Electromechanical equipment health prediction and management system based on big data

Abstract

The invention discloses an electromechanical equipment health prediction and management system based on big data, which relates to the technical field of electromechanical equipment prediction management and comprises a vibration monitoring module, a working information acquisition module, a data classification module, a standard data test module and an anomaly judgment module. The plurality of vibration monitoring modules are used for respectively monitoring the vibration condition of each part of the electromechanical equipment to obtain the vibration data of each part; according to the electromechanical equipment health prediction and management system based on big data, the vibration monitoring module, the working information acquisition module, the data classification module, the standard data test module and the abnormity judgment module are arranged, and task information and workload executed by electromechanical equipment can be monitored; whether the vibration of each part of the electromechanical equipment is abnormal under the execution of the corresponding task and the corresponding load is judged, and the abnormal part of the electromechanical equipment is judged to have health hidden dangers and need to be repaired and maintained through the abnormal prediction of the vibration of the part, so that the health judgment precision of the part of the electromechanical equipment is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of predictive management of electromechanical equipment, and particularly relates to a health prediction and management system for electromechanical equipment based on big data. Background Art

[0002] Electromechanical equipment generally refers to mechanical, electrical, and electrical automation equipment. In construction, it mostly refers to the general term for mechanical and pipeline equipment other than geotechnical engineering, carpentry, steel bars, and cement. Electromechanical equipment is composed of several components. In order to ensure the good working state of electromechanical equipment, it is necessary to monitor the working state of electromechanical equipment in order to timely discover problems that occur in electromechanical equipment.

[0003] The prior art with the publication number CN115391083A discloses an airborne electromechanical equipment health management method and system. The airborne electromechanical equipment health management method includes: Step S11, based on the airborne electromechanical equipment, obtaining the airborne data of the airborne electromechanical equipment; Step S12, based on the airborne data, the airborne processor performs a first diagnosis on the airborne electromechanical equipment to obtain first diagnosis data; Step S13, downloading the airborne data and the first diagnosis data to the ground memory; Step S14, based on the airborne data and / or the first diagnosis data, the ground processor performs a second diagnosis on the airborne electromechanical equipment to obtain second diagnosis data. In this way, the problems of timely diagnosis of the state and life prediction of airborne electromechanical equipment are solved.

[0004] However, in the actual application of electromechanical equipment, the work tasks undertaken by electromechanical equipment are different. When electromechanical equipment executes each task or process, the components participating in the work are different, the degree of participation of each component in the work is different, and the loads of electromechanical equipment and its components are different. This may lead to different judgment criteria for whether each component of electromechanical equipment is in a normal working state. The prior art still cannot judge whether the operation of electromechanical equipment is normal according to the differences in tasks or loads of electromechanical equipment, and the judgment accuracy is relatively low. Summary of the Invention

[0005] The purpose of the present invention is to provide a health prediction and management system for electromechanical equipment based on big data to solve the above deficiencies in the prior art.

[0006] To achieve the above purpose, the present invention provides the following technical solution: A health prediction and management system for electromechanical equipment based on big data, including a vibration monitoring module, a work information acquisition module, a data classification module, a standard data test module, and an abnormality judgment module;

[0007] There are multiple vibration monitoring modules, which are used to respectively monitor the vibration conditions of each part of the electromechanical equipment to obtain the vibration data of each part. Among them, the vibration sensors can be installed at the positions of each part of the electromechanical equipment to respectively monitor each part of the electromechanical equipment;

[0008] The working information acquisition module is used to acquire the task information executed by the electromechanical equipment, and the parts and loads working at each moment when each task information is executed. The task information can be set according to relevant industry rules and working experience. For example, the conventional cleaning, dehydration, self-cleaning, etc. of an automatic washing machine. Then, by acquiring the parts participating in the work when the automatic washing machine executes each task, the working parts are associated with the corresponding task information to obtain the parts working conditions at each moment of each task information;

[0009] The data classification module is used to respectively acquire the vibration data of the parts corresponding to each task information, extract the features in the time domain of the vibration data to obtain vibration time-domain data, and classify the vibration time-domain data according to different task information to obtain multiple vibration time-domain data sets;

[0010] The data classification module is also used to transform the vibration time-domain data into vibration frequency-domain data based on Fourier transform, and the vibration frequency-domain data forms multiple vibration frequency-domain data sets classified according to different task information;

[0011] The Fourier transform formula is as follows:

[0012]

[0013] Among them, X(f) represents the frequency-domain signal, x(t) represents the time-domain signal, j is the imaginary unit, f represents the frequency, and t represents the time;

[0014] The standard data testing module is used to test the vibration time-domain data and vibration frequency-domain data of the electromechanical equipment when executing each task information, and extract the real-time vibration time-domain features and real-time vibration frequency-domain features of each task information based on the real-time vibration time-domain data and vibration frequency-domain data;

[0015] The abnormality judgment module is used to acquire the vibration data and task information of the electromechanical equipment in real time, compare them with the corresponding vibration time-domain features and vibration frequency-domain features, and judge whether the electromechanical equipment is abnormal when executing the task information.

[0016] Further, the data classification module is used to respectively acquire the vibration data of the parts corresponding to each task information, extract the features in the time domain of the vibration data to obtain vibration time-domain data, and classify the vibration time-domain data according to different task information to obtain multiple vibration time-domain data sets, including the following steps:

[0017] A1. Acquire each task information and the working parts and loads corresponding to the task information;

[0018] A2. Classify the task information based on the differences of the working parts to obtain multiple task groups;

[0019] A3. Extract the vibration time-domain data under each load condition of each task group respectively;

[0020] A4. Obtain the vibration time-domain characteristics of the vibration time-domain data under each load condition of each task group;

[0021] A5. Compare the vibration time-domain characteristics, and divide the task groups and load conditions corresponding to the vibration time-domain characteristics with a similarity greater than the set time-domain characteristic threshold into a first sub-task group, that is, divide multiple first sub-task groups under each task group, so that the vibration time-domain characteristics are similar when the parts of each first sub-task group execute the task information of the corresponding task group;

[0022] A6. Divide the vibration time-domain data corresponding to the vibration time-domain characteristics in each first sub-task group into a vibration time-domain data set.

[0023] Further, the data classification module is also used to transform the vibration time-domain data into vibration frequency-domain data based on Fourier transform, and the vibration frequency-domain data forms multiple vibration frequency-domain data sets classified according to different task information, including the following steps:

[0024] B1. Respectively transform the vibration time-domain data in each vibration time-domain data set to the frequency domain based on Fourier transform to obtain the corresponding vibration frequency-domain data;

[0025] B2. Divide the vibration frequency-domain data based on the difference in the task groups of the task information corresponding to the vibration frequency-domain data;

[0026] B3. Obtain the vibration frequency-domain characteristics of the vibration frequency-domain data under each load condition of each task group;

[0027] B4. Compare the vibration frequency-domain characteristics, and divide the task groups and load conditions corresponding to the vibration frequency-domain characteristics with a similarity greater than the set frequency-domain characteristic threshold into a second sub-task group;

[0028] B5. Divide the vibration frequency-domain data corresponding to the vibration frequency-domain characteristics in each second sub-task group into a vibration frequency-domain data set.

[0029] Further, the abnormality judgment module is also used to train a machine learning model based on the vibration time-domain data set and the vibration frequency-domain data set to judge whether the device is abnormal, including the following steps:

[0030] C1. Set normal labels and abnormal labels, and mark the vibration time-domain characteristics in the vibration time-domain data set and the vibration frequency-domain characteristics in the vibration frequency-domain data set. The marking can be achieved by associating normal vibration time-domain characteristics and vibration frequency-domain characteristics with normal labels, and abnormal vibration time-domain characteristics and vibration frequency-domain characteristics with abnormal labels;

[0031] C2. Based on the vibration time-domain features, normal labels, and abnormal labels corresponding to the vibration time-domain data in the i-th vibration time-domain dataset, train the first machine learning sub-model to obtain the first abnormal prediction model i, which is used to output the predicted normal label or abnormal label based on the input real-time vibration time-domain features, where i is initially 1;

[0032] C3. Let i = i + 1, and judge whether i ≤ n holds, where n is the number of vibration time-domain datasets;

[0033] C4. If so, return to C2;

[0034] C5. If not, integrate all the first abnormal prediction models i to obtain the first abnormal prediction model.

[0035] Further, the abnormal judgment module is also used to train a machine learning model based on the vibration time-domain dataset and the vibration frequency-domain dataset to judge whether the device is abnormal, and further includes the following steps:

[0036] D1. Based on the vibration frequency-domain features, normal labels, and abnormal labels corresponding to the vibration frequency-domain data in the l-th vibration frequency-domain dataset, train the second machine learning sub-model to obtain the second abnormal prediction model l, which is used to output the predicted normal label or abnormal label based on the input real-time vibration frequency-domain features, where l is initially 1;

[0037] D2. Let l = l + 1, and judge whether l ≤ m holds, where m is the number of vibration frequency-domain datasets;

[0038] D3. If so, return to D1;

[0039] D4. If not, integrate all the second abnormal prediction models l to obtain the second abnormal prediction model;

[0040] D5. Integrate the first abnormal prediction model and the second abnormal prediction model to obtain the abnormal prediction model.

[0041] Among them, the present invention does not limit the specific machine learning model. For example, support vector machine (SVM), random forest, K-nearest neighbor (KNN), neural network, regression model, etc. can be adopted. When extracting vibration time-domain features and vibration frequency-domain features, the vibration time-domain features include: mean value: the average value of the signal; standard deviation: the degree of fluctuation of the signal; peak value: the maximum amplitude of the signal; root mean square value (RMS): a feature reflecting the signal energy; kurtosis and skewness: describing the sharpness and symmetry of the signal waveform. When extracting vibration frequency-domain features, first, through Fourier transform, the time-domain signal is converted into a frequency-domain signal to obtain spectrum information. The vibration frequency-domain features include: main frequency component: extracting the main frequencies in the spectrum; band energy: calculating the energy distribution of each frequency band, such as total energy, energy of a specific frequency band; harmonic analysis: identifying the fundamental frequency and its harmonic components. Then, through feature selection techniques (such as PCA, LDA, tree-based feature importance, etc.) to screen the features most useful for judging the device state.

[0042] During model training, the data set is divided into a training set and a test set, such as 70% training set and 30% test set. The selected machine learning model is trained using the training set. Methods such as cross-validation are used to optimize the hyperparameters of the model. Metrics such as accuracy, recall rate, F1-score, etc. are used to evaluate the performance of the model on the test set. The classification effect of the model is analyzed to check the recognition of normal and abnormal situations to ensure the prediction accuracy of the trained machine learning model.

[0043] Furthermore, the abnormal judgment module is used to obtain the vibration data and task information of the electromechanical device in real time, compare them with the corresponding vibration time-domain features and vibration frequency-domain features, and judge whether the electromechanical device is abnormal when executing the task information, including the following steps:

[0044] E1. Obtain the task information and load condition of the electromechanical device;

[0045] E2. Based on the task information and load condition, obtain the corresponding first subtask group and second subtask group;

[0046] E3. Obtain the real-time vibration data of the electromechanical device to obtain the corresponding real-time vibration time-domain data and real-time vibration frequency-domain data;

[0047] E4. Extract the features of the real-time vibration time-domain data and real-time vibration frequency-domain data respectively to obtain the real-time vibration time-domain features and real-time vibration frequency-domain features;

[0048] E5. Input the real-time vibration time-domain features into the abnormal prediction model. The abnormal prediction model inputs the real-time vibration time-domain features into the first abnormal prediction model i corresponding to the vibration frequency-domain data set corresponding to the first subtask group, and obtains the output prediction normal label or abnormal label;

[0049] E6. Input the real-time vibration frequency domain features into the anomaly prediction model. The anomaly prediction model inputs the real-time vibration frequency domain features into the second anomaly prediction model l corresponding to the vibration frequency domain dataset of the corresponding second subtask group, and obtains an output of a predicted normal label or anomaly label.

[0050] E7. Determine whether the labels predicted and output by the first anomaly prediction model i and the second anomaly prediction model l are both normal labels.

[0051] E8. If so, the electromechanical equipment is normal; if not, the electromechanical equipment is abnormal.

[0052] Compared with the prior art, a health prediction and management system for electromechanical equipment based on big data provided by the present invention can, by setting up a vibration monitoring module, a work information acquisition module, a data classification module, a standard data testing module, and an anomaly judgment module, monitor the task information and work load executed by the electromechanical equipment, judge whether the vibration of each part of the electromechanical equipment is abnormal under the execution of the corresponding task and corresponding load, and through the anomaly prediction of the part vibration, judge that there are potential health hazards in the parts with abnormal vibration of the electromechanical equipment that need to be repaired and maintained, thereby improving the judgment accuracy of the health of the parts of the electromechanical equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings.

[0054] Figure 1 It is a system structure block diagram provided by an embodiment of the present invention;

[0055] Figure 2 It is a schematic diagram of the classification steps of the vibration time domain dataset provided by an embodiment of the present invention;

[0056] Figure 3 It is a schematic diagram of the classification steps of the vibration frequency domain dataset provided by an embodiment of the present invention;

[0057] Figure 4 It is a schematic diagram of the anomaly judgment steps provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0058] In order to enable those skilled in the art to better understand the technical solutions of the present invention, the following will further introduce the present invention in detail with reference to the drawings.

[0059] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of the described features. In the description of the present invention, "a plurality of" means two or more unless otherwise specifically defined. In addition, the terms "mounted", "connected", and "coupled" should be construed broadly. For example, it may be a fixed connection, a detachable connection, or an integral connection; it may be a mechanical connection or an electrical connection; it may be directly connected or indirectly connected through an intermediate medium, and it may be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0060] Example embodiments will be described more fully hereinafter with reference to the accompanying drawings, but the example embodiments may be embodied in different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0061] In the case of no conflict, the various embodiments of the present disclosure and the features in the embodiments may be combined with each other. As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items. The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. As used herein, the singular forms "a" and "the" are also intended to include the plural forms unless the context clearly indicates otherwise. It will also be understood that when the terms "comprises" and / or "consists of" are used in this specification, it specifies the presence of the described features, integers, steps, operations, elements, and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0062] The embodiments described herein may be described with reference to the plan views and / or cross-sectional views by means of the ideal schematic diagrams of the present disclosure. Therefore, the example illustrations may be modified according to the manufacturing technology and / or tolerances. Therefore, the embodiments are not limited to the embodiments shown in the drawings, but include modifications of the configurations formed based on the manufacturing process. Therefore, the regions illustrated in the drawings have schematic properties, and the shapes of the regions shown in the drawings illustrate the specific shapes of the regions of the elements, but are not intended to be restrictive.

[0063] Please refer to Figures 1-4 , a health prediction and management system for electromechanical equipment based on big data, comprising a vibration monitoring module, a working information acquisition module, a data classification module, a standard data testing module, and an abnormality judgment module;

[0064] There are multiple vibration monitoring modules, which are used to monitor the vibration conditions of various parts of the electromechanical equipment respectively, and obtain the vibration data of each part. Among them, vibration sensors can be installed at the positions of various parts of the electromechanical equipment to monitor each part of the electromechanical equipment separately;

[0065] The working information acquisition module is used to obtain the task information executed by the electromechanical equipment, and the parts and loads working at each moment when each task information is executed. Among them, the task information can be set according to relevant industry rules and work experience. For example, the conventional cleaning, dehydration, self-cleaning, etc. of an automatic washing machine. Then, by obtaining the parts participating in the work when the automatic washing machine executes each task, the working parts are associated with the corresponding task information to obtain the part working conditions at each moment of each task information;

[0066] The data classification module is used to respectively obtain the vibration data of the parts corresponding to each task information, extract the characteristics in the time domain of the vibration data to obtain vibration time-domain data, and classify the vibration time-domain data according to different task information to obtain multiple vibration time-domain data sets, including the following steps:

[0067] A1. Obtain each task information and the working parts and loads corresponding to the task information;

[0068] A2. Classify the task information based on the differences of the working parts to obtain multiple task groups;

[0069] A3. Respectively extract the vibration time-domain data under each load condition of each task group;

[0070] A4. Obtain the vibration time-domain characteristics of the vibration time-domain data under each load condition of each task group;

[0071] A5. Compare the vibration time-domain characteristics, and divide the task groups and load conditions corresponding to the vibration time-domain characteristics whose similarity is greater than the set time-domain characteristic threshold into a first sub-task group, that is, divide multiple first sub-task groups under each task group, so that the vibration time-domain characteristics are similar when the parts of each first sub-task group execute the task information of the corresponding task group;

[0072] A6. Divide the vibration time-domain data corresponding to the vibration time-domain characteristics in each first sub-task group into a vibration time-domain data set.

[0073] The data classification module is also used to transform the vibration time-domain data into vibration frequency-domain data based on Fourier transform. The vibration frequency-domain data forms multiple vibration frequency-domain data sets classified according to different task information, including the following steps:

[0074] B1. Respectively transform the vibration time-domain data in each vibration time-domain data set to the frequency domain based on Fourier transform to obtain the corresponding vibration frequency-domain data; where the Fourier transform formula is:

[0075]

[0076] Among them, X(f) represents the frequency-domain signal, x(t) represents the time-domain signal, j is the imaginary unit, f represents the frequency, and t represents the time;

[0077] B2. Divide the vibration frequency-domain data according to the differences in the task groups corresponding to the vibration frequency-domain data.

[0078] B3. Obtain the vibration frequency-domain characteristics of the vibration frequency-domain data under various load conditions for each task group.

[0079] B4. Compare the vibration frequency-domain characteristics, and divide the task groups and load conditions corresponding to the vibration frequency-domain characteristics with a similarity greater than the set frequency-domain characteristic threshold into a second sub-task group.

[0080] B5. Divide the vibration frequency-domain data corresponding to the vibration frequency-domain characteristics in each second sub-task group into a vibration frequency-domain dataset.

[0081] The standard data test module is used to test the vibration time-domain data and vibration frequency-domain data of the electromechanical equipment when executing each task information, and extract the real-time vibration time-domain characteristics and real-time vibration frequency-domain characteristics of each task information based on the real-time vibration time-domain data and vibration frequency-domain data;

[0082] The anomaly judgment module is used to train a machine learning model based on the vibration time-domain dataset and the vibration frequency-domain dataset to determine whether the device is abnormal. In the present invention, the specific machine learning model is not limited. For example, support vector machine (SVM), random forest, K-nearest neighbor (KNN), neural network, regression model, etc. can be used. When extracting vibration time-domain features and vibration frequency-domain features, the vibration time-domain features include the mean value: the average value of the signal; the standard deviation: the degree of fluctuation of the signal; the peak value: the maximum amplitude of the signal; the root mean square value (RMS): the feature reflecting the signal energy; kurtosis and skewness: describing the sharpness and symmetry of the signal waveform. When extracting vibration frequency-domain features, first, through Fourier transform, the time-domain signal is converted into a frequency-domain signal to obtain spectral information. The vibration frequency-domain features include: the main frequency component: extracting the main frequencies in the spectrum; the band energy: calculating the energy distribution of each frequency band, such as the total energy, the energy of a specific frequency band; harmonic analysis: identifying the fundamental frequency and its harmonic components. Then, through feature selection techniques (such as PCA, LDA, tree-based feature importance, etc.) to screen the features that are most useful for judging the device state. When training the model, the dataset is divided into a training set and a test set, such as 70% training set and 30% test set; the selected machine learning model is trained using the training set; the hyperparameters of the model are optimized using methods such as cross-validation; indicators such as accuracy, recall rate, and F1-score are used to evaluate the performance of the model on the test set; the classification effect of the model is analyzed to check the recognition of normal and abnormal situations to ensure the prediction accuracy of the trained machine learning model.

[0083] including the following steps:

[0084] C1. Set normal labels and abnormal labels, and mark the vibration time-domain features in the vibration time-domain dataset and the vibration frequency-domain features in the vibration frequency-domain dataset. The marking can be achieved by associating normal vibration time-domain features and vibration frequency-domain features with normal labels, and abnormal vibration time-domain features and vibration frequency-domain features with abnormal labels;

[0085] C2. Based on the vibration time-domain features, normal labels, and abnormal labels corresponding to the vibration time-domain data in the i-th vibration time-domain dataset, train the first machine learning sub-model to obtain the first anomaly prediction model i, which is used to output the predicted normal label or abnormal label based on the input real-time vibration time-domain features, where i is initially 1;

[0086] C3. Let i = i + 1, and judge whether i ≤ n holds, where n is the number of vibration time-domain datasets;

[0087] C4. If so, return to C2;

[0088] C5. If not, integrate all the first anomaly prediction models i to obtain the first anomaly prediction model.

[0089] It further includes the following steps:

[0090] D1. Based on the vibration frequency domain features, normal labels, and abnormal labels corresponding to the vibration frequency domain data in the l-th vibration frequency domain dataset, train the second machine learning sub-model to obtain the second abnormal prediction model l, which is used to output the predicted normal label or abnormal label based on the input real-time vibration frequency domain features, where l is initially 1;

[0091] D2. Let l = l + 1, and determine whether l ≤ m holds, where m is the number of vibration frequency domain datasets;

[0092] D3. If yes, return to D1;

[0093] D4. If not, integrate all the second abnormal prediction models l to obtain the second abnormal prediction model;

[0094] D5. Integrate the first abnormal prediction model and the second abnormal prediction model to obtain the abnormal prediction model.

[0095] The abnormal judgment module is used to obtain the vibration data and task information of the electromechanical equipment in real time, compare them with the corresponding vibration time domain features and vibration frequency domain features, and judge whether the electromechanical equipment is abnormal when executing the task information, including the following steps:

[0096] E1. Obtain the task information and load condition of the electromechanical equipment;

[0097] E2. Based on the task information and load condition, obtain the corresponding first sub-task group and second sub-task group;

[0098] E3. Obtain the real-time vibration data of the electromechanical equipment to obtain the corresponding real-time vibration time domain data and real-time vibration frequency domain data;

[0099] E4. Extract the features of the real-time vibration time domain data and real-time vibration frequency domain data respectively to obtain the real-time vibration time domain features and real-time vibration frequency domain features;

[0100] E5. Input the real-time vibration time domain features into the abnormal prediction model, and the abnormal prediction model inputs the real-time vibration time domain features into the first abnormal prediction model i corresponding to the vibration frequency domain dataset corresponding to the corresponding first sub-task group to obtain the output predicted normal label or abnormal label;

[0101] E6. Input the real-time vibration frequency domain features into the abnormal prediction model, and the abnormal prediction model inputs the real-time vibration frequency domain features into the second abnormal prediction model l corresponding to the vibration frequency domain dataset corresponding to the corresponding second sub-task group to obtain the output predicted normal label or abnormal label;

[0102] E7. Judge the first abnormal prediction model i and the second abnormal prediction model l, and determine whether the output predicted labels are both normal labels;

[0103] E8. If so, the electromechanical equipment is normal; if not, the electromechanical equipment is abnormal, and there are abnormalities in the corresponding parts of the real-time vibration time domain characteristics or real-time vibration frequency domain characteristics corresponding to the further abnormal labels.

[0104] Only some exemplary embodiments of the present invention are described above by way of illustration. Without doubt, for those of ordinary skill in the art, the described embodiments can be modified in various different ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the protection scope of the claims of the present invention.

Claims

1. A health prediction and management system for electromechanical equipment based on big data, characterized in that: It includes a vibration monitoring module, a working information acquisition module, a data classification module, a standard data testing module, and an anomaly judgment module; There are multiple vibration monitoring modules, which are used to monitor the vibration conditions of various parts of the electromechanical equipment respectively, and obtain the vibration data of each part; The working information acquisition module is used to acquire the task information executed by the electromechanical equipment, and the parts and loads working at each moment when each task information is executed; The data classification module is used to respectively obtain the vibration data of the parts corresponding to each task information, extract the features in the time domain of the vibration data to obtain vibration time-domain data, and classify the vibration time-domain data according to different task information to obtain multiple vibration time-domain data sets; The data classification module is also used to transform the vibration time-domain data into vibration frequency-domain data based on Fourier transform, and the vibration frequency-domain data forms multiple vibration frequency-domain data sets classified according to different task information; The standard data testing module is used to test the vibration time-domain data and vibration frequency-domain data of the electromechanical equipment when each task information is executed, and extract the real-time vibration time-domain features and real-time vibration frequency-domain features of each task information based on the real-time vibration time-domain data and vibration frequency-domain data; The anomaly judgment module is used to obtain the vibration data and task information of the electromechanical equipment in real time, compare them with the corresponding vibration time-domain features and vibration frequency-domain features, and judge whether the electromechanical equipment is abnormal when executing the task information.

2. The health prediction and management system for electromechanical equipment based on big data according to claim 1, wherein: The data classification module is used to respectively obtain the vibration data of the parts corresponding to each task information, extract the features in the time domain of the vibration data to obtain vibration time-domain data, and classify the vibration time-domain data according to different task information to obtain multiple vibration time-domain data sets, including the following steps: A1. Obtain each task information and the parts and loads working corresponding to the task information; A2. Classify the task information based on the differences of the working parts to obtain multiple task groups; A3. Respectively extract the vibration time-domain data under each load condition of each task group; A4. Obtain the vibration time-domain features of the vibration time-domain data under each load condition of each task group; A5. Compare the vibration time-domain features, and divide the task groups and load conditions corresponding to the vibration time-domain features with a similarity greater than the set time-domain feature threshold into a first sub-task group; A6. Divide the vibration time-domain data corresponding to the vibration time-domain features in each first sub-task group into a vibration time-domain data set.

3. The health prediction and management system for electromechanical equipment based on big data according to claim 2, characterized in that: The data classification module is also used to transform the vibration time-domain data into vibration frequency-domain data based on Fourier transform, and the vibration frequency-domain data forms multiple vibration frequency-domain data sets classified according to different task information, including the following steps: B1. Respectively transform the vibration time-domain data in each vibration time-domain data set to the frequency domain based on Fourier transform to obtain the corresponding vibration frequency-domain data; B2. Divide the vibration frequency-domain data based on the differences of the task groups of the task information corresponding to the vibration frequency-domain data; B3. Obtain the vibration frequency-domain features of the vibration frequency-domain data under each load condition of each task group; B4. Compare the vibration frequency domain features, and divide the task groups and load conditions corresponding to the vibration frequency domain features with a similarity greater than the set frequency domain feature threshold into a second sub-task group; B5. Divide the vibration frequency domain data corresponding to the vibration frequency domain features in each second sub-task group into a vibration frequency domain data set.

4. An electromechanical equipment health prediction and management system based on big data according to claim 3, characterized in that: The abnormality judgment module is also used to train a machine learning model based on the vibration time domain data set and the vibration frequency domain data set to judge whether the device is abnormal, including the following steps: C1. Set normal labels and abnormal labels, and mark the vibration time domain features in the vibration time domain data set and the vibration frequency domain features in the vibration frequency domain data set; C2. Based on the vibration time domain features, normal labels, and abnormal labels corresponding to the vibration time domain data in the i-th vibration time domain data set, train a first machine learning sub-model to obtain a first abnormality prediction model i, which is used to output a predicted normal label or abnormal label based on the input real-time vibration time domain features, where i is initially 1; C3. Let i = i + 1, and judge whether i ≤ n holds, where n is the number of vibration time domain data sets; C4. If so, return to C2; C5. If not, integrate all the first abnormality prediction models i to obtain a first abnormality prediction model.

5. The health prediction and management system for electromechanical equipment based on big data according to claim 4, wherein: The abnormality judgment module is also used to train a machine learning model based on the vibration time domain data set and the vibration frequency domain data set to judge whether the device is abnormal, and further includes the following steps: D1. Based on the vibration frequency domain features, normal labels, and abnormal labels corresponding to the vibration frequency domain data in the l-th vibration frequency domain data set, train a second machine learning sub-model to obtain a second abnormality prediction model l, which is used to output a predicted normal label or abnormal label based on the input real-time vibration frequency domain features, where l is initially 1; D2. Let l = l + 1, and judge whether l ≤ m holds, where m is the number of vibration frequency domain data sets; D3. If so, return to D1; D4. If not, integrate all the second abnormality prediction models l to obtain a second abnormality prediction model; D5. Integrate the first abnormality prediction model and the second abnormality prediction model to obtain an abnormality prediction model.

6. The health prediction and management system for electromechanical equipment based on big data according to claim 5, characterized in that: The abnormality judgment module is used to obtain the vibration data and task information of the electromechanical device in real time, compare them with the corresponding vibration time domain features and vibration frequency domain features, and judge whether the electromechanical device is abnormal when executing the task information, including the following steps: E1. Obtain the task information and load condition of the electromechanical device; E2. Based on the task information and load condition, obtain the corresponding first sub-task group and second sub-task group; E3. Obtain the real-time vibration data of the electromechanical device to obtain the corresponding real-time vibration time domain data and real-time vibration frequency domain data; E4. Extract the features of the real-time vibration time domain data and the real-time vibration frequency domain data respectively to obtain the real-time vibration time domain features and the real-time vibration frequency domain features; E5. Input the real-time vibration time domain features into the abnormality prediction model, and the abnormality prediction model inputs the real-time vibration time domain features into the first abnormality prediction model i corresponding to the vibration frequency domain data set corresponding to the corresponding first sub-task group to obtain an output predicted normal label or abnormal label; E6. Input the real-time vibration frequency domain features into the anomaly prediction model. The anomaly prediction model inputs the real-time vibration frequency domain features into the second anomaly prediction model l corresponding to the vibration frequency domain dataset of the corresponding second subtask group, and obtains an output prediction of a normal label or an abnormal label. E7. Determine whether the labels predicted by the first anomaly prediction model i and the second anomaly prediction model l are both normal labels. E8. If so, the electromechanical equipment is normal; if not, the electromechanical equipment is abnormal.

Images