An online fault diagnosis method and device for a bearing of a coal mine mechanical equipment

Abstract

The application provides a coal mine mechanical equipment bearing online fault diagnosis method and device, and belongs to the technical field of bearing fault online diagnosis; the application solves the problem of complex algorithm of existing bearing online diagnosis or detection, which is not conducive to the online arrangement of diagnosis capacity; and comprises the following steps: bearing vibration related signal collection: collecting the vibration signal and rotation speed of the bearing, and dividing according to frames; bearing vibration dimension feature extraction: according to the vibration signal, rotation speed and sampling frequency of the bearing, the bearing vibration features-energy level and energy factor are sequentially calculated; bearing vibration time feature extraction: according to the energy factor, the time information of the energy factor in a certain time interval is sequentially calculated; bearing fault diagnosis of fused dimension-time features: according to the dimension information-energy level and the time information of the energy factor in a certain time interval, whether the bearing is faulty at a certain moment, and the damage type and severity of the bearing are given; the application is applied to bearing fault diagnosis.

Description

Technical Field

[0001] This invention provides a method and device for online fault diagnosis of bearings in coal mining machinery and equipment, belonging to the field of online bearing fault diagnosis technology. Background Technology

[0002] Vibration signal analysis has been widely applied to the detection and diagnosis of bearing faults in coal mining machinery. Numerous time-domain and frequency-domain analysis techniques, as well as information fusion techniques, have been proposed to extract and synthesize relevant information about bearing health from vibration signals for the detection and diagnosis of bearing faults. However, in mine operations, due to the strong broadband background noise of coal mining machinery and the inherent uncertainty in the occurrence and development of bearing faults, bearing vibration signals often become exceptionally complex. A single vibration feature is insufficient to accurately reflect the bearing's operational health, and different vibration features exhibit significant dispersion or performance differences in capturing the dynamic response of the bearing to damage development. Information fusion techniques based purely on big data not only have complex algorithms that hinder the online deployment of diagnostic capabilities but may even lead to mediocre diagnostic results. Diagnostic capability refers to the hardware and software system required to implement fault diagnosis functions. Different fault diagnosis algorithms require varying levels of complexity in their corresponding hardware and software implementation systems, such as the necessary storage, computing, and communication chips. Therefore, how to track the actual health status of bearings in coal mine machinery and equipment in real time, and improve the accuracy of bearing fault detection and diagnosis under a given false alarm rate, remains a key technical issue that needs to be addressed to avoid secondary damage and unplanned downtime caused by early bearing failures in mine operations, and to achieve safe, reliable, and economical operation of coal mine machinery and equipment. Summary of the Invention

[0003] To address the problem of complex algorithms hindering diagnostic capabilities in existing online bearing diagnosis or detection methods, this invention proposes an online fault diagnosis method and device for bearings in coal mine machinery.

[0004] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: an online fault diagnosis method for bearings of coal mine machinery and equipment, comprising the following steps:

[0005] S1: Bearing vibration related signal acquisition: Acquire the bearing vibration signal, rotational speed, and sampling frequency, and divide it into frames;

[0006] S2: Dimensional feature extraction of bearing vibration: Based on the bearing vibration signal, rotational speed, and sampling frequency, sequentially calculate the bearing vibration characteristics - energy level RMSband, E(f) and energy factor EI;

[0007] S3: Time Feature Extraction of Bearing Vibration: Based on the energy factor EI, sequentially calculate the time features from time t1 to t2. nTemporal information of the energy factor EI within the time interval of time;

[0008] S4: Bearing fault diagnosis based on fusion dimension - time characteristics: according to t n Bearing vibration characteristics at time t1 - energy level RMSband, E(f), and time from t1 to t2 n The time information of the energy factor EI within the time interval of time t, giving t n It is important to know whether the bearing has failed at any time, and the type and severity of the bearing damage.

[0009] The steps for extracting the dimensional features of bearing vibration in step S2 are as follows:

[0010] S21: Bandpass filtering is used to extract the vibration signal of the bearing's resonant frequency band;

[0011] S22: Calculate the energy level RMSband of the bearing resonant frequency band based on the vibration signal of the bearing resonant frequency band;

[0012] S23: The envelope spectrum of the bearing resonance frequency band vibration signal is extracted using Hilbert transform and Fourier transform;

[0013] S24: Calculate the energy level E(f) of the bearing characteristic frequency based on the envelope spectrum of the bearing resonance frequency band vibration signal;

[0014] S25: Calculate the energy level RMS of the entire frequency band of the bearing based on the bearing vibration signal;

[0015] S26: Calculate the energy factor EI based on the energy level RMSband of the bearing resonant frequency band and the energy level RMS of the entire bearing frequency band.

[0016] The steps for extracting the time features of bearing vibration in step S3 are as follows:

[0017] S31: Calculate t using the Trigg tracking variable algorithm with smoothing parameters that change over time. n The tracking variable for the time-matter energy factor EI;

[0018] S32: The smoothing parameter is approximated by numerically calculating the normalized variance of the predicted signal in the Trigg tracking variable algorithm to a very small constant.

[0019] S33: According to t n The tracking variable of the time-energy factor EI is given from time t1 to t n The temporal information of the energy factor EI within the time interval of time.

[0020] The steps for fusion of dimension-time features in bearing fault diagnosis in step S4 are as follows:

[0021] S41: Using a Naive Bayesian network model for bearing fault diagnosis;

[0022] S42: Define the health status of the bearing as the root node of the Naive Bayesian network model, and define the dimensional-time characteristics of the bearing vibration as the leaf nodes of the Naive Bayesian network model.

[0023] S43: Use maximum likelihood estimation on complete data to train the conditional probability distribution parameters of the model;

[0024] S44: t predicted based on the Naive Bayes network model n The conditional probability distribution of the bearing's health status at time t gives t n The uncertainty includes whether the bearing has failed, the type and severity of bearing damage, and the corresponding uncertainty.

[0025] An online fault diagnosis device for bearings in coal mining machinery includes:

[0026] Bearing vibration related signal acquisition module: used to acquire the vibration signal, rotational speed, and sampling frequency of the bearing, and to segment it into frames and send it to the bearing vibration dimensional feature extraction module;

[0027] The bearing vibration dimensional feature extraction module is used to sequentially calculate the bearing vibration features - energy level RMSband, E(f) and energy factor EI based on the bearing vibration signal, rotational speed and sampling frequency. The bearing vibration features - energy level RMSband and E(f) are sent as the dimensional features of bearing vibration to the bearing fault diagnosis module that integrates dimensional-time features, and the energy factor is sent to the bearing vibration time feature extraction module.

[0028] The bearing vibration time feature extraction module: Based on the energy factor EI, it sequentially calculates the time feature from time t1 to t2. n The time information of the energy factor EI within the time interval is sent to the bearing fault diagnosis module that integrates the dimension-time features.

[0029] The bearing fault diagnosis module integrates time-based features: used to diagnose bearing faults based on time characteristics. n Bearing vibration characteristics at time t1 - energy levels RMSband, E(f), and from time t1 to t2. n The time information of the energy factor EI within the time interval of time t, giving t n It is important to know whether the bearing has failed at any time, and the type and severity of the bearing damage.

[0030] The bearing vibration dimensional feature extraction module uses bandpass filtering to extract the vibration signal of the bearing resonant frequency band;

[0031] The bearing vibration dimensional feature extraction module calculates the energy level RMSband of the bearing resonance frequency band based on the vibration signal of the bearing resonance frequency band.

[0032] The bearing vibration dimensional feature extraction module uses Hilbert transform and Fourier transform to extract the envelope spectrum of the bearing resonance frequency band vibration signal;

[0033] The bearing vibration dimensional feature extraction module calculates the energy level E(f) of the bearing characteristic frequency based on the envelope spectrum of the bearing resonance frequency band vibration signal.

[0034] The bearing vibration dimensional feature extraction module calculates the energy level RMS of the entire frequency band of the bearing based on the bearing vibration signal.

[0035] The bearing vibration dimensional feature extraction module calculates the energy factor EI based on the energy level RMSband of the bearing resonance frequency band and the energy level RMS of the entire bearing frequency band.

[0036] The bearing vibration dimensional feature extraction module uses a Trigg tracking variable algorithm with smooth parameters changing over time to calculate t. n The tracking variable for the time-matter energy factor EI;

[0037] The bearing vibration dimensional feature extraction module performs approximate numerical calculation of the smoothing parameter by limiting the normalized variance of the predicted signal in the Trigg tracking variable algorithm to a very small constant.

[0038] The bearing vibration dimensional feature extraction module is based on t n The tracking variable of the time-energy factor EI is given from time t1 to t n The temporal information of the energy factor EI within the time interval of time.

[0039] The bearing fault diagnosis module that integrates dimensional and temporal features uses a Naive Bayes network model for bearing fault diagnosis.

[0040] The bearing fault diagnosis module that integrates dimension-time features defines the bearing health status as the root node of the Naive Bayes network model and the dimension-time features of bearing vibration as the leaf nodes of the Naive Bayes network model.

[0041] The bearing fault diagnosis module that integrates dimensional and temporal features uses maximum likelihood estimation of the conditional probability distribution parameters of the training model under complete data.

[0042] The bearing fault diagnosis module, which integrates dimensional and temporal features, predicts t based on the Naive Bayes network model. n The conditional probability distribution of the bearing's health status at time t gives t nThe uncertainty includes whether the bearing has failed, the type and severity of bearing damage, and the corresponding uncertainty.

[0043] The advantages of this invention compared to existing technologies are as follows: This invention constructs dimensional features reflecting the current vibration mode of the bearing and temporal features reflecting the changing trend of the bearing vibration mode. These vibration features are complementary in information and have clear physical meaning in engineering, which can accurately track the actual health status of bearings in coal mining machinery and improve the accuracy of bearing fault detection and diagnosis under a given false alarm rate. The feature extraction and information fusion algorithms used in this invention are simple and have low computational resource requirements. The constructed fault diagnosis network model has strong generalization ability, which makes it easy for coal mining machinery manufacturers and users to flexibly configure fault diagnosis capabilities online as needed. Attached Figure Description

[0044] The present invention will be further described below with reference to the accompanying drawings:

[0045] Figure 1 This is a flowchart of the method of the present invention;

[0046] Figure 2 This is a schematic diagram of the Naive Bayes network model structure used in this invention. Detailed Implementation

[0047] This invention proposes an online fault diagnosis method for bearings in coal mine machinery equipment, the flowchart of which is shown below. Figure 1 As shown, it includes the following steps:

[0048] S1, Bearing vibration related signal acquisition: Acquires the bearing vibration signal and rotational speed, segments them into frames, and sends them to the bearing vibration dimensional feature extraction module.

[0049] S2, Dimensional feature extraction of bearing vibration: Based on the bearing vibration signal, rotational speed, and sampling frequency, sequentially calculate the bearing vibration features - energy level RMSband, E(f) and energy factor EI. The bearing vibration features - energy level RMSband and E(f) are sent as dimensional features of bearing vibration to the bearing fault diagnosis module that integrates dimensional-time features, and the energy factor EI is sent to the bearing vibration time feature extraction module.

[0050] The dimensional feature extraction of bearing vibration in step S2, when calculating the bearing vibration features—energy level RMSband, E(f), and energy factor EI—includes the following steps:

[0051] S21, the dimensional feature extraction of bearing vibration is performed by using bandpass filtering to extract the vibration signal of the bearing resonant frequency band.

[0052] The cutoff frequency of the bandpass filter is set according to the resonant frequency range of the bearing and support structure in actual applications.

[0053] S22, Dimensional feature extraction of bearing vibration: Calculate the energy level RMSband of the bearing resonant frequency band based on the vibration signal of the bearing resonant frequency band.

[0054] The energy level RMSband of the bearing resonant frequency band is achieved by the following formula:

[0055]

[0056] In the above formula: x(i) is the bandpass filtered signal. K is the mean of this signal, and K is the number of samples of this signal.

[0057] S23, the dimensional feature extraction of bearing vibration is performed by using Hilbert transform and Fourier transform to extract the envelope spectrum of the bearing resonant frequency band vibration signal.

[0058] S24, Dimensional feature extraction of bearing vibration: Calculate the energy level E(f) of the bearing characteristic frequency based on the envelope spectrum of the bearing resonant frequency band vibration signal.

[0059] The energy level E(f) of the bearing's characteristic frequency is achieved by the following formula:

[0060]

[0061] Where Env(f) is the envelope spectrum of the bearing resonant frequency band vibration signal, and Δf / 2 is the neighborhood radius of the bearing characteristic frequency f.

[0062] The bearing has characteristic frequencies f corresponding to the inner ring, outer ring, and rolling elements. bpfi f bpfo and 2f bsf This can be achieved using the following formula:

[0063]

[0064]

[0065]

[0066] In the above formula: f s ζ is the frequency of the rotating shaft, ζ is the contact angle of the bearing rolling element, D is the diameter of the bearing rolling element, and d is the bearing rolling element diameter. m Z is the bearing pitch diameter, and Z is the number of rolling elements in the bearing.

[0067] S25, Dimensional feature extraction of bearing vibration: Calculate the energy level RMS of the entire frequency band of the bearing based on the bearing vibration signal.

[0068] The energy level RMS of the bearing across the entire frequency band is achieved by formula (1).

[0069] S26, Dimensional feature extraction of bearing vibration: The energy factor EI is calculated based on the energy level RMSband of the bearing resonance frequency band and the energy level RMS of the entire bearing frequency band.

[0070] The energy factor EI is achieved through the following formula:

[0071] EI = RMSband / RMS (6).

[0072] S3, Time feature extraction of bearing vibration: Based on the energy factor EI, the time features from time t1 to t2 are calculated sequentially. n The time information of the energy factor EI within the time interval is sent to the bearing fault diagnosis module that integrates the dimension-time features.

[0073] In step S3, the extraction of the bearing vibration time characteristics is performed when calculating the time from time t1 to t2. n When obtaining the time information of the energy factor EI within a time interval, the following steps are included:

[0074] S31, the bearing vibration dimensional feature extraction module uses the Trigg tracking variable algorithm with smooth parameters changing over time to calculate t. n The tracking variable for the time-matter energy factor EI.

[0075] The t n The tracking variable for the time-matter energy factor EI is achieved through the following formula:

[0076] T(t i )=s(t i ) / M(t j (7);

[0077] s(t i )=αe(t i )+(1-α)s(t i-1 (8);

[0078] M(t i )=α|e(t i )|+(1-α)M(t i-1 (9);

[0079] e(t i )=d(t i )-u(t i (10);

[0080] u(t i )=αd(t i-1)+(1-α)u(t i-1 (11);

[0081] Where T(t) i ) is t i The tracking variable at time t, s(t) i ) is t i Smoothing error at time step, M(t) i ) is t i The absolute smoothing error at time t, e(t) i ) is t i Signal prediction error at time t, d(t) i ) is t i The original signal at time t, u(t) i ) is t i The predicted signal at time t is α, which is a smoothing parameter between 0 and 1.

[0082] S32, the dimensional feature extraction of bearing vibration is approximated by numerically calculating the smoothing parameter by limiting the normalized variance of the predicted signal in the Trigg tracking variable algorithm to a very small constant.

[0083] The smoothing parameter is achieved through the following formula:

[0084]

[0085]

[0086]

[0087] Where α(t) i-1 ) is t i-1 The smoothing parameter at time t, d(t) i-1 ),d 2 (t i-1 ) are respectively t i-1 The original signal at time t, the square of the original signal, u(t) i-1 ) is t i-1 The predicted signal at time t, E(d(t) i-1 )), E(d 2 (t i-1 )) are respectively d(t i-1 ),d 2 (t i-1 The mean of ), var(d(t) i-1 )) is d(t i-1 The variance of ε is a very small constant, 0 < ε << 1.

[0088] S33, Dimensional feature extraction of bearing vibration based on t nThe tracking variable of the time-energy factor EI is given from time t1 to t n The temporal information of the energy factor EI within the time interval of time.

[0089] The time from t1 to t n The time information of the energy factor EI within the time interval is realized by the following formula:

[0090]

[0091] Bearing vibration characteristics – energy level RMSband, E(f) and energy factor EI – are in a complementary form, and together they can effectively indicate the type and severity of bearing damage.

[0092] Bearing vibration characteristics - energy level RMSband represents the energy level of the vibration signal within the bearing's resonant frequency band.

[0093] Bearing vibration characteristics - energy level E(f), f = f bpfi ,f bpfo ,2f bsf This reflects the periodic repetitive vibration pulses in the bearing vibration signal. It can not only indicate the degree of bearing damage, but also infer the failure mode of bearing damage, especially in the early stages of bearing failure.

[0094] The energy factor (EI) captures the dynamic response of a bearing to damage, and its temporal information can be used to enhance the ability of other vibration characteristics to infer bearing damage. Specifically, in the early stages of bearing failure, the significant upward trend of the energy factor EI can be combined with the energy level E(f) to improve the diagnostic capability of early bearing damage. In the later stages of bearing failure, as the bearing vibration mode tends to be random, the variance of the bearing vibration characteristic-energy level RMSband becomes large, and E(f) becomes indistinguishable. At this time, the significant downward trend of the energy factor EI can be used to compensate for these negative impacts on bearing failure diagnosis.

[0095] S4, bearing fault diagnosis integrating time-dimensional features, based on t n Bearing vibration characteristics at time t1 - energy level RMSband, E(f), and from time t1 to t2 n The time information of the energy factor EI within the time interval of time t, giving t n It is important to know whether the bearing has failed at any time, and the type and severity of the bearing damage.

[0096] In step S4, the bearing fault diagnosis that integrates dimension-time features is given by t. n When determining whether a bearing has failed, and the type and severity of bearing damage, the following steps are included:

[0097] S41, the bearing fault diagnosis module that integrates dimensional and temporal features uses a Naive Bayes network model for bearing fault diagnosis.

[0098] The Naive Bayes network can effectively solve the following three problems: the need to simultaneously integrate information from multiple vibration features [RMSband,T(EI),E(f)] bpfi ),E(f bpfo ),E(2f bsf It is necessary to fuse two different types of data, namely continuous dimensional features and discrete temporal features. When using information fusion technology to integrate the dimensional and temporal features of a bearing, it is necessary not only to provide whether the bearing has failed and the severity of the damage, but also to provide the uncertainty of its diagnostic results.

[0099] The Naive Bayes network describes the distribution characteristics of data through a simple model, and is therefore not sensitive to data fluctuations (such as noise). The learned model has good generalization ability.

[0100] S42, Dimension-Time Features Fusion Bearing Fault Diagnosis: The health status of the bearing is defined as the root node of the Naive Bayesian network model, and the dimension-time features of bearing vibration are defined as the leaf nodes of the Naive Bayesian network model.

[0101] The Naive Bayes network model for bearing fault diagnosis is as follows: Figure 2 As shown, State indicates the health status of the bearing, RMSband, T(EI), E(f) bpfi ), E(f bpfo) E(2f) bsf The indicator shows the dimensional-time characteristics of bearing vibration, with gray representing observed variables, white representing implicit variables, squares representing discrete variables, and circles representing continuous variables.

[0102] The bearing health status includes seven states: bearing in good condition, early failure of inner ring, serious failure of inner ring, early failure of outer ring, serious failure of outer ring, early failure of rolling element, and serious failure of rolling element.

[0103] S43, the bearing fault diagnosis module that integrates dimensional and temporal features uses maximum likelihood estimation of the conditional probability distribution parameters of the training model under complete data.

[0104] The conditional probability distribution parameters of the bearing fault diagnosis model are achieved through the following formula:

[0105]

[0106]

[0107] L i (h)=p{State i RMSband i T i (EI), E i (f bpfi E i (f bpfo E i (2f bsf )}

[0108] =p{State i}p{RMSband i State i}p{T i (EI)|State i}p{E i (f bpfi )|State i}

[0109] p{E i (f bpfo )|State i}p{E i (2f bsf )|State i} (18);

[0110] Where h is a parameter of the conditional probability distribution p{t|h} in the network model, h ML It is the distribution parameter that maximizes p{t|h}; L(h) is the training data of the network model. The likelihood function, X n The bearing vibration characteristics are represented by [RMSband,T(EI),E(f)]. bpfi ),E(f bpfo ),E(2f bsf The observed value of t] n The observed value representing the bearing's health condition (State); L i (h) is given an observation [State] i ,RMSband i ,T i (EI),E i (f bpfi ),E i (f bpfo ),E i (2f bsf The likelihood function under [ )].

[0111] In the bearing fault diagnosis model, the continuous random variable RMSband, E(f)bpfi ), E(f bpfo ), E(2f bsf Assume that it follows a Gaussian distribution.

[0112] S44, bearing fault diagnosis integrating dimensional-temporal features based on t predicted by a Naive Bayes network model. n The conditional probability distribution of the bearing's health status at time t gives t n The uncertainty includes whether the bearing has failed, the type and severity of bearing damage, and the corresponding uncertainty.

[0113] The conditional probability distribution of the health status of the bearing is achieved by the following formula:

[0114] p{State * |t}≈p{State * |h ML} (19);

[0115] Where h ML It is the distribution parameter that maximizes p{t|h}.

[0116] This invention constructs a set of vibration features that are complementary in information and have clear physical meaning in engineering, in order to accurately track the actual health status of bearings in coal mining machinery. Then, the vibration features of the bearings are integrated by an information fusion algorithm with low computational resource requirements and strong generalization ability to infer whether the bearing has failed, the type and severity of the bearing damage, and the corresponding uncertainty, so as to achieve robust online fault diagnosis of bearings in coal mining machinery.

[0117] This invention also proposes an online fault diagnosis device for bearings in coal mining machinery, including a bearing vibration-related signal acquisition module, a bearing vibration dimensional feature extraction module, a bearing vibration temporal feature extraction module, and a bearing fault diagnosis module that integrates dimensional and temporal features. Addressing the strong broadband background noise in coal mining machinery and the inherent uncertainty in the occurrence and development of bearing faults, the dimensional and temporal feature extraction modules respectively construct dimensional features reflecting the current vibration mode of the bearing and temporal features reflecting the changing trend of the bearing vibration mode. Then, the bearing fault diagnosis module, integrating the dimensional and temporal features of the bearing vibration, provides information on whether a bearing fault has occurred, as well as the type and severity of bearing damage, thereby improving the accuracy of bearing fault diagnosis under a given false alarm rate.

[0118] Regarding the specific structure of this invention, it should be noted that the connection relationships between the various component modules used in this invention are definite and achievable. Except as specifically described in the embodiments, their specific connection relationships can bring about corresponding technical effects and solve the technical problems proposed by this invention without relying on the execution of corresponding software programs. The models of the components, modules, and specific components appearing in this invention, the connection methods between them, and the conventional usage methods and expected technical effects brought about by the above technical features, unless specifically described, are all publicly disclosed content in patents, journal articles, technical manuals, technical dictionaries, and textbooks that can be obtained by those skilled in the art before the application date, or belong to conventional technology, common knowledge, and other existing technologies in this field. There is no need to elaborate, which makes the technical solution provided in this case clear, complete, and achievable, and can reproduce or obtain corresponding physical products based on this technical means.

[0119] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for on-line fault diagnosis of a bearing of a coal mine mechanical equipment, characterized in that: The method comprises the following steps: S1: collecting bearing vibration related signals: collecting the vibration signals, rotating speed, sampling frequency of the bearing, and dividing them by frames; S2: extracting the dimensional characteristics of bearing vibration: according to the vibration signals, rotating speed, sampling frequency of the bearing, sequentially calculating the bearing vibration characteristics-energy level RMSband, E(f) and energy factor EI; The step S2 of extracting the dimensional characteristics of bearing vibration is as follows: S21: extracting the vibration signals of the bearing resonance frequency band by using band-pass filtering; S22: calculating the energy level RMSband of the bearing resonance frequency band according to the vibration signals of the bearing resonance frequency band; S23: extracting the envelope spectrum of the bearing resonance frequency band vibration signals by using Hilbert transform and Fourier transform; S24: calculating the energy level E(f) of the bearing characteristic frequency according to the envelope spectrum of the bearing resonance frequency band vibration signals; S25: calculating the energy level RMS of the whole frequency band of the bearing according to the bearing vibration signals; S26: calculating the energy factor EI according to the energy level RMSband of the bearing resonance frequency band and the energy level RMS of the whole frequency band of the bearing; S3: Time feature extraction of bearing vibration: According to the energy factor EI, the time information of the energy factor EI in the time interval from t1 time to t n time is calculated sequentially. S4: Bearing fault diagnosis based on fusion dimension - time characteristics: according to t n Bearing vibration characteristics at time t1 - energy level RMSband, E(f), and time from t1 to t2 n The time information of the energy factor EI within the time interval of time t, giving t n It is important to know whether the bearing has failed at any time, and the type and severity of the bearing damage.

2. The online fault diagnosis method for coal mine mechanical equipment bearing according to claim 1, characterized in that: The step S3 of extracting the time characteristics of bearing vibration is as follows: S31: Calculate t using the Trigg tracking variable algorithm with time-varying smoothing parameters n Tracking variable for the moment energy factor E; S32: performing approximate numerical calculation on the smoothing parameter by limiting the normalized variance of the predicted signal in the Trigg tracking variable algorithm to a very small constant; S33: According to t n The tracking variable of the energy factor EI gives the time information of the energy factor EI in the time interval from the time instant t1 to the time instant t n The tracking variable of the energy factor EI gives the time information of the energy factor EI in the time interval from the time instant t1 to the time instant t 3. The online fault diagnosis method for coal mine mechanical equipment bearing according to claim 2, characterized in that: The step S4 of fusing the dimensional-time characteristics for bearing fault diagnosis is as follows: S41: adopting the naive Bayes network model to perform bearing fault diagnosis; S42: defining the health condition of the bearing as the root node of the naive Bayes network model, and defining the dimensional-time characteristics of the bearing vibration as the leaf node of the naive Bayes network model; S43: training the conditional probability distribution parameters of the model by using the maximum likelihood estimation under complete data; S44: t predicted according to the Naive Bayes network model n The conditional probability distribution of the health condition of the bearing at time t gives t n whether the bearing fails at time t, the type and severity of the damage of the bearing, and the corresponding uncertainty.

4. An apparatus for on-line fault diagnosis of a bearing of a coal mine mechanical equipment, characterized in that: It comprises: A bearing vibration related signal collection module: used for collecting the vibration signals, rotating speed, sampling frequency of the bearing, and dividing them by frames, and sending them to the bearing vibration dimensional characteristic extraction module; A bearing vibration dimensional characteristic extraction module: used for sequentially calculating the bearing vibration characteristics-energy level RMSband, E(f) and energy factor EI according to the vibration signals, rotating speed, sampling frequency of the bearing, wherein the bearing vibration characteristics-energy level RMSband, E(f) are sent to the bearing fault diagnosis module fusing the dimensional-time characteristics as the dimensional characteristics of the bearing vibration, and the energy factor EI is sent to the bearing vibration time characteristic extraction module; The bearing vibration dimensional characteristic extraction module extracts the vibration signals of the bearing resonance frequency band by using band-pass filtering; The bearing vibration dimensional characteristic extraction module calculates the energy level RMSband of the bearing resonance frequency band according to the vibration signals of the bearing resonance frequency band; The bearing vibration dimensional characteristic extraction module extracts the envelope spectrum of the bearing resonance frequency band vibration signals by using Hilbert transform and Fourier transform; The bearing vibration dimensional characteristic extraction module calculates the energy level E(f) of the bearing characteristic frequency according to the envelope spectrum of the bearing resonance frequency band vibration signals; The dimension feature extraction module of the bearing vibration calculates the energy level RMS of the whole frequency band of the bearing vibration signal; The dimension feature extraction module of the bearing vibration calculates the energy factor EI according to the energy level RMSband of the resonance frequency band of the bearing and the energy level RMS of the whole frequency band of the bearing; The time feature extraction module of bearing vibration: according to the energy factor EI, the time information of energy factor EI in the time interval from t1 moment to t n The fusion dimension-time feature bearing fault diagnosis module is sent to the fusion dimension-time feature bearing fault diagnosis module. A bearing fault diagnosis module fusing dimension-time features: for determining whether a bearing has failed at time t n and the time information of the energy factor EI within the time interval from time t n to time t n indicates whether the bearing has failed at time t and the type and severity of the damage of the bearing.

5. The online fault diagnosis device for bearings of coal mine machinery equipment according to claim 4, characterized in that: The dimension feature extraction module of the bearing vibration adopts the Trigg tracking variable algorithm with time-varying smoothing parameters to calculate t n the tracking variable of the moment energy factor EI; The dimension feature extraction module of the bearing vibration performs approximate numerical calculation on the smoothing parameter by limiting the normalized variance of the prediction signal in the Trigg tracking variable algorithm to a very small constant; The dimension feature extraction module of the bearing vibration extracts the time information of the energy factor EI in the time interval from t1 to t n The tracking variable of the energy factor EI at time t gives the time information of the energy factor EI in the time interval from t1 to t n t.

6. The online fault diagnosis device for bearings of coal mine machinery equipment according to claim 5, characterized in that: The bearing fault diagnosis module fusing the dimension-time features adopts a naive Bayes network model to perform bearing fault diagnosis; The bearing fault diagnosis module fusing the dimension-time features defines the health condition of the bearing as the root node of the naive Bayes network model and defines the dimension-time features of the bearing vibration as the leaf node of the naive Bayes network model; The bearing fault diagnosis module fusing the dimension-time features adopts maximum likelihood estimation under complete data to train the conditional probability distribution parameters of the model; The fusion dimension-time feature bearing fault diagnosis module predicts the condition probability distribution of the health condition of the bearing at time t according to the naive Bayesian network model n The condition probability distribution of the health condition of the bearing at time t gives whether the bearing fails at time t, the damage type and severity of the bearing, and the corresponding uncertainty. n The condition probability distribution of the health condition of the bearing at time t gives whether the bearing fails at time t, the damage type and severity of the bearing, and the corresponding uncertainty.

Images