude;

[0161] Obtain the highest frequency value in the historical vibration data, and determine the sampling frequency value according to the highest frequency value; wherein the sampling frequency value is greater than or equal to twice the highest frequency value;

[0162] Acquire amplitude sampling points according to the sampling frequency value and the preset sampling duration, and calculate the amplitude effective value according to all amplitude sampling points and the preset first formula;

[0163] The amplitude data is obtained according to the maximum amplitude, the minimum amplitude, the peak-to-peak amplitude and the effective amplitude;

[0164] The amplitude data is used as feature data.

[0165] In some possible embodiments, performing feature extraction on the acquired historical vibration data to obtain feature data further includes:

[0166] The first processing module 201 is specifically used for:

[0167] The historical vibration data is transformed and processed by a preset algorithm to obtain frequency domain data;

[0168] Obtaining a main frequency value according to the frequency domain data and a preset first function;

[0169] Identify the harmonic frequency according to the frequency domain data and a preset second function; wherein the value of the harmonic frequency is an integer multiple of the main frequency value;

[0170] According to the main frequency value and the recognition result, the frequency data is obtained;

[0171] The frequency data is used as the feature data.

[0172] In some possible embodiments, obtaining a predicted vibration curve by using a vibration curve prediction model and first vibration data corresponding to a first time period includes:

[0173] The second processing module 202 is specifically used for:

[0174] Inputting first vibration data corresponding to a first time period into a vibration curve prediction model to obtain second vibration data corresponding to a second time period; wherein the first time period is before the second time period;

[0175] A predicted vibration curve is drawn according to the second vibration data and a preset third function.

[0176] In some possible embodiments, the optimal weight parameters are calculated based on the hybrid rice algorithm, and a wind turbine fault prediction model is obtained through the optimal weight parameters and historical vibration data, including:

[0177] The third processing module 203 is specifically used for:

[0178] Generate an initial gene sequence corresponding to each set of initial weight parameters; wherein each set of initial weight parameters includes at least two initial weight parameters, each initial gene sequence includes at least two initial genes corresponding to the initial weight parameters, and the number of initial weight parameters is the same as the number of initial genes;

[0179] Generate at least two initial models corresponding to each set of initial weight parameters through a preset algorithm, and evaluate the classification accuracy of each initial model;

[0180] The fitness function value of each initial gene sequence is determined according to all classification correct rates, and the rice seeds corresponding to each initial gene sequence are grouped according to the fitness function value to obtain a maintenance group, a recovery group and a sterile group;

[0181] The optimal weight parameter is determined by hybridizing the rice seeds in the maintenance group with the rice seeds in the sterile group and selfing the rice seeds in the recovery group;

[0182] According to the initial model, optimal weight parameters and historical vibration data, a fan fault prediction model is obtained.

[0183] In some possible embodiments, the optimal weight parameters are determined by hybridizing the rice seeds in the maintenance group with the rice seeds in the sterile group, and selfing the rice seeds in the recovery group, including:

[0184] The third processing module 203 is specifically used for:

[0185] Performing a hybridization operation on the rice seeds in the maintenance group and the rice seeds in the sterile group according to a preset second formula to obtain a first gene sequence;

[0186] Performing a self-pollination operation on rice seeds in the recovery group according to a preset third formula to obtain a second gene sequence;

[0187] When it is detected that the first fitness function value corresponding to the first gene sequence is greater than or equal to the second fitness function value corresponding to the second gene sequence, the first gene sequence is used as the optimal gene sequence;

[0188] When it is detected that the first fitness function value is less than the second fitness function value, the second gene sequence is used as the optimal gene sequence;

[0189] The optimal weight parameters are determined based on the optimal gene sequence.

[0190] In some possible embodiments, the predicted vibration curve is input into a fan fault prediction model to obtain a fault type, and a corresponding maintenance suggestion is determined according to the fault type, including:

[0191] The fourth processing module 204 is specifically used for:

[0192] Input the predicted vibration curve and the corresponding environmental data into the fan fault prediction model to obtain the fault type;

[0193] When it is identified that a fault type exists in the preset database, a maintenance suggestion corresponding to the fault type is determined in the preset database; wherein the preset database includes at least one fault type and a maintenance suggestion corresponding to each fault type;

[0194] When it is identified that the fault type does not exist in the preset database, the fault type and the first vibration data are sent to the administrator, so that the administrator can respond to the dust removal fan fault corresponding to the fault type in a timely manner.

[0195] See also Figure 3 , Figure 3 A structural schematic diagram of another dust removal fan fault prediction device based on a hybrid rice algorithm provided in an embodiment of the present application is shown.

[0196] like Figure 3As shown, the dust removal fan fault prediction device 300 based on the hybrid rice algorithm may include at least one processor 301 , at least one network interface 304 , a user interface 303 , a memory 305 and at least one communication bus 302 .

[0197] The communication bus 302 may be used to realize the connection and communication among the above-mentioned components.

[0198] The user interface 303 may include buttons, and the optional user interface may also include a standard wired interface or a wireless interface.

[0199] The network interface 304 may include, but is not limited to, a Bluetooth module, an NFC module, a Wi-Fi module, etc.

[0200] Among them, the processor 301 may include one or more processing cores. The processor 301 uses various interfaces and lines to connect the various parts of the dust removal fan fault prediction device 300 based on the hybrid rice algorithm, and executes various functions and processes data of the dust removal fan fault prediction device 300 based on the hybrid rice algorithm by running or executing instructions, programs, code sets or instruction sets stored in the memory 305, and calling data stored in the memory 305. Optionally, the processor 301 can be implemented in at least one hardware form of DSP, FPGA, and PLA. The processor 301 can integrate one or a combination of CPU, GPU, modem, etc. Among them, the CPU mainly processes the operating system, user interface, and application program; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; the modem is used to process wireless communication. It can be understood that the above-mentioned modem may not be integrated into the processor 301, and it can be implemented by a single chip.

[0201] The memory 305 may include a RAM or a ROM. Optionally, the memory 305 includes a non-transitory computer-readable medium. The memory 305 may be used to store instructions, programs, codes, code sets or instruction sets. The memory 305 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store data involved in the above-mentioned various method embodiments, etc. The memory 305 may optionally be at least one storage device located away from the aforementioned processor 301. As Figure 3 As shown, the memory 305 as a computer storage medium may include an operating system, a network communication module, a user interface module, and a dust removal fan fault prediction application based on a hybrid rice algorithm.

[0202] Specifically, the processor 301 can be used to call the dust removal fan fault prediction application based on the hybrid rice algorithm stored in the memory 305, and specifically perform the following operations:

[0203] Perform feature extraction on the acquired historical vibration data to obtain feature data;

[0204] Establishing a vibration curve prediction model based on the characteristic data, and obtaining a predicted vibration curve through the vibration curve prediction model and the first vibration data corresponding to the first time period; wherein the output data of the vibration curve prediction model is the predicted vibration curve;

[0205] The optimal weight parameters are calculated based on the hybrid rice algorithm, and the wind turbine fault prediction model is obtained through the optimal weight parameters and historical vibration data; wherein the input data of the wind turbine fault prediction model includes the predicted vibration curve;

[0206] The predicted vibration curve is input into the fan fault prediction model to obtain the fault type, and the corresponding maintenance suggestions are determined according to the fault type.

[0207] The present application also provides a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, the steps of the above method are implemented. The computer-readable storage medium may include, but is not limited to, any type of disk, including a floppy disk, an optical disk, a DVD, a CD-ROM, a micro drive, and a magneto-optical disk, a ROM, a RAM, an EPROM, an EEPROM, a DRAM, a VRAM, a flash memory device, a magnetic card or an optical card, a nanosystem (including a molecular memory IC), or any type of medium or device suitable for storing instructions and / or data.

[0208] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0209] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0210] In the several embodiments provided in the present application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0211] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0212] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0213] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a memory and includes several instructions for a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of each embodiment method of the present application. The aforementioned memory includes: U disk, read-only memory (ROM), random access memory (RAM), mobile hard disk, disk or optical disk, etc., and other media that can store program codes.

[0214] A person skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by entering a program to instruct related hardware. The program can be stored in a computer-readable memory, and the memory can include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0215] The above are only exemplary embodiments of the present disclosure and cannot be used to limit the scope of the present disclosure. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the specification and practicing the disclosure here, those skilled in the art will easily think of the implementation scheme of the present disclosure. This application is intended to cover any modification, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the technical field that are not recorded in the present disclosure. The description and examples are only regarded as exemplary, and the scope and spirit of the present disclosure are defined by the claims.

Claims

1. A dust removal fan fault prediction method based on hybrid rice algorithm, characterized in that: include: Perform feature extraction on the acquired historical vibration data to obtain feature data; A vibration curve prediction model is established based on the characteristic data, and a predicted vibration curve is obtained through the vibration curve prediction model and the first vibration data corresponding to the first time period; wherein the output data of the vibration curve prediction model is the predicted vibration curve; Calculating the optimal weight parameter based on the hybrid rice algorithm, and obtaining a wind turbine failure prediction model through the optimal weight parameter and the historical vibration data; wherein the input data of the wind turbine failure prediction model includes the predicted vibration curve; The predicted vibration curve is input into the fan fault prediction model to obtain the fault type, and a corresponding maintenance suggestion is determined according to the fault type.

2. The method according to claim 1, characterized in that The feature extraction of the acquired historical vibration data to obtain feature data includes: Determine a maximum amplitude and a minimum amplitude according to the historical vibration data; Calculate the peak-to-peak value of the amplitude according to the maximum amplitude and the minimum amplitude; Obtaining the highest frequency value in the historical vibration data, and determining a sampling frequency value according to the highest frequency value; wherein the sampling frequency value is greater than or equal to twice the highest frequency value; Acquire amplitude sampling points according to the sampling frequency value and the preset sampling duration, and calculate the amplitude effective value according to all the amplitude sampling points and the preset first formula; Obtaining amplitude data according to the maximum amplitude value, the minimum amplitude value, the peak-to-peak amplitude value and the effective amplitude value; The amplitude data is used as feature data.

3. The method according to claim 1, characterized in that The feature extraction of the acquired historical vibration data to obtain feature data also includes: The historical vibration data is transformed and processed by a preset algorithm to obtain frequency domain data; Obtaining a main frequency value according to the frequency domain data and a preset first function; Identify the harmonic frequency according to the frequency domain data and a preset second function; wherein the value of the harmonic frequency is an integer multiple of the main frequency value; Obtaining frequency data according to the main frequency value and the identification result; The frequency data is used as feature data.

4. The method according to claim 1, characterized in that: The step of obtaining a predicted vibration curve by using the vibration curve prediction model and first vibration data corresponding to a first time period includes: Inputting first vibration data corresponding to a first time period into the vibration curve prediction model to obtain second vibration data corresponding to a second time period; wherein the first time period is before the second time period; A predicted vibration curve is drawn according to the second vibration data and a preset third function.

5. The method according to claim 1, characterized in that The hybrid rice algorithm is used to calculate the optimal weight parameter, and a wind turbine fault prediction model is obtained through the optimal weight parameter and the historical vibration data, including: Generate an initial gene sequence corresponding to each set of initial weight parameters; wherein each set of initial weight parameters includes at least two initial weight parameters, each of the initial gene sequences includes at least two initial genes corresponding to the initial weight parameters, and the number of the initial weight parameters is the same as the number of the initial genes; Generate at least two initial models corresponding to each group of the initial weight parameters by a preset algorithm, and evaluate the classification accuracy of each of the initial models; Determining the fitness function value of each of the initial gene sequences according to all the classification correct rates, and grouping the rice seeds corresponding to each of the initial gene sequences according to the fitness function value to obtain a maintenance group, a recovery group and a sterile group; Determining the optimal weight parameter by hybridizing the rice seeds in the maintenance group and the rice seeds in the sterile group, and selfing the rice seeds in the recovery group; A fan fault prediction model is obtained according to the initial model, the optimal weight parameters and the historical vibration data.

6. The method according to claim 5, characterized in that The method of determining the optimal weight parameter by hybridizing the rice seeds in the maintenance group with the rice seeds in the sterile group and selfing the rice seeds in the recovery group comprises: Performing a hybridization operation on the rice seeds in the maintenance group and the rice seeds in the sterile group according to a preset second formula to obtain a first gene sequence; Performing a selfing operation on the rice seeds in the recovery group according to a preset third formula to obtain a second gene sequence; When it is detected that the first fitness function value corresponding to the first gene sequence is greater than or equal to the second fitness function value corresponding to the second gene sequence, taking the first gene sequence as the optimal gene sequence; When it is detected that the first fitness function value is less than the second fitness function value, taking the second gene sequence as the optimal gene sequence; An optimal weight parameter is determined according to the optimal gene sequence.

7. The method according to claim 1, characterized in that The step of inputting the predicted vibration curve into the fan fault prediction model to obtain a fault type, and determining a corresponding maintenance suggestion according to the fault type, comprises: Inputting the predicted vibration curve and corresponding environmental data into the fan fault prediction model to obtain the fault type; When it is identified that the fault type exists in a preset database, a maintenance suggestion corresponding to the fault type is determined in the preset database; wherein the preset database includes at least one of the fault types and a maintenance suggestion corresponding to each of the fault types; When it is identified that the fault type does not exist in the preset database, the fault type and the first vibration data are sent to an administrator, so that the administrator can respond to the dust removal fan fault corresponding to the fault type in a timely manner.

8. A dust removal fan fault prediction device based on hybrid rice algorithm, characterized in that: include: A first processing module is used to extract features from the acquired historical vibration data to obtain feature data; a second processing module, configured to establish a vibration curve prediction model based on the characteristic data, and obtain a predicted vibration curve through the vibration curve prediction model and the first vibration data corresponding to the first time period; wherein the output data of the vibration curve prediction model is the predicted vibration curve; A third processing module is used to calculate an optimal weight parameter based on a hybrid rice algorithm, and obtain a fan fault prediction model through the optimal weight parameter and the historical vibration data; wherein the input data of the fan fault prediction model includes the predicted vibration curve; The fourth processing module is used to input the predicted vibration curve into the fan fault prediction model to obtain the fault type, and determine corresponding maintenance suggestions according to the fault type.

9. A dust removal fan fault prediction device based on hybrid rice algorithm, characterized in that: including a processor and a memory; The processor is connected to the memory; The memory is used to store executable program code; The processor runs a program corresponding to the executable program code by reading the executable program code stored in the memory, so as to execute the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed on a computer or a processor, the computer or the processor executes the steps of the method according to any one of claims 1 to 7.

Images