CN112183290B - Mechanical fault diagnosis system based on SAsFFT algorithm - Google Patents
Mechanical fault diagnosis system based on SAsFFT algorithm Download PDFInfo
- Publication number
- CN112183290B CN112183290B CN202011002546.2A CN202011002546A CN112183290B CN 112183290 B CN112183290 B CN 112183290B CN 202011002546 A CN202011002546 A CN 202011002546A CN 112183290 B CN112183290 B CN 112183290B
- Authority
- CN
- China
- Prior art keywords
- data
- module
- user
- submodule
- fault diagnosis
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2136—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on sparsity criteria, e.g. with an overcomplete basis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/30—Authentication, i.e. establishing the identity or authorisation of security principals
- G06F21/31—User authentication
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
- G06F2218/04—Denoising
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computer Security & Cryptography (AREA)
- Data Mining & Analysis (AREA)
- Software Systems (AREA)
- Computer Hardware Design (AREA)
- Signal Processing (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention discloses a mechanical fault diagnosis system based on an SAsFFT algorithm, which comprises a user management module, a data preprocessing module, a fault diagnosis module, a data storage module and a display module, wherein the user management module is used for managing the mechanical fault; the data preprocessing module receives original vibration data input from the outside and preprocesses the vibration data; and the fault diagnosis module analyzes and processes the preprocessed vibration data to obtain frequency spectrum data, and performs classification diagnosis on the frequency spectrum data to obtain a fault diagnosis result. The system improves the signal-to-noise ratio by preprocessing the input original vibration signal, restores the actual condition as truly as possible, judges the running state of the machine through the fault diagnosis module, distinguishes the normal running state and the fault state, identifies the fault type, and provides accurate and reliable data support for the maintainers to know the running state of the machine.
Description
Technical Field
The invention relates to the technical field of cloud computing and fault diagnosis, in particular to a mechanical fault diagnosis system based on an SAsFFT algorithm.
Background
At present, with the continuous promotion of industrialization, industrial safety problems are more and more concerned, and due to the fact that mechanical failure monitoring is not in place, casualties and property loss caused by mechanical failure happen frequently. In the operation process of a machine, vibration data are generated every moment, thousands of records can be generated every second, the most important link of fault diagnosis of the machine is frequency domain analysis on the vibration data, the traditional machine fault diagnosis method is difficult to obtain accurate analysis results after analyzing a large amount of vibration data in a short time, reliable data support is difficult to provide for the diagnosis results, the authenticity and the reliability of the obtained diagnosis results are difficult to guarantee, and the reference value is not high.
Therefore, how to provide a stable and reliable mechanical failure diagnosis system is a problem that needs to be solved by those skilled in the art.
Disclosure of Invention
In view of this, the invention provides a mechanical fault diagnosis system based on the SAsFFT algorithm, which implements efficient and reliable diagnosis of mechanical faults by preprocessing, analyzing and classifying and diagnosing vibration data, and solves the problem that the diagnosis result of the existing mechanical fault diagnosis method is not true and accurate enough.
In order to achieve the purpose, the invention adopts the following technical scheme:
a mechanical fault diagnosis system based on the SAsFFT algorithm, the system comprising:
the user management module is used for registering and logging in the system by a user and managing user authority;
the data preprocessing module reads original vibration data input from the outside and preprocesses the vibration data;
the fault diagnosis module analyzes and processes the vibration data preprocessed by the data preprocessing module to obtain frequency spectrum data, and performs classification diagnosis on the frequency spectrum data to obtain a fault diagnosis result;
the data storage module is used for storing user registration login data, original vibration data, preprocessed vibration data and fault diagnosis results in a classified manner; and
and the display module is used for displaying the fault diagnosis result in real time.
Further, the user management module includes:
the registration submodule is used for the user to fill in registration information for registration verification;
the login submodule is used for inputting login information and verification information for a user to log in;
the registration information maintenance submodule is used for auditing and verifying the registration information uploaded by the user and feeding back a verification result; and
and the authority maintenance submodule is used for managing the authority state of the successfully logged user and recording a log of the logged user.
The user management module of the invention provides the functions of login authentication, authority management, information modification and the like of the system, is responsible for the login authentication and exit management of users with different roles, judges the legality of the login and exit according to the different roles of the users, records the login and exit logs of the users and determines the authority of the users.
Further, the data preprocessing module comprises:
the data acquisition submodule is used for reading original vibration data from an external system;
the data slicing submodule is used for carrying out data slicing on the original vibration data;
the data sorting submodule is used for carrying out weighting operation on the fragmented data and sorting the data after the weighting operation; and
and the smoothing sub-module is used for smoothing the sorted data.
In the working process of mechanical equipment, due to the influence of natural frequency response and electromagnetic wave interference of a sensor, some noise signals are doped in vibration data signals obtained through a data acquisition system, the noise signals comprise periodic interference signals and also comprise a part of irregular random interference signals, for example, high-frequency signals higher than sampling frequency occur, so that the acquired vibration signals need to be preprocessed to improve the signal-to-noise ratio and restore the signal-to-noise ratio to the actual situation as real as possible.
Further, the fault diagnosis module includes:
the SAsFFT algorithm processing submodule is used for carrying out SAsFFT algorithm processing on the preprocessed vibration data to obtain frequency spectrum data representing time domain signal characteristics in the vibration data; and
and the classification submodule is used for performing classification diagnosis on various kinds of frequency spectrum information contained in the frequency spectrum data to obtain a fault diagnosis result.
Further, the SAsFFT algorithm processing sub-module includes:
the data supplementing unit is used for performing data supplementing processing on the input preprocessed vibration data;
the sparsity estimation unit is used for estimating the sparsity of the spectrum sparse signal in the vibration data after the data completion processing to obtain the prediction sparsity;
the frequency spectrum rearrangement unit is used for carrying out time domain rearrangement on the time domain signals in the vibration data after the data padding processing;
the window function filtering unit is used for adding a window function to the time domain signal after the time domain rearrangement to obtain a smooth time domain signal;
a down-sampling unit, configured to perform down-sampling fast fourier transform processing on the smoothed time-domain signal;
the positioning unit is used for performing positioning cyclic operation on a frequency point with the maximum signal sparsity in the time domain signal after the down-sampling fast Fourier transform processing; and
and the estimation unit is used for estimating the frequency domain amplitude of the time domain signal subjected to the down-sampling fast Fourier transform processing according to the positioning cycle operation result to obtain accurate frequency spectrum data.
Further, the data storage module includes: the system comprises a user data storage submodule, an original vibration data storage submodule, an intermediate data storage submodule and a diagnosis result storage submodule;
the user data storage submodule is used for storing user registration login data obtained by the user management module, the original vibration data storage submodule is used for storing original vibration data, the intermediate data storage submodule is used for storing preprocessed vibration data obtained by the data preprocessing module, and the diagnosis result storage submodule is used for storing diagnosis result data obtained by the fault diagnosis module.
The data storage module is divided into three parts, wherein the first part is used for storing user registration login information, the second part is used for storing the most original time domain signal, the third part is used for storing data and fault results after preprocessing and SAsFFT algorithm processing, HDFS is used for storing the vibration signals due to the fact that the data volume of the vibration signals is large, the data and the diagnosis result data volume after SAsFFT algorithm processing are relatively small, a relational database is used for storing the vibration signals, and the results are convenient to display.
According to the technical scheme, compared with the prior art, the mechanical fault diagnosis system based on the SAsFFT algorithm is disclosed and provided, the system improves the signal-to-noise ratio by preprocessing the input original vibration signal, restores the actual condition as real as possible, judges the operation state of the machine through the fault diagnosis module, distinguishes the normal operation state and the fault state, identifies the fault type, and provides accurate and reliable data support for maintenance personnel to know the operation condition of the machine.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.
Fig. 1 is a schematic structural diagram of a mechanical fault diagnosis system based on an SAsFFT algorithm provided in the present invention;
FIG. 2 is a schematic diagram of a flow chart of an implementation of the SAsFFT algorithm in the embodiment of the present invention;
FIG. 3 is a schematic diagram of an implementation flow of an sFFT algorithm in an embodiment of the present invention;
fig. 4 is a flowchart illustrating a sparsity prediction process according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1, the embodiment of the invention discloses a mechanical fault diagnosis system based on a SAsFFT algorithm, which comprises:
the user management module 1 is used for a user to register and log in the system and manage user authority;
the data preprocessing module 2 is used for receiving original vibration data input from the outside and preprocessing the vibration data;
the fault diagnosis module 3 is used for analyzing and processing the vibration data preprocessed by the data preprocessing module to obtain frequency spectrum data, and performing classification diagnosis on the frequency spectrum data to obtain a fault diagnosis result;
the data storage module 4 is used for storing user registration login data, original vibration data, preprocessed vibration data and fault diagnosis results in a classified manner; and
and the display module 5 is used for displaying the fault diagnosis result in real time.
Specifically, the user management module 1 includes:
the registration submodule 11 is used for the user to fill in registration information for registration verification;
the login submodule 12 is used for inputting login information and verification information for a user to login;
the registration information maintenance submodule 13 is used for auditing and verifying the registration information uploaded by the user and feeding back a verification result; and
and the authority maintenance submodule 14 is used for managing the authority state of the successfully logged user and recording a log of the logged user.
When a user registers, information such as a user name, an encrypted password, authority level and the like of the user is stored in a database, in order to ensure the safety of a registration process, the user needs to perform verification twice during registration, whether the user information meets the specification or not and whether the passwords input twice are the same or not are respectively verified, the embodiment verifies the user name and the passwords input by the user through a regular expression, whether the passwords input twice are the same or not is verified in a character string matching mode, and if the passwords input twice are matched, the verification is successful.
In this example, the specific process of user registration is:
step 1: the user inputs a user name, a password and a confirmation password;
step 2: verifying whether the information input by the user meets the specification;
and step 3: if the user information meets the specification, further verifying whether the confirmed password is the same as the set password, if so, successfully registering, and if not, feeding back error information.
The specific process of user login is as follows:
step 1: a user submits a user name, a password and an authentication code;
step 2: verifying the verification code;
and step 3: if the verification code is successfully verified, acquiring user data information for comparison, otherwise, feeding back error information;
and 4, step 4: and if the database comparison is successful, acquiring the user authority and recording a user login log.
Specifically, the data preprocessing module 2 specifically includes:
a data acquisition sub-module 21, configured to read original vibration data from an external system;
the data slicing submodule 22 is used for carrying out data slicing on the original vibration data;
the data sorting submodule 23 is configured to perform weighting operation on the sliced data and sort the weighted data; and
and a smoothing sub-module 24, configured to perform smoothing on the sorted data.
The vibration data preprocessing module firstly reads original vibration data from a distributed file system (HDFS), processes the original vibration data by utilizing the principle of a five-time three-point smoothing method, and stores an obtained processing result, wherein the specific process comprises the following steps:
the original vibration data is subjected to data slicing, the slicing principle is that 5 data are selected at equal intervals according to the length N of the original vibration data, and are stored in key1, value1, list 2, list 3, value3, list 4, list 5 and key value pair, and each key value pair is used as an element in the preprocessing set. A cubic curve can be fitted by 5 data points, the value of the corresponding position on the cubic curve is taken as the result after data smoothing processing, the coefficient of a cubic equation can be solved by a least square method, and a cubic smoothing formula can be obtained by a simultaneous equation, wherein the formula is as follows:
according to the above formula, a weighting coefficient of each value can be obtained, and a value after smoothing processing, that is, a value after smoothing processing can be obtained by calculation
And finally, sequencing the elements in the new preprocessing set through a sortBYKey function according to the key value in each element.
Specifically, the fault diagnosis module 3 includes:
the SAsFFT algorithm processing submodule 31 is configured to perform SAsFFT algorithm processing on the preprocessed vibration data to obtain frequency spectrum data representing time domain signal characteristics in the vibration data; and
and the classification submodule 32 is used for performing classification diagnosis on various kinds of spectrum information contained in the spectrum data to obtain a fault diagnosis result.
Specifically, the SAsFFT algorithm processing sub-module 31 includes:
the data supplementing unit is used for performing data supplementing processing on the input preprocessed vibration data;
the sparsity estimation unit is used for estimating the sparsity of the spectrum sparse signal in the vibration data after the data completion processing to obtain the prediction sparsity;
the frequency spectrum rearrangement unit is used for carrying out time domain rearrangement on the time domain signals in the vibration data after the data padding processing;
the window function filtering unit is used for adding a window function to the time domain signal after the time domain rearrangement to obtain a smooth time domain signal;
the down-sampling unit is used for carrying out down-sampling fast Fourier transform processing on the smooth time domain signal;
the positioning unit is used for performing positioning cyclic operation on a frequency point with the maximum signal sparsity in the time domain signal after the down-sampling fast Fourier transform processing; and
and the estimation unit is used for estimating the frequency domain amplitude of the time domain signal subjected to the down-sampling fast Fourier transform processing according to the positioning cycle operation result to obtain accurate frequency spectrum data.
The SAsFFT (sparse fast Fourier transform with self-adaptive sparsity) algorithm utilizes the frequency domain sparsity of signals, and obtains the sparsity of the signals through iterative gradual estimation on the premise that the sparsity of the signals is unknown, so that the algorithm is suitable for the signals with unknown sparsity. Compared with the existing sFFT algorithm, the SAsFFT algorithm is improved by adding the sparsity estimation and detection process. The algorithm can be divided into 2 steps, and the overall algorithm flow is shown in fig. 2.
1. And (5) estimating sparsity. Over-estimating the sparsity of the signal to a certain extent, i.e. predicting the sparsity
Slightly larger than the true sparsity k of the signal. Wherein the sparsity is predicted
Is to estimate the k-sparse signalThe resulting sparsity.
2. Sparse fast fourier transform. Will predict sparsity
The signals being obtained by performing sparse fast Fourier transform on the signals as known parameters
Information of the Fourier domain.
The main idea of the sparse fast fourier transform (sFFT) algorithm is to map signal frequency points into a group of "baskets" through a certain strategy. Because the spectrum of the signal is sparse, each effective frequency point of the signal will exist uniquely in each "basket" as long as the strategy is proper. By superimposing the frequency points in each "basket", the N-point sequence will be compressed into a short sequence, a process referred to as time domain dimensionality reduction. And after FFT, obtaining frequency domain information of the signal, and reconstructing the frequency spectrum by combining a time domain dimension reduction strategy to recover the frequency spectrum of the original signal of the N point. The sparse Fourier transform algorithm flow chart is shown in FIG. 3, wherein X (n) is input, and is output after spectral rearrangement, window function filtering, down-sampling FFT, positioning and estimation processing.
The fin-based SAsFFT (sparse adaptive for sparse Fast Fourier Transform) algorithm mainly comprises the following parts:
1. data completion parallel algorithm
The SAsFFT operation is performed on the premise that the length of the input time domain sequence must be an integer power of 2, and therefore, it is necessary to perform data padding processing so that the length of the input sequence satisfies the integer power of 2 with respect to an input sequence that does not satisfy this requirement.
The data completion implementation idea based on DataSet is as follows: the data files stored in the HDFS are first read by lines and distributed in a DataSet. Obtaining the number of elements by a count operator, taking the logarithm of the number of elements, obtaining the minimum integer (namely power m) which is larger than or equal to the value of the number of elements by a ceil function, and calculating 2 m Obtaining the sequence length satisfying the FFT algorithm, and obtaining the difference between the two lengthsThe number of d to be zero-filled. And constructing a DataSet of all zeros with the length d, and using an union operator to combine the DataSet with the initial DataSet data set and generate a new DataSet data set.
2. Signal sparsity estimation algorithm
And estimating the sparsity of the spectrum sparse signal by utilizing the sparse characteristic of the sparse signal and the property of time domain dimension reduction in the sFFT algorithm, thereby obtaining the prediction sparsity. The flowchart for predicting sparsity is shown in fig. 4.
In the signal sparsity determination algorithm, the final prediction sparsity is determined through multiple iterations, and the specific process is to preset a series of bucket numbers K b· Initializing a smaller prediction sparsity, then performing the dimensionality reduction operation of a signal time domain (namely frequency spectrum rearrangement, window function filtering and signal downsampling) by using the sparsity k', then performing FFT (fast Fourier transform) of a downsampling domain, and finally counting the effective components of a downsampling domain vector by an energy detection method to obtain the prediction sparsity of the mth iteration. If the number of effective fractions is not satisfactory, i.e.
The step is iterated until the requirements are met. At this time
Namely the final prediction sparsity of the signal, and is used for an sFFT algorithm.
3. Parallel algorithm for frequency spectrum rearrangement
In the time domain dimension reduction operation in this embodiment, frequency domain signals need to be rearranged in frequency spectrum. From the shift and scale properties of the signal, a frequency domain rearrangement of the signal corresponds to a time domain rearrangement of the signal. Therefore, the time domain signal needs to be rearranged. In this embodiment, the time domain signals are rearranged by using a modulo inverse operation.
The time domain rearrangement realization idea based on the DataSet data set is as follows: and taking the time domain signal obtained after the data are filled as input, and adding an index to the time domain signal as a key value of the corresponding data. And performing a modulo operation on the key to obtain a new key value as a new index value and form a new DataSet data set. And finally, sequencing the new DataSet data sets according to the new index value to obtain the time domain signal after time domain rearrangement.
4. Window function filtering parallel algorithm
The input time domain signal is a finite-length signal, and in order to avoid frequency spectrum leakage caused by window effect, a window function needs to be added to the rearranged time domain signal, so that the time domain signal is smoother at the cut-off position. Therefore, a hamming window is selected as a window function, and the weighting coefficients thereof can make the side lobe smaller.
And window function filtering based on the DataSet data set only needs to operate the time domain signal value after frequency spectrum rearrangement through a map () function to obtain the time domain signal after windowing. Denoted by winDataSet. The DataSet transition is as follows:
val winDataSet＝rearrangeDataSet.map(hanning)
the hanning is a self-defined Hanning window function, the function of the hanning is to window an input time domain signal, and the realization idea of the hanning is to multiply the rearranged key values by h, wherein
h＝0.5·[1-cos(2·pi·k/(N-1)]
Where N is the signal length and k is the rearranged bond.
5. Down-sampling FFT parallel algorithm
(1) Indexing operation
In order to make the output of the FTT algorithm a natural sequence, the input sequence needs to be indexed, and the output is made a natural sequence by changing the order of the input sequence.
Firstly, converting the key of an input sequence into a binary system through a map (DecToBinary) function, then constructing a DataSet of an Array [ Char ] type, realizing code bit inversion through the map (reindex) function, and finally converting the binary system into a decimal system through the map (BinaryToDec) function.
The realization idea of the DecToBinary function is as follows: dividing the decimal integer by 2 integer to obtain a quotient and a remainder; and removing the quotient by using 2 to obtain a quotient and a remainder, and carrying out the operation until the quotient is less than 1, then using the remainder obtained firstly as the low-order significant bit of the binary number, and using the remainder obtained later as the high-order significant bit of the binary number, and sequentially arranging the remainder.
The realization idea of the reindex function is as follows: two shaping variables i, j are defined, where i is the number of bits of the converted binary number and j =1. Starting from j =1, the ith and jth bits of the binary number are exchanged in sequence until i > j.
The realization idea of the BinaryToDec function is as follows: binary numbers are first written as a weighted coefficient expansion and then summed according to a decimal addition rule.
(2) Butterfly operation
The butterfly operation in the FFT parallel algorithm based on the Flink adopts the recursive idea, the parity grouping is carried out on the input sequence, and the recursive butterfly operation is carried out on the grouped sequence. The algorithm is designed as follows:
(1) complex type Complex design
In the FFT calculation process, complex numbers are involved in the calculation, so a class capable of representing complex numbers needs to be designed. The Complex class is designed to represent the Complex number, and the serialization and deserialization can be quickly and conveniently carried out. The design idea of the complete class is: two variables r and theta are defined, where r is the modulus of the complex number and theta is the value of the arc between the line segment from the origin (0, 0) to the (imaginary, real) point and the positive x-axis direction.
(2) Recursive computation process
Because the sequence length for performing the FFT satisfies N =2 m The butterfly requires recursive m layers. In each layer of butterfly operation, a complex array is required to be constructed to store complex type data blocks. Firstly, whether the input sequence length N meets the power of 2 or not needs to be judged, if yes, butterfly operation is carried out, and if not, the operation is quitted. Then, after the sequence parity grouping is input, the butterfly operation is continuously iterated in the range of N/2 until m times of operation is finished.
6. Positioning
Obtaining B frequency points by down-sampling FTT, selecting K (signal sparsity) larger frequency points, and solving Hash mapping h for the frequency points σ (k) The set I is obtained by inverse mapping of = round (sigma kB/N), and the set I is used as the result of the frequency point positioning at this time, namely, the frequency point positioning is finished onceAnd (5) positioning operation.
In order to ensure the accuracy of the positioning result and reduce the influence of collision on the result in the basket dividing operation, four steps of frequency spectrum rearrangement, window function filtering, down-sampling FFT and positioning operation are circularly executed to obtain a plurality of sets I. And taking the K frequency points with the largest occurrence frequency in the plurality of sets as a final positioning result.
7. Valuation
Because the window function filtering changes the amplitude of the original time domain signal, a certain error is brought to the result of the down-sampling FFT. In order to make the final FFT operation more accurate, the frequency domain amplitude of the final signal needs to be estimated.
K frequency points obtained through multiple positioning cycles have the largest occurrence frequency, and K sets are constructed to store multiple amplitudes of the corresponding frequency points. And then, the median of each set is obtained as the final amplitude of the frequency point.
The spectrum data obtained after the SAsFFT algorithm processing comprises multi-class spectrum information, and taking a wind turbine generator as an example, the spectrum data comprises a large number of normal bearing spectrums, single-point drive end bearing defects and fan end bearing defects, and the spectrum data is used as a classification sample.
In the embodiment, the libsvm classification software is adopted to classify the running state of the wind turbine generator, samples need to be divided into three classes, and a one-to-one multi-classification mode is adopted, namely, each two samples construct a libsvm classifier, so that 3 sub-classifiers need to be constructed.
In this embodiment, the format of the training sample is (label 1 index1: value1 index2: value2 \8230; label2 index1: value1 index2: value2 \8230; where label is a mark, index is a feature, and value is a feature value.
Specifically, the data storage module 4 includes: a user data storage submodule 41, an original vibration data storage submodule 42, an intermediate data storage submodule 43, and a diagnosis result storage submodule 44;
the user data storage submodule 41 is configured to store user registration login data obtained by the user management module 1, the original vibration data storage submodule 42 is configured to store original vibration data, the intermediate data storage submodule 43 is configured to store preprocessed vibration data obtained by the data preprocessing module 2, and the diagnosis result storage submodule 44 is configured to store diagnosis result data obtained by the fault diagnosis module 3.
In this embodiment, the data storage module mainly stores user basic information, mechanical failure information and system operation status information, and in order to ensure that the classification of the data storage process is clearer and is not prone to error, this embodiment provides the content of the relational database, and for the system information stored in the relational database, in combination with the relationship of each entity, the designed database relational model is as follows:
(1) User information table: user ID, user name, password, authority level; the tabular format is shown in table 1 below:
TABLE 1 user information Table
| Serial number | Name (R) | Type (B) | Brief description of the drawings |
| 1 | User_id | Varchar(20) | User ID (Main Key) |
| 2 | User_name | Varchar(20) | User name |
| 3 | password | Varchar(20) | |
| 4 | User_right | integer | Permission level |
(2) Device information table: the method comprises the following steps of equipment number, unit number, equipment name, original vibration data address link, preprocessed vibration data address link, SAsFFT algorithm operation result address link, diagnosis time, lowest frequency, highest frequency, running state and fault type; the tabular format is shown in table 2 below;
TABLE 2 Equipment information Table
(3) System information table: a primary key ID, an event name, event detailed information, and an event occurrence time; the format of the table can be seen in table 3 below:
TABLE 3 System information Table
| Serial number | Name (R) | Types of | Brief description of the drawings |
| 1 | event_id | Varchar(20) | Event ID (Main Key) |
| 2 | event_name | Varchar(20) | Event name |
| 3 | information | TEXT | |
| 4 | time | Datatime | Time of occurrence |
(4) User-mechanical events table: user ID, event name, time of occurrence, success or failure; the format of this table can be seen in table 4 below:
table 4 user-device event table
| Serial number | Name (R) | Type (B) | Brief description of the drawings |
| 1 | User_id | Varchar(20) | User ID (Main Key) |
| 2 | machine_id | Varchar(20) | Equipment ID (Main Key) |
| 3 | event_name | Varchar(20) | |
| 4 | issuccess | Varchar(20) | Success or failure |
| 5 | time | Datatime | Time of occurrence |
The system disclosed by the embodiment is mainly aimed at a user who is a mechanical maintenance worker, the wind turbine generator is taken as an object, the main purpose is to enable the mechanical maintenance worker to obtain the relevant operation information of the wind turbine generator with a fault in the first time, and the maintenance worker can conveniently confirm the fault in time and maintain the fault, so that the maintenance worker can download the relevant information of the wind turbine generator with the fault on a front-end page, wherein the relevant information comprises the name of the wind turbine generator, the use condition, the fault, the maintenance history and other information, and the state information of the wind turbine generator can be modified, and comprises the information of repaired, difficult repair, misinformation and the like. The whole system has stable work, complete data and accurate and reliable diagnosis result, and is more suitable for large-scale popularization and application.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. The device disclosed in the embodiment corresponds to the method disclosed in the embodiment, so that the description is simple, and the relevant points can be referred to the description of the method part.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (4)
1. A mechanical fault diagnosis system based on an SAsFFT algorithm is characterized by comprising:
the user management module is used for registering and logging in the system by a user and managing user authority;
the data preprocessing module reads original vibration data input from the outside and preprocesses the vibration data;
the data preprocessing module comprises:
the data acquisition submodule is used for reading original vibration data from an external system;
the data slicing submodule is used for carrying out data slicing on the original vibration data; selecting 5 data at equal intervals according to the length N of the original vibration data, and storing the data in key value pairs < (key 1, value 1), lists (key 2, value 2), lists (key 3, value 3), lists (key 4, value 4) and lists (key 5, value 5) >;
the data sorting submodule is used for carrying out weighting operation on the fragmented data and sorting the data after the weighting operation; fitting a cubic curve by using 5 data points; solving the coefficient of a cubic equation by a least square method, and obtaining a cubic smoothing formula by a simultaneous equation; obtaining a value weighting coefficient in each key value pair according to a cubic smoothing formula, and obtaining a value after smoothing processing through calculation; sorting the elements in the new preprocessing set through a sortBykey function according to the key value in each element; and
the smoothing sub-module is used for smoothing the sorted data;
the fault diagnosis module analyzes and processes the vibration data preprocessed by the data preprocessing module to obtain frequency spectrum data, and performs classification diagnosis on the frequency spectrum data to obtain a fault diagnosis result; the fault diagnosis module includes:
the SAsFFT algorithm processing submodule is used for carrying out SAsFFT algorithm processing on the preprocessed vibration data to obtain frequency spectrum data representing time domain signal characteristics in the vibration data; and
the classification submodule is used for performing classification diagnosis on various kinds of frequency spectrum information contained in the frequency spectrum data to obtain a fault diagnosis result;
the data storage module is used for storing user registration login data, original vibration data, preprocessed vibration data and fault diagnosis results in a classified manner; and
and the display module is used for displaying the fault diagnosis result in real time.
2. The system of claim 1, wherein the customer management module comprises:
the registration submodule is used for the user to fill in registration information for registration verification;
the login submodule is used for inputting login information and verification information for a user to log in;
the registration information maintenance submodule is used for auditing and verifying the registration information uploaded by the user and feeding back a verification result; and
and the authority maintenance submodule is used for managing the authority state of the successfully logged user and recording a log of the logged user.
3. The system of claim 1, wherein the processing sub-module of the SAsFFT algorithm comprises:
the data supplementing unit is used for performing data supplementing processing on the input preprocessed vibration data;
the sparsity estimation unit is used for estimating the sparsity of the spectrum sparse signal in the vibration data after the data completion processing to obtain the prediction sparsity;
the frequency spectrum rearrangement unit is used for carrying out time domain rearrangement on the time domain signals in the vibration data after the data padding processing;
the window function filtering unit is used for adding a window function to the time domain signal after the time domain rearrangement to obtain a smooth time domain signal;
a down-sampling unit, configured to perform down-sampling fast fourier transform processing on the smoothed time domain signal;
the positioning unit is used for performing positioning cyclic operation on a frequency point with the maximum signal sparsity in the time domain signal after the down-sampling fast Fourier transform processing; and
and the estimation unit is used for estimating the frequency domain amplitude of the time domain signal subjected to the down-sampling fast Fourier transform processing according to the positioning cycle operation result to obtain accurate frequency spectrum data.
4. The system for diagnosing mechanical failure based on the SAsFFT algorithm of claim 1, wherein the data storage module comprises: the system comprises a user data storage submodule, an original vibration data storage submodule, an intermediate data storage submodule and a diagnosis result storage submodule;
the user data storage submodule is used for storing user registration login data obtained by the user management module, the original vibration data storage submodule is used for storing original vibration data, the intermediate data storage submodule is used for storing preprocessed vibration data obtained by the data preprocessing module, and the diagnosis result storage submodule is used for storing diagnosis result data obtained by the fault diagnosis module.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011002546.2A CN112183290B (en) | 2020-09-22 | 2020-09-22 | Mechanical fault diagnosis system based on SAsFFT algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011002546.2A CN112183290B (en) | 2020-09-22 | 2020-09-22 | Mechanical fault diagnosis system based on SAsFFT algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112183290A CN112183290A (en) | 2021-01-05 |
| CN112183290B true CN112183290B (en) | 2023-02-24 |
Family
ID=73956311
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011002546.2A Active CN112183290B (en) | 2020-09-22 | 2020-09-22 | Mechanical fault diagnosis system based on SAsFFT algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112183290B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114047010A (en) * | 2021-11-10 | 2022-02-15 | 宜昌达瑞机电科技有限公司 | Vibration and noise detection method and device for automobile air conditioner and computer equipment |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108716989A (en) * | 2018-05-17 | 2018-10-30 | 长沙理工大学 | Turbo generator set shafting torsional vibration fault monitoring method, monitor and system |
| CN109238735A (en) * | 2018-08-08 | 2019-01-18 | 上海交通大学 | The malfunction monitoring diagnostic system of the electronic AGV of port cargo |
| CN110160791A (en) * | 2019-06-27 | 2019-08-23 | 郑州轻工业学院 | Based on small echo-spectrum kurtosis induction machine bearing failure diagnosis system and diagnostic method |
| CN110849626A (en) * | 2019-11-18 | 2020-02-28 | 东南大学 | Self-adaptive sparse compression self-coding rolling bearing fault diagnosis system |
-
2020
- 2020-09-22 CN CN202011002546.2A patent/CN112183290B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108716989A (en) * | 2018-05-17 | 2018-10-30 | 长沙理工大学 | Turbo generator set shafting torsional vibration fault monitoring method, monitor and system |
| CN109238735A (en) * | 2018-08-08 | 2019-01-18 | 上海交通大学 | The malfunction monitoring diagnostic system of the electronic AGV of port cargo |
| CN110160791A (en) * | 2019-06-27 | 2019-08-23 | 郑州轻工业学院 | Based on small echo-spectrum kurtosis induction machine bearing failure diagnosis system and diagnostic method |
| CN110849626A (en) * | 2019-11-18 | 2020-02-28 | 东南大学 | Self-adaptive sparse compression self-coding rolling bearing fault diagnosis system |
Non-Patent Citations (1)
| Title |
|---|
| 稀疏信号快速傅里叶变换研究;刘仲;《中国优秀硕博士学位论文全文数据库（硕士）信息科技辑》;20180615;摘要，第2,4章 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112183290A (en) | 2021-01-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10713597B2 (en) | Systems and methods for preparing data for use by machine learning algorithms | |
| Mishra | Handling imbalanced data: SMOTE vs. random undersampling | |
| Bommidi et al. | Hybrid wind speed forecasting using ICEEMDAN and transformer model with novel loss function | |
| Yin et al. | Wasserstein Generative Adversarial Network and Convolutional Neural Network (WG‐CNN) for Bearing Fault Diagnosis | |
| CN113884290B (en) | Voltage regulator fault diagnosis method based on self-training semi-supervised generation countermeasure network | |
| CN104729853A (en) | Rolling bearing performance degradation evaluation device and method | |
| CN104504403B (en) | A kind of rotating machinery fault Forecasting Methodology based on scattering conversion | |
| Gao et al. | Rolling bearing fault diagnosis of PSO–LSSVM based on CEEMD entropy fusion | |
| CN110175541B (en) | Method for extracting sea level change nonlinear trend | |
| CN112183290B (en) | Mechanical fault diagnosis system based on SAsFFT algorithm | |
| CN114548199A (en) | Multi-sensor data fusion method based on deep migration network | |
| CN115114965A (en) | Wind turbine generator gearbox fault diagnosis model, method, equipment and storage medium | |
| CN115204368A (en) | Aircraft engine fault diagnosis method based on intelligent chip technology | |
| CN114546797A (en) | Cloud virtual machine aging prediction method based on time sequence decomposition | |
| CN112819141A (en) | Information processing system based on deep neural network | |
| CN106940688B (en) | A kind of mathematical model modeling method of EMD chaff component identification | |
| CN114386452B (en) | Nuclear power circulating water pump sun gear fault detection method | |
| Hao et al. | Gear fault detection in a planetary gearbox using deep belief network | |
| Zhang et al. | Bearing Fault Diagnosis with Kernel Sparse Representation Classification Based on Adaptive Local Iterative Filtering‐Enhanced Multiscale Entropy Features | |
| Sun et al. | A compound structure for wind speed forecasting using MKLSSVM with feature selection and parameter optimization | |
| Shuai et al. | Adaptive morphological feature extraction and support vector regressive classification for bearing fault diagnosis | |
| CN113408804A (en) | Electricity stealing behavior detection method, system, terminal equipment and storage medium | |
| An et al. | A Parameter‐Optimized Variational Mode Decomposition Investigation for Fault Feature Extraction of Rolling Element Bearings | |
| Huang et al. | An Accurate Prediction Algorithm of RUL for Bearings: Time‐Frequency Analysis Based on MRCNN | |
| CN117949821B (en) | Asynchronous motor fault diagnosis method and related equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |