CN114861827B - Coal mining machine predictive diagnosis and health management method based on multi-source data fusion - Google Patents
Coal mining machine predictive diagnosis and health management method based on multi-source data fusion Download PDFInfo
- Publication number
- CN114861827B CN114861827B CN202210614486.2A CN202210614486A CN114861827B CN 114861827 B CN114861827 B CN 114861827B CN 202210614486 A CN202210614486 A CN 202210614486A CN 114861827 B CN114861827 B CN 114861827B
- Authority
- CN
- China
- Prior art keywords
- coal mining
- mining machine
- maintenance
- module
- health index
- 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
- 239000003245 coal Substances 0.000 title claims abstract description 170
- 230000036541 health Effects 0.000 title claims abstract description 157
- 238000005065 mining Methods 0.000 title claims abstract description 155
- 238000003745 diagnosis Methods 0.000 title claims abstract description 45
- 230000004927 fusion Effects 0.000 title claims abstract description 40
- 238000007726 management method Methods 0.000 title claims abstract description 19
- 238000012423 maintenance Methods 0.000 claims abstract description 105
- 230000002159 abnormal effect Effects 0.000 claims abstract description 74
- 238000012544 monitoring process Methods 0.000 claims abstract description 32
- 238000000034 method Methods 0.000 claims description 83
- 239000011159 matrix material Substances 0.000 claims description 73
- 230000008569 process Effects 0.000 claims description 65
- 238000006731 degradation reaction Methods 0.000 claims description 60
- 230000015556 catabolic process Effects 0.000 claims description 59
- 238000012549 training Methods 0.000 claims description 36
- 238000012795 verification Methods 0.000 claims description 33
- 238000007689 inspection Methods 0.000 claims description 24
- 239000013598 vector Substances 0.000 claims description 22
- 230000008376 long-term health Effects 0.000 claims description 18
- 230000000737 periodic effect Effects 0.000 claims description 14
- 238000012216 screening Methods 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000007781 pre-processing Methods 0.000 claims description 4
- 230000001419 dependent effect Effects 0.000 claims description 3
- 230000005856 abnormality Effects 0.000 claims description 2
- 238000007621 cluster analysis Methods 0.000 claims description 2
- 206010063385 Intellectualisation Diseases 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 11
- 238000012545 processing Methods 0.000 description 6
- 238000009826 distribution Methods 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000008439 repair process Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000010200 validation analysis Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000002817 coal dust Substances 0.000 description 1
- 239000003034 coal gas Substances 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000001035 drying Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000005461 lubrication Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000011435 rock Substances 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000002759 z-score normalization Methods 0.000 description 1
Classifications
-
- 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
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- E—FIXED CONSTRUCTIONS
- E21—EARTH OR ROCK DRILLING; MINING
- E21C—MINING OR QUARRYING
- E21C35/00—Details of, or accessories for, machines for slitting or completely freeing the mineral from the seam, not provided for in groups E21C25/00 - E21C33/00, E21C37/00 or E21C39/00
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/20—Administration of product repair or maintenance
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Human Resources & Organizations (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Mining & Mineral Resources (AREA)
- Operations Research (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- Quality & Reliability (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Mechanical Engineering (AREA)
- Economics (AREA)
- General Life Sciences & Earth Sciences (AREA)
- Geochemistry & Mineralogy (AREA)
- Geology (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a coal mining machine predictive diagnosis and health management method based on multi-source data fusion, and belongs to the field of coal mine intellectualization. The system comprises a multi-source data acquisition and fusion module, an abnormal working condition real-time monitoring module, a fault diagnosis module, a residual life prediction module and a maintenance decision suggestion module. The invention can timely and accurately evaluate and forecast the running health state of the coal mining machine, so that the coal mining machine is in the optimal working state, the reliability of the coal mining machine can be greatly improved, unnecessary shutdown is avoided, the maximum working capacity of the coal mining machine is ensured, and the invention has important significance for guaranteeing the running reliability and safety of a coal mining system.
Description
Technical Field
The invention relates to the technical field of coal mine intellectualization, in particular to a coal mining machine predictive diagnosis and health management method based on multi-source data fusion.
Background
The normal operation of the electric traction coal mining machine determines the production efficiency of the fully-mechanized coal mining face, however, the underground working environment is very bad, the coal mining machine is not only subjected to huge impact load from coal, rock and the like, but also is polluted by coal dust, gas and the like during operation, and each key part of the coal mining machine is inevitably worn and destroyed in the long-term operation process, so that the coal mining machine is in fault. Therefore, the research of the health management technology of the coal mining machine is carried out, the running health state of the coal mining machine is timely and accurately estimated and forecasted, the coal mining machine is scheduled in real time according to the running condition, the coal mining machine is in the optimal working state, the reliability of the coal mining machine can be greatly improved, unnecessary shutdown is avoided, and the maximum working capacity of the coal mining machine is ensured. The coal mining machine fault prediction and health management PHM (Prognostics Health Management) technology is an important technical guarantee for realizing a 'less-humanization' fully-mechanized coal mining working face, and has important significance for guaranteeing the operation reliability and safety of a coal mining system only by realizing reliable health monitoring of the coal mining machine and reducing casualties of coal mining operators of the fully-mechanized coal mining face during accidents.
Disclosure of Invention
In order to solve the technical problems, the invention provides a coal mining machine prediction diagnosis and health management method based on multi-source data fusion. The technical scheme adopted by the invention is as follows:
a coal mining machine predictive diagnosis and health management method based on multi-source data fusion comprises the following steps:
s1, a multi-source data acquisition and fusion module acquires and transmits a single health index feature matrix of a single operation process of a coal mining machine to an abnormal working condition real-time monitoring module, and acquires and transmits a long-time health index feature matrix of a plurality of operation processes to a residual life prediction module;
s2, the abnormal condition real-time monitoring module judges whether the coal mining machine is abnormal in the operation process according to the single health index feature matrix and the pre-trained abnormal condition monitoring model, and when the abnormal condition real-time monitoring module determines that the coal mining machine is abnormal in the operation process, the abnormal condition real-time monitoring module sends the health index feature matrix under the abnormal condition to the fault diagnosis module and sends an abnormal alarm and inspection information to the maintenance decision suggestion module;
s3, the fault diagnosis module performs fault positioning and fault type identification according to the health index feature matrix under the abnormal working condition, and sends targeted maintenance suggestions to the maintenance decision suggestion module;
s4, a residual life prediction module predicts the residual life of the coal cutter according to the long-term health index feature matrix and a pre-trained coal cutter residual life prediction model, and when the predicted residual life of the coal cutter is smaller than a preset threshold value, the single health index feature matrix of the current single operation process in the long-term health index feature matrix is sent to a fault diagnosis module so as to locate faults and identify fault types by the fault diagnosis module, and the predicted residual life of the coal cutter is sent to a maintenance decision suggestion module;
and S5, the maintenance decision suggestion module carries out maintenance of the coal mining machine according to the abnormal alarm and inspection information, the targeted maintenance suggestion and the predicted residual life prompt of the coal mining machine.
Optionally, when the multi-source data acquisition and fusion module in S1 acquires a single health index feature matrix of a single operation process of the coal mining machine, the multi-source data acquisition and fusion module includes:
s11, acquiring initial multi-source signals acquired in a single operation process of the coal mining machine by a multi-source data acquisition and fusion module;
s12, preprocessing an initial multi-source signal acquired in a single operation process by a multi-source data acquisition and fusion module to obtain a target multi-source signal in the single operation process;
s13, extracting a plurality of health index feature vectors in a target multi-source signal of a single operation process by a multi-source data acquisition and fusion module;
s14, the multi-source data acquisition and fusion module fuses the plurality of health index feature vectors to obtain a single health index feature matrix of the single operation process of the coal mining machine.
Optionally, the abnormal condition real-time monitoring model is a GSA-LSSVM model, and the determining, by the abnormal condition real-time monitoring module in S2, whether the coal mining machine is abnormal in the running process according to the single health index feature matrix and the pre-trained abnormal condition monitoring model includes:
the abnormal working condition real-time monitoring module normalizes the single health index feature matrix, inputs the normalized single health index feature matrix into a pre-trained GSA-LSSVM model, and judges whether the coal mining machine is abnormal or not in the running process according to the output result of the GSA-LSSVM model.
Optionally, when the fault diagnosis module in S3 performs fault location and fault type identification according to the health indicator feature matrix under the abnormal working condition, the fault diagnosis module includes:
the fault diagnosis module compares the health index feature matrix under the abnormal working condition with the pre-stored typical fault features, and positions faults and identifies fault types according to the comparison result.
Optionally, when the remaining life prediction module in S4 predicts the remaining life of the coal mining machine according to the long-term health index feature matrix and the pre-trained remaining life prediction model of the coal mining machine, the method includes:
the residual life prediction module inputs the long-term health index feature matrix into a pre-trained residual life prediction model, and predicts the residual life of the coal mining machine according to the output result of the residual life prediction model.
Optionally, before the remaining life prediction module inputs the long-term health index feature matrix into the pre-trained coal mining machine remaining life prediction model, the remaining life prediction module further includes:
s41, a remaining life prediction module acquires a health index feature set determined according to an initial multi-source signal acquired in a historical operation process of the coal mining machine;
s42, dividing the health index features in the health index feature set into a training set and a verification set;
s43, performing cluster analysis on target health index features in the training set and the verification set respectively to divide the health stage of the coal mining machine into three health stages, namely a steady operation stage, an initial degradation stage and an accelerated degradation stage;
s44, respectively calculating the identifiable indexes between all the health index features and the health stages in the training set and the verification set, and screening the health index features in the training set and the verification set according to the calculation results to obtain a representative health index feature matrix;
s45, calculating the residual life and the health state of the coal mining machine according to the health stage;
s46, fitting a training set degradation curve group and a verification set degradation curve group of the coal mining machine according to the representative health index feature matrixes and the health states of the training set and the verification set respectively;
s47, calculating the distance and similarity between each degradation curve in the degradation curve family of the training set and each degradation curve in the degradation curve family of the verification set, and screening a plurality of groups of degradation curves which are most similar to the degradation curves in the degradation curve family of the verification set from the degradation curve family of the training set according to the distance and the similarity to serve as a residual life prediction set;
s48, obtaining the residual life of each degradation curve in the residual life prediction set, and giving weight to the residual life of each degradation curve in the residual life prediction set according to the similarity to obtain a coal mining machine residual life prediction model.
Optionally, when the maintenance decision suggestion module performs maintenance on the coal mining machine according to the abnormality alarm and the inspection information, the targeted maintenance suggestion and the predicted remaining life prompt of the coal mining machine in S5, the maintenance decision suggestion module includes:
the maintenance decision suggestion module performs periodic maintenance or on-line maintenance of the coal mining machine according to the abnormal alarm and inspection information, the targeted maintenance suggestion and the predicted remaining life prompt of the coal mining machine, wherein the on-line maintenance comprises manual inspection, conventional maintenance and fault maintenance.
Optionally, the maintenance decision suggestion module performs periodic maintenance or optionally repair of the coal mining machine according to the abnormal alarm and inspection information, the targeted repair suggestion and the predicted remaining life prompt of the coal mining machine, including:
setting a time interval of periodic maintenance, and prompting the periodic maintenance of the coal mining machine when the condition-dependent maintenance does not occur in the time interval of the periodic maintenance;
when the on-condition maintenance occurs in the time interval of the regular maintenance, prompting the on-condition maintenance of the coal mining machine, and updating the time interval of the regular maintenance.
Optionally, the prompting the coal mining machine to be maintained optionally includes:
sending out an abnormal alarm signal according to the abnormal alarm and the inspection information to prompt for manual inspection;
performing fault maintenance according to the targeted maintenance suggestion prompt;
and carrying out conventional maintenance according to the predicted residual life prompt of the coal mining machine.
The beneficial effects of the invention are as follows:
the working state of the coal mining machine can be monitored in real time through the abnormal working condition real-time monitoring module, faults can be located and the fault type can be determined in time through the fault diagnosis module, the residual life of the coal mining machine can be predicted in time through the residual life prediction module, and the maintenance or maintenance strategy for the coal mining machine can be provided in time through the maintenance decision suggestion module.
Drawings
Fig. 1 is a flow chart of the present invention.
FIG. 2 is a schematic diagram of the operation of the multi-source data acquisition and fusion module of the present invention.
FIG. 3 is a schematic diagram of the working process of the real-time monitoring module for abnormal working conditions in the invention.
Fig. 4 is a schematic diagram illustrating the operation of the fault diagnosis module according to the present invention.
FIG. 5 is a schematic diagram illustrating the operation of the residual life prediction module of the present invention.
FIG. 6 is a schematic diagram of the operation of the maintenance decision suggestion module of the present invention.
Fig. 7 is a schematic diagram of the overall operation of the modules of the present invention.
Detailed Description
The invention will be described in further detail below with reference to the drawings and examples.
As shown in fig. 1, the method for predicting diagnosis and health management of a coal mining machine based on multi-source data fusion in this embodiment includes the following steps:
s1, a multi-source data acquisition and fusion module acquires and transmits a single health index feature matrix of a single operation process of the coal mining machine to an abnormal working condition real-time monitoring module, and acquires and transmits a long-time health index feature matrix of a plurality of operation processes to a residual life prediction module.
Optionally, when the multi-source data acquisition and fusion module in S1 acquires a single health index feature matrix of a single operation process of the coal mining machine, the method includes the following steps:
s11, acquiring initial multi-source signals acquired in a single operation process of the coal mining machine by a multi-source data acquisition and fusion module.
The initial multisource signal includes an initial vibration signal, an initial temperature signal, and an initial current signal. The initial multi-source signal can be acquired from vibration sensors, temperature sensors and ampere meters configured on the shearer.
By acquiring initial multi-source signals acquired in the single operation process of the coal mining machine, state data acquired in the single operation process of the coal mining machine can be acquired, multiple groups of state data capable of indicating the operation state of the coal mining machine are obtained, and the accuracy of the method for carrying out predictive diagnosis and health management on the coal mining machine according to the state data is high.
S12, the multi-source data acquisition and fusion module preprocesses initial multi-source signals acquired in a single operation process to obtain target multi-source signals in the single operation process.
On the basis of the initial multi-source signal types, the step S12 is to preprocess the initial multi-source signal collected in the single operation process to obtain the target multi-source signal in the single operation process, and includes the following steps:
s121, performing de-drying filtering processing on the initial vibration signals acquired in the single operation process to obtain target vibration signals in the single operation process.
S122, performing outlier processing on the initial temperature signal and the initial current signal acquired in the single operation process to obtain a target temperature signal and a target current signal in the single operation process.
By preprocessing, the initial multi-source signals which obviously have no research value can be removed, so that the data volume of the initial multi-source signals can be reduced, invalid calculation is reduced, and the calculation efficiency is improved.
S13, the multi-source data acquisition and fusion module extracts a plurality of health index feature vectors in a target multi-source signal of a single operation process.
On the basis of S11 and S12, the step S13 includes the following steps when extracting the health index feature vector in the target multi-source signal of the single operation process:
s131, extracting a time domain feature vector, a frequency domain feature vector and a time-frequency domain feature vector of a target vibration signal as health index feature vectors of the target vibration signal, wherein the time domain feature vector comprises vibration variance, skewness, peak value factor, kurtosis, peak value, kurtosis factor, average value, waveform factor, skewness factor, standard deviation and root mean square, the frequency domain feature vector comprises amplitude average value, frequency variance, center of gravity frequency, root mean square frequency, mean square frequency and frequency amplitude variance, and the time-frequency domain feature vector comprises Hilbert Huang Putu interval mean variance.
The frequency domain feature vector can be obtained by performing fourier transform on the target vibration signal when extracting the frequency domain feature vector. The frequency domain feature vector can be obtained by performing time-frequency joint processing based on Empirical Mode Decomposition (EMD) and Hilbert-Huang transform (HHT) on the target vibration signal.
S132, extracting a temperature maximum value, a temperature minimum value, a temperature mean value and a temperature variance as health index feature vectors of the target temperature signal.
S133, extracting a current maximum value, a current minimum value, a current mean value, a current variance and a current effective value as health index feature vectors of the target current signal.
S14, the multi-source data acquisition and fusion module fuses the plurality of health index feature vectors to obtain a single health index feature matrix of the single operation process of the coal mining machine.
Fig. 2 is a schematic diagram of the working process of the multi-source data acquisition and fusion module. Each column in the single health indicator feature matrix shown in fig. 2 includes the acquisition time (ti) of the target multi-source signal and the feature vector (xij) in the target multi-source signal.
Further, when the multi-source data acquisition and fusion module acquires the long-term health index feature matrix of the multi-time operation process, the single health index feature matrix of each time operation process is combined according to the time sequence.
S2, the abnormal working condition real-time monitoring module judges whether the coal mining machine is abnormal in the operation process according to the single health index feature matrix and the pre-trained abnormal working condition monitoring model, and when the abnormal working condition real-time monitoring module determines that the coal mining machine is abnormal in the operation process, the abnormal working condition real-time monitoring module sends the health index feature matrix under the abnormal working condition to the fault diagnosis module, and sends the abnormal alarm and the inspection information to the maintenance decision suggestion module.
Optionally, the abnormal working condition real-time monitoring model is a GSA-LSSVM model. As shown in fig. 3, the abnormal condition real-time monitoring module in S2 may be implemented by the following manner when determining whether the coal mining machine is abnormal in the operation process according to the single health index feature matrix and the pre-trained abnormal condition monitoring model:
the abnormal working condition real-time monitoring module normalizes the single health index feature matrix, inputs the normalized single health index feature matrix into a pre-trained GSA-LSSVM model, and judges whether the coal mining machine is abnormal or not in the running process according to the output result of the GSA-LSSVM model. Wherein, when the single health index feature matrix is normalized, a Z-Score normalization processing algorithm can be adopted. The GSA-LSSVM model is obtained by training an unsupervised learning algorithm based on a single health index feature matrix of a historical operation process.
S3, the fault diagnosis module performs fault positioning and fault type identification according to the health index feature matrix under the abnormal working condition, and sends the targeted maintenance suggestion to the maintenance decision suggestion module.
Optionally, when the fault diagnosis module in S3 performs fault location and fault type identification according to the health indicator feature matrix under the abnormal working condition, the following manner may be implemented:
the fault diagnosis module compares the health index feature matrix under the abnormal working condition with the pre-stored typical fault features, and positions faults and identifies fault types according to the comparison result. The fault diagnosis module collects and stores typical fault characteristics corresponding to various fault types provided by mature part manufacturers, monitoring software manufacturers, fault diagnosis specialists of universities and the like in advance.
And when the health index feature matrix under the abnormal working condition is successfully compared with any typical fault feature, determining the typical fault as the current fault type of the coal mining machine. Whether the health index feature matrix is successfully compared with any typical fault feature can be determined by calculating the similarity between the health index feature matrix and any typical fault feature.
S4, the residual life prediction module predicts the residual life of the coal cutter according to the long-term health index feature matrix and a pre-trained coal cutter residual life prediction model, and when the predicted residual life of the coal cutter is smaller than a preset threshold value, the single health index feature matrix of the current single operation process in the long-term health index feature matrix is sent to the fault diagnosis module so as to locate faults and identify fault types by the fault diagnosis module, and the predicted residual life of the coal cutter is sent to the maintenance decision suggestion module.
Wherein the preset threshold is a certain value tending to 0. The single health index feature matrix of the current single operation process in the long-term health index feature matrix refers to a single health index feature matrix acquired in the time closest to the current moment in the long-term health index feature matrix.
When the predicted remaining life of the coal mining machine is smaller than a preset threshold value, the coal mining machine is indicated to have a larger fault, so that the embodiment of the invention sends the single health index feature matrix of the current single operation process to the fault diagnosis module so as to locate the fault and determine the fault type in time.
As shown in fig. 4, a schematic diagram of the operation of the fault diagnosis module is shown. As shown in fig. 5, a schematic diagram of the operation of the remaining life prediction module is shown.
Optionally, when the remaining life prediction module in S4 predicts the remaining life of the coal mining machine according to the long-term health index feature matrix and the pre-trained remaining life prediction model of the coal mining machine, the following manner may be implemented:
the residual life prediction module inputs the long-term health index feature matrix into a pre-trained residual life prediction model, and predicts the residual life of the coal mining machine according to the output result of the residual life prediction model.
Optionally, before the remaining life prediction module inputs the long-term health index feature matrix into the pre-trained remaining life prediction model of the coal mining machine, the remaining life prediction model of the coal mining machine needs to be trained, and the following description will describe the training process of the remaining life prediction model of the coal mining machine. Specifically, when training the residual life prediction model of the coal mining machine, the method comprises the following steps:
s41, a remaining life prediction module acquires a health index feature set determined according to an initial multi-source signal acquired in a historical operation process of the coal mining machine.
Wherein, the step S41, when acquiring the health index feature set determined according to the initial multi-source signal acquired during the historical operation of the coal mining machine, includes:
s411, acquiring initial multi-source signals acquired respectively in each historical operation process of the coal mining machine.
S412, preprocessing the initial multi-source signals collected by each historical operation process to obtain target multi-source signals of each historical operation process.
S413, extracting the health index feature vector in each target multi-source signal, and combining the health index feature vectors in each target multi-source signal to obtain the health index feature set of the coal mining machine.
Specifically, the principles of steps S411 to S413 are the same as those of steps S11 to S14, and the details of steps S11 to S14 are specifically referred to, and are not described herein.
S42, dividing the health index features in the health index feature set into a training set and a verification set.
Wherein, the step S42 includes the following steps when dividing the health index features in the health index feature set into a training set and a verification set:
s421, performing Z-Score standardization processing on the health index features in the health index feature set so as to uniformly measure the Z-Score values uniformly calculated by the health index features with different magnitudes.
S422, dividing the standardized health index feature set into a training set and a verification set by adopting a k-fold cross verification method.
S43, clustering analysis is carried out on the target health index characteristics in the training set and the verification set respectively, so that the health stage of the coal mining machine is divided into three health stages, namely a stable operation stage, an initial degradation stage and an accelerated degradation stage.
Wherein the target health indicator characteristic includes, but is not limited to, root mean square. Simulation experiments show that the target health index feature is selected to be root mean square, so that the divided health stages can be accurate.
S44, respectively calculating the identifiable indexes between all the health index features and the health stages in the training set and the verification set, and screening the health index features in the training set and the verification set according to the calculation results to obtain a representative health index feature matrix.
Specifically, the step S44 may be implemented by the following formula when calculating the recognizability index between all the health index features and the health phases in the training set and the verification set, respectively:
wherein:in order to be an inter-class scatter matrix,is an intra-class scatter matrix, n s Is the sample number of the s class, m s For the mean of the health index features of class s, m is the mean of all the health index features, Ω represents the health index feature set of class s.
Further, when the health index features in the training set and the verification set are screened according to the calculation result, the calculated identifiable indexes can be ranked, and a plurality of health index features ranked at the tail end can be removed according to the requirement. For example, when the current maximum, current minimum, temperature maximum, temperature minimum, and the recognizability index between the health phases are ranked behind the recognizability index of other health index features, it is demonstrated that these health value index features have less impact on the health phases, which can be filtered out of the training set and validation set to reduce the training set and validation set, reducing ineffective computations.
S45, calculating the remaining service life and the health state of the coal mining machine according to the health stage.
Wherein, the step S45 can be realized by the following formula when calculating the remaining life and health state of the coal mining machine according to the health stage:
wherein:maximum service life->To accelerate the initial remaining life of the degradation phase; epsilon (t) is random noise and accords with normal distribution with the mean value of 0; θ (t) is the slope of the initial degradation phase, conforming to a normal distribution with average value θ; θ 0 (t) conforms to a lognormal distribution with average value θ, and β (t) conforms to a normal distribution with average value β, σ 2 Is the noise variance; hs=0, 1,2 represent the healthy phase as a stationary operation phase, an initial degradation phase and an accelerated degradation phase, RUL as a remaining life, HD as a healthy state, respectively.
S46, fitting a training set degradation curve group and a verification set degradation curve group of the coal mining machine according to the representative health index feature matrixes and the health states of the training set and the verification set respectively.
As can be seen from the formula (4), the health state of the coal mining machine is always 1 in the steady operation stage, then gradually decays to 0 along with the degradation of the coal mining machine, the health state HD is taken as output, the representative health index feature matrix of the training set and the verification set obtained by screening in S54 is taken as input, the input and the output are fitted in segments according to the health stage, and smoothing processing is performed to obtain the training set degradation curve family and the verification set degradation curve family of the coal mining machine.
S47, calculating the distance and the similarity between each degradation curve in the degradation curve family of the training set and each degradation curve in the degradation curve family of the verification set, and screening a plurality of groups of degradation curves which are most similar to the degradation curves in the degradation curve family of the verification set from the degradation curve family of the training set according to the distance and the similarity to serve as a residual life prediction set.
The distance can be calculated by adopting a Euclidean distance equidistant calculation formula. The similarity is calculated using a similarity algorithm. The number of degradation curves in the remaining life prediction set may be selected as desired. The degradation curves in the residual life prediction set are a plurality of degradation curves with minimum distance and highest similarity between the training set and the verification set degradation curve family.
S48, obtaining the residual life of each degradation curve in the residual life prediction set, and giving weight to the residual life of each degradation curve in the residual life prediction set according to the similarity to obtain a coal mining machine residual life prediction model.
The remaining life of each degradation curve in the remaining life prediction set may be obtained from the degradation curve. The sum of the weights of all degradation curves in the remaining life prediction set is 1, and the higher the similarity is, the higher the weight is for the degradation curve.
According to the embodiment of the invention, a degradation curve family of the coal mining machine is constructed according to the initial multi-source data of the running process of the coal mining machines with the same type, and the remaining service life of the coal mining machine is predicted by calculating the distance and similarity screening verification set and giving weight. The segmented degradation model of the coal mining machine is provided according to theoretical analysis and field practical experience, the degradation process of the coal mining machine is scientifically and reasonably summarized, and a new thought is provided for predicting the residual life and managing the health of the coal mining machine.
And S5, the maintenance decision suggestion module carries out maintenance of the coal mining machine according to the abnormal alarm and inspection information, the targeted maintenance suggestion and the predicted residual life prompt of the coal mining machine.
Alternatively, as shown in fig. 6, it is a schematic diagram of the operation of the maintenance decision suggestion module. The maintenance decision suggestion module in S5 performs maintenance of the coal mining machine according to the abnormal alarm and inspection information, the targeted maintenance suggestion and the predicted remaining life prompt of the coal mining machine, and includes: and carrying out periodical maintenance or optionally maintenance of the coal mining machine according to the abnormal alarm and inspection information, the targeted maintenance advice and the predicted residual life prompt of the coal mining machine, wherein the optionally maintenance comprises manual inspection, conventional maintenance and fault maintenance. The method specifically comprises the steps of carrying out balance checking on a rotor, repairing or replacing a bearing, enhancing the rigidity of a foundation, repairing or replacing a gear, re-checking and centering, fastening bolts, increasing lubrication, maintaining a gearbox and other accurate maintenance strategies.
Further, the maintenance decision suggestion module performs periodic maintenance or optionally maintenance of the coal mining machine according to the abnormal alarm and inspection information, the targeted maintenance suggestion and the predicted remaining life prompt of the coal mining machine, including: setting a time interval of periodic maintenance, and prompting the periodic maintenance of the coal mining machine when the condition-dependent maintenance does not occur in the time interval of the periodic maintenance; when the on-condition maintenance occurs in the time interval of the regular maintenance, prompting the on-condition maintenance of the coal mining machine, and updating the time interval of the regular maintenance.
Optionally, the prompting the coal mining machine to be maintained optionally includes: sending out an abnormal alarm signal according to the abnormal alarm and the inspection information to prompt for manual inspection; performing fault maintenance according to the targeted maintenance suggestion prompt; and carrying out conventional maintenance according to the predicted residual life prompt of the coal mining machine.
In summary, as shown in fig. 7, an overall working process schematic diagram of a coal mining machine predictive diagnosis and health management method based on multi-source data fusion according to an embodiment of the present invention is shown. According to the embodiment of the invention, the working state of the coal mining machine can be monitored in real time through the abnormal working condition real-time monitoring module, the fault can be timely positioned and the fault type can be determined through the fault diagnosis module, the residual life of the coal mining machine can be timely predicted through the residual life prediction module, and the maintenance or maintenance strategy for the coal mining machine can be timely provided through the maintenance decision suggestion module, so that the working health state of the coal mining machine can be timely and accurately estimated and predicted, the coal mining machine is in the optimal working state, the reliability of the coal mining machine can be greatly improved, unnecessary shutdown is avoided, the maximum working capacity of the coal mining machine is ensured, and the method has important significance for guaranteeing the operation reliability and safety of a coal mining system.
It is to be understood that the above embodiments are merely illustrative of the application of the principles of the present invention, but not in limitation thereof. Various modifications and improvements may be made by those skilled in the art without departing from the spirit and substance of the invention, and are also considered to be within the scope of the invention.
Claims (9)
1. The coal mining machine predictive diagnosis and health management method based on multi-source data fusion is characterized by comprising the following steps of:
s1, a multi-source data acquisition and fusion module acquires and transmits a single health index feature matrix of a single operation process of a coal mining machine to an abnormal working condition real-time monitoring module, and acquires and transmits a long-time health index feature matrix of a plurality of operation processes to a residual life prediction module;
s2, the abnormal condition real-time monitoring module judges whether the coal mining machine is abnormal in the operation process according to the single health index feature matrix and the pre-trained abnormal condition monitoring model, and when the abnormal condition real-time monitoring module determines that the coal mining machine is abnormal in the operation process, the abnormal condition real-time monitoring module sends the health index feature matrix under the abnormal condition to the fault diagnosis module and sends an abnormal alarm and inspection information to the maintenance decision suggestion module;
s3, the fault diagnosis module performs fault positioning and fault type identification according to the health index feature matrix under the abnormal working condition, and sends targeted maintenance suggestions to the maintenance decision suggestion module;
s4, a residual life prediction module predicts the residual life of the coal cutter according to the long-term health index feature matrix and a pre-trained coal cutter residual life prediction model, and when the predicted residual life of the coal cutter is smaller than a preset threshold value, the single health index feature matrix of the current single operation process in the long-term health index feature matrix is sent to a fault diagnosis module so as to locate faults and identify fault types by the fault diagnosis module, and the predicted residual life of the coal cutter is sent to a maintenance decision suggestion module;
and S5, the maintenance decision suggestion module carries out maintenance of the coal mining machine according to the abnormal alarm and inspection information, the targeted maintenance suggestion and the predicted residual life prompt of the coal mining machine.
2. The method for predicting diagnosis and health management of a coal mining machine based on multi-source data fusion according to claim 1, wherein the multi-source data acquisition and fusion module in S1, when acquiring a single health index feature matrix of a single operation process of the coal mining machine, comprises:
s11, acquiring initial multi-source signals acquired in a single operation process of the coal mining machine by a multi-source data acquisition and fusion module;
s12, preprocessing an initial multi-source signal acquired in a single operation process by a multi-source data acquisition and fusion module to obtain a target multi-source signal in the single operation process;
s13, extracting a plurality of health index feature vectors in a target multi-source signal of a single operation process by a multi-source data acquisition and fusion module;
s14, the multi-source data acquisition and fusion module fuses the plurality of health index feature vectors to obtain a single health index feature matrix of the single operation process of the coal mining machine.
3. The method for predicting diagnosis and health management of a coal mining machine based on multi-source data fusion according to claim 1, wherein the abnormal condition real-time monitoring model is a GSA-LSSVM model, and the step S2 of determining whether the coal mining machine is abnormal in the operation process by the abnormal condition real-time monitoring module according to the single health index feature matrix and the pre-trained abnormal condition monitoring model comprises the following steps:
the abnormal working condition real-time monitoring module normalizes the single health index feature matrix, inputs the normalized single health index feature matrix into a pre-trained GSA-LSSVM model, and judges whether the coal mining machine is abnormal or not in the running process according to the output result of the GSA-LSSVM model.
4. The method for predicting diagnosis and health management of coal mining machine based on multi-source data fusion according to claim 1, wherein the step S3 of the fault diagnosis module performing fault location and fault type identification according to the health index feature matrix under abnormal working conditions comprises:
the fault diagnosis module compares the health index feature matrix under the abnormal working condition with the pre-stored typical fault features, and positions faults and identifies fault types according to the comparison result.
5. The method for predicting diagnosis and health management of a coal mining machine based on multi-source data fusion according to claim 1, wherein when the residual life prediction module predicts the residual life of the coal mining machine according to the long-term health index feature matrix and the pre-trained residual life prediction model of the coal mining machine in S4, the method comprises:
the residual life prediction module inputs the long-term health index feature matrix into a pre-trained residual life prediction model, and predicts the residual life of the coal mining machine according to the output result of the residual life prediction model.
6. The method for predicting diagnosis and health of a coal mining machine based on multi-source data fusion according to claim 5, wherein before the residual life prediction module inputs the long-term health index feature matrix into the pre-trained coal mining machine residual life prediction model, the method further comprises:
s41, a remaining life prediction module acquires a health index feature set determined according to an initial multi-source signal acquired in a historical operation process of the coal mining machine;
s42, dividing the health index features in the health index feature set into a training set and a verification set;
s43, performing cluster analysis on target health index features in the training set and the verification set respectively to divide the health stage of the coal mining machine into three health stages, namely a steady operation stage, an initial degradation stage and an accelerated degradation stage;
s44, respectively calculating the identifiable indexes between all the health index features and the health stages in the training set and the verification set, and screening the health index features in the training set and the verification set according to the calculation results to obtain a representative health index feature matrix;
s45, calculating the residual life and the health state of the coal mining machine according to the health stage;
s46, fitting a training set degradation curve group and a verification set degradation curve group of the coal mining machine according to the representative health index feature matrixes and the health states of the training set and the verification set respectively;
s47, calculating the distance and similarity between each degradation curve in the degradation curve family of the training set and each degradation curve in the degradation curve family of the verification set, and screening a plurality of groups of degradation curves which are most similar to the degradation curves in the degradation curve family of the verification set from the degradation curve family of the training set according to the distance and the similarity to serve as a residual life prediction set;
s48, obtaining the residual life of each degradation curve in the residual life prediction set, and giving weight to the residual life of each degradation curve in the residual life prediction set according to the similarity to obtain a coal mining machine residual life prediction model.
7. The method for predicting diagnosis and health management of a coal mining machine based on multi-source data fusion according to claim 1, wherein the maintenance decision suggestion module in S5 performs maintenance of the coal mining machine according to the abnormality alarm and inspection information, the targeted maintenance suggestion and the predicted remaining life prompt of the coal mining machine, and comprises:
the maintenance decision suggestion module performs periodic maintenance or on-line maintenance of the coal mining machine according to the abnormal alarm and inspection information, the targeted maintenance suggestion and the predicted remaining life prompt of the coal mining machine, wherein the on-line maintenance comprises manual inspection, conventional maintenance and fault maintenance.
8. The method for predicting diagnosis and health of a shearer based on multi-source data fusion of claim 7, wherein the maintenance decision suggestion module performs periodic maintenance or on-demand maintenance of the shearer according to the anomaly alarm and inspection information, the targeted maintenance suggestions and the predicted shearer remaining life cues, comprising:
setting a time interval of periodic maintenance, and prompting the periodic maintenance of the coal mining machine when the condition-dependent maintenance does not occur in the time interval of the periodic maintenance;
when the on-condition maintenance occurs in the time interval of the regular maintenance, prompting the on-condition maintenance of the coal mining machine, and updating the time interval of the regular maintenance.
9. The method for predictive diagnosis and health management of a shearer based on multi-source data fusion of claim 8, wherein said prompting an on-line maintenance of the shearer comprises:
sending out an abnormal alarm signal according to the abnormal alarm and the inspection information to prompt for manual inspection;
performing fault maintenance according to the targeted maintenance suggestion prompt;
and carrying out conventional maintenance according to the predicted residual life prompt of the coal mining machine.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210614486.2A CN114861827B (en) | 2022-05-27 | 2022-05-27 | Coal mining machine predictive diagnosis and health management method based on multi-source data fusion |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210614486.2A CN114861827B (en) | 2022-05-27 | 2022-05-27 | Coal mining machine predictive diagnosis and health management method based on multi-source data fusion |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114861827A CN114861827A (en) | 2022-08-05 |
| CN114861827B true CN114861827B (en) | 2024-03-26 |
Family
ID=82640359
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210614486.2A Active CN114861827B (en) | 2022-05-27 | 2022-05-27 | Coal mining machine predictive diagnosis and health management method based on multi-source data fusion |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114861827B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116825243B (en) * | 2023-05-09 | 2024-01-16 | 安徽工程大学 | Multi-source data-based thermal barrier coating service life prediction method and system |
| CN117572295B (en) * | 2024-01-12 | 2024-04-12 | 山东和兑智能科技有限公司 | Multi-mode on-line monitoring and early warning method for high-voltage sleeve |
| CN117668498B (en) * | 2024-01-31 | 2024-04-26 | 和尘自仪(嘉兴)科技有限公司 | Pump health assessment method based on reliability distribution and anomaly detection |
| CN117710759B (en) * | 2024-02-06 | 2024-04-26 | 青岛哈尔滨工程大学创新发展中心 | System and method for predicting faults and service lives of manned submersible vehicles |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104153816A (en) * | 2014-09-04 | 2014-11-19 | 中国矿业大学(北京) | Coal cutter state monitoring method based on black box |
| KR101936283B1 (en) * | 2017-12-28 | 2019-01-08 | 주식회사 한국가스기술공사 | Diagnostic and prognostics method for machine fault |
| CN113093693A (en) * | 2021-03-18 | 2021-07-09 | 北京天地玛珂电液控制系统有限公司 | Online fault diagnosis method for operation state of coal mining machine |
| CN114282434A (en) * | 2021-12-16 | 2022-04-05 | 成都航天科工大数据研究院有限公司 | Industrial equipment health management system and method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113298278B (en) * | 2020-02-19 | 2023-12-05 | 硕天科技股份有限公司 | Power equipment with self-health state prediction function, self-health state prediction method thereof and cloud server |
-
2022
- 2022-05-27 CN CN202210614486.2A patent/CN114861827B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104153816A (en) * | 2014-09-04 | 2014-11-19 | 中国矿业大学(北京) | Coal cutter state monitoring method based on black box |
| KR101936283B1 (en) * | 2017-12-28 | 2019-01-08 | 주식회사 한국가스기술공사 | Diagnostic and prognostics method for machine fault |
| CN113093693A (en) * | 2021-03-18 | 2021-07-09 | 北京天地玛珂电液控制系统有限公司 | Online fault diagnosis method for operation state of coal mining machine |
| CN114282434A (en) * | 2021-12-16 | 2022-04-05 | 成都航天科工大数据研究院有限公司 | Industrial equipment health management system and method |
Non-Patent Citations (2)
| Title |
|---|
| 数控设备故障预测和健康管理的维修保障系统;赵中敏;;林业机械与木工设备;20130410(第04期);全文 * |
| 采煤机在线监测与故障诊断系统研究;桑建伟;刘赟;郜彤;;山西煤炭;20200615(第02期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114861827A (en) | 2022-08-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114861827B (en) | Coal mining machine predictive diagnosis and health management method based on multi-source data fusion | |
| CN112784373B (en) | Fault early warning method for wind turbine generator gearbox | |
| CN111325410B (en) | Universal fault early warning system based on sample distribution and early warning method thereof | |
| CN117074965B (en) | Lithium ion battery remaining life prediction method and system | |
| CN110865924A (en) | Health degree diagnosis method and health diagnosis framework for internal server of power information system | |
| CN116910680B (en) | Remote fault detection method and system for fitness equipment | |
| CN117331790A (en) | Machine room fault detection method and device for data center | |
| KR102204199B1 (en) | Prediction method and diagonoisis method of vehicle for autonomous public bus | |
| US11339763B2 (en) | Method for windmill farm monitoring | |
| CN116562437A (en) | Track circuit compensation capacitor fault prediction method and device | |
| CN113955149B (en) | Health diagnosis method and device for motor system | |
| CN113435228A (en) | Motor bearing service life prediction and analysis method based on vibration signal modeling | |
| CN110796282A (en) | Industry fluctuation period monitoring method based on correlation coefficient | |
| CN115358533A (en) | Predictive maintenance management system based on big data | |
| Trstenjak et al. | A Decision Support System for the Prediction of Wastewater Pumping Station Failures Based on CBR Continuous Learning Model. | |
| CN115326393A (en) | Wind turbine generator bearing pair fault diagnosis method based on temperature information | |
| CN114046228B (en) | Wind turbine generator abnormality diagnosis method and system | |
| CN117521084B (en) | Active safety early warning method for complex system | |
| CN117808456B (en) | Equipment fault early warning method and device based on intelligent operation management | |
| CN117514653A (en) | Wind turbine generator system fault early warning method and device based on abnormal deviation detection | |
| CN117521084A (en) | Active safety early warning method for complex system | |
| CN117110741A (en) | Equipment fault diagnosis method, device, equipment and storage medium | |
| CN117807400A (en) | Wind power plant energy efficiency analysis method, device, equipment and readable storage medium | |
| CN114548186A (en) | Fault diagnosis method for rotor of gearbox of air compressor high-speed machine | |
| CN118030409A (en) | Method and system for detecting abnormal operation performance of fan unit |
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 |