CN113672859B - Fault acoustic diagnosis system for switch machine - Google Patents
Fault acoustic diagnosis system for switch machine Download PDFInfo
- Publication number
- CN113672859B CN113672859B CN202110944630.4A CN202110944630A CN113672859B CN 113672859 B CN113672859 B CN 113672859B CN 202110944630 A CN202110944630 A CN 202110944630A CN 113672859 B CN113672859 B CN 113672859B
- Authority
- CN
- China
- Prior art keywords
- module
- environmental
- signal
- acquisition time
- information
- 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
- 238000003745 diagnosis Methods 0.000 title abstract description 8
- 238000001514 detection method Methods 0.000 claims abstract description 88
- 230000007613 environmental effect Effects 0.000 claims abstract description 79
- 238000007781 pre-processing Methods 0.000 claims abstract description 66
- 238000007405 data analysis Methods 0.000 claims abstract description 44
- 238000004458 analytical method Methods 0.000 claims abstract description 43
- 238000011156 evaluation Methods 0.000 claims abstract description 31
- 239000013598 vector Substances 0.000 claims description 33
- 238000000034 method Methods 0.000 claims description 30
- 238000012545 processing Methods 0.000 claims description 20
- 238000012423 maintenance Methods 0.000 claims description 12
- 238000003672 processing method Methods 0.000 claims description 12
- 239000011159 matrix material Substances 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 7
- 239000000284 extract Substances 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims description 3
- 238000007689 inspection Methods 0.000 claims description 3
- 238000010801 machine learning Methods 0.000 claims description 3
- 230000002159 abnormal effect Effects 0.000 description 4
- 238000000513 principal component analysis Methods 0.000 description 3
- 238000011946 reduction process Methods 0.000 description 3
- 230000008030 elimination Effects 0.000 description 2
- 238000003379 elimination reaction Methods 0.000 description 2
- 230000005856 abnormality Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000007477 logistic regression Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B61—RAILWAYS
- B61K—AUXILIARY EQUIPMENT SPECIALLY ADAPTED FOR RAILWAYS, NOT OTHERWISE PROVIDED FOR
- B61K9/00—Railway vehicle profile gauges; Detecting or indicating overheating of components; Apparatus on locomotives or cars to indicate bad track sections; General design of track recording vehicles
- B61K9/08—Measuring installations for surveying permanent way
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B61—RAILWAYS
- B61L—GUIDING RAILWAY TRAFFIC; ENSURING THE SAFETY OF RAILWAY TRAFFIC
- B61L23/00—Control, warning, or like safety means along the route or between vehicles or vehicle trains
- B61L23/04—Control, warning, or like safety means along the route or between vehicles or vehicle trains for monitoring the mechanical state of the route
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
Abstract
The invention discloses a fault acoustic diagnosis system of a switch machine, which comprises a signal acquisition module, a storage module, a data analysis module, a signal preprocessing module, an acoustic detection module and an evaluation module, wherein the signal acquisition module is used for acquiring detection data information and storing the detection data information in the storage module, the evaluation module obtains main influence factors through analysis of the existing environmental information and detection results of faults, the signal preprocessing module is used for preprocessing to obtain a preprocessing result, the data analysis module is used for analyzing the preprocessing result of the signal preprocessing module to obtain noise influence parameters, the environmental information in an acquisition time period is used for analyzing to obtain environmental influence parameters, finally, a threshold value is obtained through the preprocessing result, the noise influence parameters and the environmental influence parameters, a feature subset of acoustic signals is selected through the threshold value, and the accuracy of acoustic signal detection is improved through selection of the threshold value on the features of the acoustic signals.
Description
Technical Field
The invention relates to the technical field of acoustic diagnosis, in particular to a fault acoustic diagnosis system of a switch machine.
Background
The turnout switch machine is used for converting different running directions of a train, the turnout switch machine, the signal machine and the track circuit are the most critical outdoor three pieces in a railway signal system, and as the turnout switch machine is arranged outdoors for a long time, the working state of the turnout switch machine is easily influenced by factors such as external environment (foreign matters clamped between pointed base rails and the like) and extreme weather (rain, snow, icing and the like), and the problem of the fault of the turnout switch machine is one of main reasons for influencing the running efficiency and the running safety of the train; when the mechanical fault occurs in the switch machine, the physical structure of the switch machine is changed, from the acoustic perspective, on the premise that the acquisition equipment is unchanged, the characteristic of an acoustic signal is determined by a signal source and a propagation medium, and the change of the physical structure of the switch machine can cause the change of the signal source or the propagation medium, so that the characteristic of the acoustic signal is changed, and as long as the acoustic signal characteristic points which change along with the change of the physical structure can be found and detected, whether the switch machine operates normally or fails can be judged; however, the problems of incomplete noise elimination, excessive noise elimination, inaccurate selection of the characteristic subsets of the acoustic signals, various fault types and the like exist in the process of processing the acoustic signals, so that the acoustic signals of a normal switch machine and the acoustic signals of an abnormal switch machine cannot be accurately distinguished, mechanical faults of the switch machine can be timely found, measures can be timely taken to reduce the running risk of a train and improve the running efficiency of the train in order to improve the accuracy of acoustic detection signals of the switch machine, and the system provides an acoustic detection and diagnosis system for the faults of the switch machine.
Disclosure of Invention
In view of the above situation, in order to overcome the defects of the prior art, the present invention aims to provide a fault acoustic diagnosis system for a switch machine, wherein an evaluation module compares the characteristics of an acoustic signal with the characteristics of an acoustic signal when the acoustic signal is collected and the acoustic signal is detected by the characteristics of the acoustic signal when the acoustic signal is in a fault type, a data analysis module combines the noise reduction process and the detection process of the acoustic signal with the analysis result of the environmental information to obtain a threshold value when the acoustic signal is detected, and a feature subset of the acoustic signal in different collection time periods is selected according to the threshold value, so that the accuracy of detecting the acoustic signal of the switch machine is improved.
The technical scheme is that the fault acoustic diagnosis system of the turnout switch machine comprises a signal acquisition module, a storage module, a data analysis module, a signal preprocessing module, an acoustic detection module and an evaluation module, wherein the signal acquisition module is used for acquiring detection data information of the turnout switch machine, and the data analysis module is used for analyzing the detection data information to obtain environmental influence parameters and threshold values; the system management process is specifically as follows:
(1) The signal acquisition module uploads the acquired detection data information of the detection turnout point machine to the storage module;
(2) The data analysis module can analyze the detection data information, firstly analyze the environmental information in the detection data information to obtain noise influence parameters and environmental influence parameters, and then obtain a selection threshold value when the characteristics of the acoustic signals in the detection data information are selected through the analysis of the environmental influence parameters and the noise influence parameters, wherein the specific analysis process is as follows:
step one, a data analysis module calls all detection data information of the turnout point machine, calls acquisition time information in the detection data information, and marks an acquisition time period in the acquisition time information as t 1 ，t 2 ，t 3 ...t n N represents the collection times, a signal preprocessing module preprocesses acoustic signals collected in different collection time periods according to main influence factors obtained by an evaluation module on evaluation analysis results of detection data information in the collection time periods to obtain preprocessing results, the preprocessing results are sent to a data analysis module, and the data analysis module analyzes filtered noise in the preprocessing results to obtain noise influence parameters;
step two, the data analysis module extracts the environmental characteristic vector (X) of the acquisition time period from the environmental information 1 ，X 2 ，X 3 ...X s ) S represents the number of environmental feature variables, the main component analysis method analyzes the environmental feature vector of each acquisition time period to obtain main environmental features, and the specific process has the following formula:
F i ＝a 1i X 1 +a 2i X 2 +…+a si X s ；
wherein a is 1i 、a 2i 、a si I epsilon [1, n ] is the influence coefficient of the element in the environment characteristic vector]I represents a subscript of the acquisition time period, and F is analyzed by principal component analysis i The main environmental characteristics in n corresponding to all the acquisition time periods are obtained, each main environmental characteristic represents the environmental characteristic with the largest influence in the corresponding acquisition time period, and an environmental characteristic matrix is obtained according to the influence degree of the largest environmental characteristic on acoustic signals in all the acquisition time periodsCalculating the trace of the environment characteristic matrix X as an environment influence parameter T;
step three, the data analysis module further analyzes according to the preprocessing result of the signal preprocessing module, the environment characteristic matrix X, the noise influence parameter H and the environment influence parameter to obtain a threshold G when the characteristics of the acoustic signals are selected, the detection data information also comprises the noise intensity, the signal to noise ratio and the acoustic signals of each acquisition time period, and the calculation formula is as follows:
wherein G represents a vector, each element in G represents a threshold value of a corresponding acquisition time period, i represents a label of the acquisition time period, and i epsilon [1, n ];
(3) The acoustic detection module selects a proper feature subset according to the threshold value obtained by the analysis of the data analysis module, the acoustic detection module carries out joint detection on acoustic signals to obtain a detection result, and when the detection result is that the turnout switch machine is faulty, the detection module sends maintenance information to maintenance personnel, and the maintenance personnel carries out timely inspection and maintenance on the turnout switch machine.
The evaluation module evaluates and analyzes the environment information in the detection data information and the detected fault information to obtain main influence factors, and the signal preprocessing module preprocesses the acoustic signals of different acquired time periods according to the main influence factors to obtain preprocessing results, wherein the specific analysis process is as follows:
step 1, an evaluation analysis module evaluates and analyzes environment information in detection data information and detected fault information to obtain main influence factors corresponding to each acquisition time period;
step 2, using the first acquisition time period t 1 The method comprises the steps of performing iteration for a starting point, dividing n acquisition time periods into n-1 acquisition tuples, wherein two adjacent acquisition time periods are one acquisition tuple;
step 3, when main influencing factors corresponding to one acquisition tuple are the same, a signal preprocessing module invokes different noise reduction processing methods of acoustic signals from a storage module, and processes the acoustic signals in two acquisition time periods in one acquisition tuple by using different noise reduction processing methods to obtain two noise reduction processing results, wherein the two noise reduction processing results are noise reduction acoustic signals after noise reduction processing, and correlation coefficients of the two noise reduction acoustic signals are calculated;
step 4, when the main influencing factors corresponding to one acquisition tuple are different, adopting the same noise reduction processing method to obtain two noise reduction processing results, calculating correlation coefficients of the two noise reduction processing results, and obtaining a correlation vector R= (R) according to the correlation coefficients 1 ，r 2 ，r 3 ...r n-1 ) The calculation equation of the correlation coefficient is as follows:
wherein N is the number of features in the feature set of the noise-reduced acoustic signal after noise reduction, X and Y are feature vectors of the signal, X (j), Y (j) is a parameter value of the j-th feature, and the preprocessing result obtained by the signal preprocessing module through analysis comprises a correlation vector, filtered noise and a filtered acoustic signal and is sent to the data analysis module.
The evaluation analysis module evaluates and analyzes the environment information in the detection data information and the detected fault information to obtain main influence factors corresponding to each acquisition time period, the fault information comprises acoustic signals corresponding to the faults, the acquisition time periods, a noise reduction processing method, the environment information and fault types, the environment information in the fault information of the same fault type is analyzed, the influence degree of all environment influence factors on the faults is calculated, the fixed ratio of the influence degree of different environment influence factors on the faults is calculated, the fixed ratio corresponds to the fault types, when the evaluation module analyzes the different acquisition time periods, the ratio of each acquisition time period is analyzed and compared with the fixed ratio, the environment influence factor with the largest ratio in the ratio is determined to be the main influence factor, and the main influence factor in each acquisition time period is sent to the signal preprocessing module.
The data analysis module analyzes the filtered noise and the correlation vector in the preprocessing result to obtain noise influence parameters, firstly calculates the signal-to-noise ratio difference of the acoustic signals of each acquisition time period before and after filtering, then weights each element in the correlation vector according to the signal-to-noise ratio difference to obtain the noise influence parameters, and the data acquisition module monitors each part of the switch machine and acquires the acoustic signals of the switch machine running once every other time period.
Due to the adoption of the technical scheme, compared with the prior art, the invention has the following advantages;
1. the system evaluation module obtains an environmental influence factor ratio by analyzing the existing environmental information and fault detection results, then obtains main influence factors by utilizing the environmental influence factor ratio, then sends the main influence factors to the signal preprocessing module, the signal preprocessing module firstly judges the environmental information in each acquisition time period by utilizing the main influence factors, selects different noise reduction methods according to the judging results to preprocess acoustic signals, obtains a correlation coefficient by comparing the preprocessing results of the acoustic signals in the acquisition time periods of the same main influence factors by comparing different methods in one acquisition tuple, and reflects the generation of noise reduction errors of different signals in the noise reduction process by utilizing the correlation coefficient.
2. The data analysis module analyzes the evaluation result of the evaluation module, the preprocessing result of the signal preprocessing module and the environmental signal to obtain noise influence parameters, analyzes the environmental information in the acquisition time period to obtain the environmental influence parameters, obtains a threshold value through the preprocessing result, the noise influence parameters and the environmental influence parameters, selects a feature subset of the acoustic signal through the threshold value, can correspond to different feature subsets under the selection of different threshold values, improves the accuracy of detecting the fault acoustic signal of the switch machine through the selection of the threshold value, and solves the errors caused by the feature selection of the acoustic signal.
Drawings
FIG. 1 is an overall block diagram of the present system;
FIG. 2 is a flow chart of the overall calculation of the present system;
FIG. 3 is an analysis flow chart of the signal preprocessing module;
FIG. 4 is an analysis flow chart of the data analysis module.
Detailed Description
The foregoing and other features, aspects and advantages of the present invention will become more apparent from the following detailed description of the embodiments with reference to the accompanying drawings, 1-4. The following embodiments are described in detail with reference to the drawings.
The system comprises a signal acquisition module, a storage module, a data analysis module, a signal preprocessing module, an acoustic detection module and an evaluation module, wherein the signal acquisition module is used for acquiring detection data information of the switch machine, and the data analysis module is used for analyzing the detection data information to obtain environmental influence parameters and threshold values; in acoustic detection, the accuracy of detecting the fault of the switch machine is affected between the accuracy of detecting the acoustic signals, but the acquisition, transmission, pretreatment and detection of the acoustic signals are affected by a plurality of external factors, and the accuracy of detecting the acoustic signals can be improved by selecting the noise reduction and the characteristics of the acoustic signals from the environmental information during the acquisition of the acoustic signals and the detected detection information, wherein the system management process is as follows:
(1) The signal acquisition module uploads the acquired detection data information of the detection turnout point machine to the storage module;
(2) The data analysis module can analyze the detection data information, firstly analyze the environmental information in the detection data information to obtain noise influence parameters and environmental influence parameters, and then obtain a selection threshold value when the characteristics of the acoustic signals in the detection data information are selected through the analysis of the environmental influence parameters and the noise influence parameters, wherein the specific analysis process is as follows:
step one, a data analysis module calls all detection data information of the turnout point machine, calls acquisition time information in the detection data information, and marks an acquisition time period in the acquisition time information as t 1 ，t 2 ，t 3 ...t n N represents the collection times, a signal preprocessing module preprocesses acoustic signals collected in different collection time periods according to main influence factors obtained by an evaluation module on evaluation analysis results of detection data information in the collection time periods to obtain preprocessing results, the preprocessing results are sent to a data analysis module, the data analysis module analyzes filtered noise in the preprocessing results to obtain noise influence parameters, noise in the acoustic signals is impossible to be completely eliminated, abnormal signals from the acoustic signals to the point switch machine when fault abnormality occurs can be more accurately detected by improving the signal-to-noise ratio of the acoustic signals, the noise of the acoustic signals is influenced by environmental information, and the environment during collection has a certain influence on the fault of the point switch machine;
step two, the data analysis module extracts the environmental characteristic vector (X) of the acquisition time period from the environmental information 1 ，X 2 ，X 3 ...X s ) S represents the number of environmental feature variables, the main component analysis method analyzes the environmental feature vector of each acquisition time period to obtain main environmental features, and the specific process has the following formula:
F i ＝a 1i X 1 +a 2i X 2 +…+a si X s ；
wherein a is 11 For the influence coefficient of each element in the environment feature vector, i E [1, n]I represents a subscript of the acquisition time period, the formula is one of formulas corresponding to an environmental feature vector, and the corresponding F is obtained by performing principal component analysis through s corresponding formulas of the environmental feature vector i Analysis of F by principal component analysis i The method comprises the steps of obtaining main environmental features in n corresponding to all acquisition time periods, wherein each main environmental feature represents the environmental feature with the largest influence in the corresponding acquisition time period, determining the different influence degree of the largest environmental feature on acoustic signals, calculating the influence degree of the environmental feature with the largest influence on each acoustic feature of the acoustic signals to obtain an environmental influence vector, obtaining an environmental feature matrix according to the environmental influence vector of each acquisition time period, and obtaining the environmental feature matrix according to the influence degree of the main environmental feature on the acoustic signals in all the acquisition time periodsCalculating the trace of the environment characteristic matrix X as an environment influence parameter T;
step three, the data analysis module further analyzes according to the preprocessing result of the signal preprocessing module, the environment characteristic matrix X, the noise influence parameter H and the environment influence parameter to obtain a threshold G when the characteristics of the acoustic signals are selected, the detection data information also comprises the noise intensity, the signal to noise ratio and the acoustic signals of each acquisition time period, and the calculation formula is as follows:
wherein G represents a vector, each element in G represents a threshold value of a corresponding acquisition time period, T represents an environmental impact parameter, H represents a noise impact parameter, C represents a weighting of the environment, i represents a label of the acquisition time period, and i epsilon [1, n ];
(3) The acoustic detection module selects a proper feature subset according to the threshold value obtained by the analysis of the data analysis module, the acoustic detection module carries out joint detection on acoustic signals to obtain a detection result, and when the detection result is that the turnout switch machine is faulty, the detection module sends maintenance information to maintenance personnel, and the maintenance personnel carries out timely inspection and maintenance on the turnout switch machine.
The evaluation module evaluates and analyzes the environment information in the detection data information and the detected fault information to obtain main influence factors, the signal preprocessing module preprocesses the acoustic signals of different acquisition time periods according to the main influence factors to obtain preprocessing results, the signal preprocessing module performs noise reduction processing on the acoustic signals acquired by the signal acquisition module, the noise reduction mode is a noise reduction method stored in the storage module, different acoustic signals are subjected to noise reduction by adopting different noise reduction methods, so that the noise reduction degree of the acoustic signals cannot be damaged by the dynamic noise reduction process adopting different noise reduction methods, and the specific analysis process is as follows:
step 1, an evaluation analysis module evaluates and analyzes environment information in detection data information and detected fault information to obtain main influence factors corresponding to each acquisition time period;
step 2, using the first acquisition time period t 1 The method comprises the steps of performing iteration for a starting point, dividing n acquisition time periods into n-1 acquisition tuples, wherein two adjacent acquisition time periods are one acquisition tuple;
step 3, when main influencing factors corresponding to one acquisition tuple are the same, a signal preprocessing module invokes different noise reduction processing methods of acoustic signals from a storage module, and processes the acoustic signals in two acquisition time periods in one acquisition tuple by using different noise reduction processing methods to obtain two noise reduction processing results, wherein the two noise reduction processing results are noise reduction acoustic signals after noise reduction processing, and correlation coefficients of the two noise reduction acoustic signals are calculated;
step 4, when the main influencing factors corresponding to one acquisition tuple are different, adopting the same noise reduction processing method to obtain two noise reduction processing results, calculating correlation coefficients of the two noise reduction processing results, and obtaining a correlation vector R= (R) according to the correlation coefficients 1 ，r 2 ，r 3 ...r n-1 ) The calculation equation of the correlation coefficient is as follows:
wherein N is the number of features in the feature set of the noise-reduced acoustic signal after noise reduction, X and Y are feature vectors of the signal, X (j), Y (j) is a parameter value of the j-th feature, and the preprocessing result obtained by the signal preprocessing module through analysis comprises a correlation vector, filtered noise and a filtered acoustic signal and is sent to the data analysis module.
The evaluation analysis module is used for evaluating and analyzing the environment information in the detection data information and the detected fault information to obtain main influence factors corresponding to each acquisition time period, the fault information comprises acoustic signals corresponding to the faults, the acquisition time periods, a noise reduction processing method, the environment information and fault types, the evaluation module is used for analyzing the detection information comprising detection results stored in the storage module, the detection information comprises the fault information, the fault information and the environment information are enabled to correspond to the environment information through the analysis of the fault information and the environment information, the environment information in the fault information of the same fault type is analyzed, the influence degree of all environment influence factors on the faults is calculated, the influence coefficient corresponding to each environment influence factor is calculated through a logistic regression method, the fixed ratio of the influence degree of different environment influence factors on the faults is calculated through the influence coefficient, the fixed ratio is enabled to correspond to the fault type, when the evaluation module analyzes the different acquisition time periods, the ratio of each acquisition time period is analyzed, the environment influence factor with the largest ratio is determined to be the main influence factor, and the influence factor with the largest ratio is sent to the preprocessing module in each main influence time period.
The data analysis module analyzes the filtered noise and the correlation vector in the preprocessing result to obtain noise influence parameters, firstly calculates the signal-to-noise ratio difference of the acoustic signals of each acquisition time period before and after filtering, then weights each element in the correlation vector according to the signal-to-noise ratio difference to obtain the noise influence parameters, and the data acquisition module monitors each part of the switch machine and acquires the acoustic signals of the switch machine running once every other time period.
The acoustic detection module performs joint detection on the acoustic signals after the noise reduction processing of the signal preprocessing module according to the analysis result of the data analysis module, selects different acoustic signal feature subsets according to the threshold value analyzed and processed by the data analysis module, performs detection and analysis on the acoustic signals by utilizing the acoustic signal feature subsets and an analysis method in machine learning, judges the acoustic signals, trains the feature subsets of the acoustic signals by utilizing the analysis method in machine learning, further performs detection processing on the acoustic signals, and detects abnormal signals with faults of the switch machine.
When the system is particularly used, the system mainly comprises a signal acquisition module, a storage module, a data analysis module, a signal preprocessing module, an acoustic detection module and an evaluation module, wherein the signal acquisition module is used for acquiring detection data information of the turnout point machine and storing the detection data information in the storage module, the evaluation module obtains an environmental influence factor ratio by analyzing the existing environmental information and the detection result of faults, the environmental influence factor ratio is utilized to obtain main influence factors, the main influence factors are transmitted to the signal preprocessing module, the signal preprocessing module firstly judges the environmental information in each acquisition time period by utilizing the main influence factors, selects different noise reduction methods according to the judgment result to preprocess the acoustic signals, the correlation vector is obtained by comparing the preprocessing results of the acoustic signals in the acquisition time period in the acquisition tuple of the same main influence factors by comparing different methods in the acquisition tuple, the data analysis module analyzes the correlation vector of the preprocessing results of the signal preprocessing module to obtain noise influence parameters, environmental information in the acquisition time period is analyzed to obtain environmental influence parameters, finally a threshold value is obtained by analyzing the preprocessing results, the noise influence parameters and the environmental influence parameters, the characteristic subset of the acoustic signals is selected by the threshold value, the threshold values of the characteristic subset of the acoustic signals in different acquisition time periods can be different, the detection accuracy of fault acoustic signals of the turnout switch machine can be improved by selecting the threshold values, the error caused by characteristic selection of the acoustic signals is solved, and when the selected characteristic parameters are too much, the problems of complex model and reduced popularization capability caused by long time required for feature analysis and model training are avoided, and the problems of remarkably reduced accuracy in covering different abnormal point machine sound types and classification due to too few feature parameters are also avoided.
While the invention has been described in connection with certain embodiments, it is not intended that the invention be limited thereto; for those skilled in the art to which the present invention pertains and the related art, on the premise of based on the technical scheme of the present invention, the expansion, the operation method and the data replacement should all fall within the protection scope of the present invention.
Claims (5)
1. The system is characterized by comprising a signal acquisition module, a storage module, a data analysis module, a signal preprocessing module, an acoustic detection module and an evaluation module, wherein the signal acquisition module is used for acquiring detection data information of the switch machine, and the data analysis module is used for analyzing the detection data information to obtain environmental influence parameters and thresholds; the system management process is specifically as follows:
(1) The signal acquisition module uploads the acquired detection data information of the detection turnout point machine to the storage module;
(2) The data analysis module can analyze the detection data information, firstly analyze the environmental information in the detection data information to obtain noise influence parameters and environmental influence parameters, and then obtain a selection threshold value when the characteristics of the acoustic signals in the detection data information are selected through the analysis of the environmental influence parameters and the noise influence parameters, wherein the specific analysis process is as follows:
step one, a data analysis module calls all detection data information of the turnout point machine, calls acquisition time information in the detection data information, and marks an acquisition time period in the acquisition time information as t 1 ，t 2 ，t 3 ...t n N represents the collection times, a signal preprocessing module preprocesses acoustic signals collected in different collection time periods according to main influence factors obtained by an evaluation module on evaluation analysis results of detection data information in the collection time periods to obtain preprocessing results, the preprocessing results are sent to a data analysis module, and the data analysis module analyzes filtered noise in the preprocessing results to obtain noise influence parameters;
step two, the data analysis module extracts the environmental characteristic vector (X) of the acquisition time period from the environmental information 1 ，X 2 ，X 3 ...X s ) S represents the number of environmental feature variables, the main component analysis method analyzes the environmental feature vector of each acquisition time period to obtain main environmental features, and the specific process has the following formula:
F i ＝a 1i X 1 +a 2i X 2 +…+a si X s ；
wherein a is 1i 、a 2i 、a si I epsilon [1, n ] is the influence coefficient of the element in the environment characteristic vector]I represents a subscript of the acquisition time period, by the masterAnalysis by component analysis method F i The main environmental characteristics in n corresponding to all the acquisition time periods are obtained, each main environmental characteristic represents the environmental characteristic with the largest influence in the corresponding acquisition time period, and an environmental characteristic matrix is obtained according to the main environmental characteristic and the influence degree of the analysis result of the evaluation module on the acoustic signals in all the acquisition time periodsCalculating the trace of the environment characteristic matrix X as an environment influence parameter T;
step three, the data analysis module further analyzes according to the preprocessing result of the signal preprocessing module, the environment characteristic matrix X, the noise influence parameter H and the environment influence parameter to obtain a threshold G when the characteristics of the acoustic signals are selected, the detection data information also comprises the noise intensity, the signal to noise ratio and the acoustic signals of each acquisition time period, and the calculation formula is as follows:
wherein G represents a vector, each element in G represents a threshold value of a corresponding acquisition time period, C represents the weighting of the environment, i represents the label of the acquisition time period, and i epsilon [1, n ];
(3) The acoustic detection module selects a proper feature subset according to the threshold value obtained by the analysis of the data analysis module, the acoustic detection module carries out joint detection on acoustic signals to obtain a detection result, and when the detection result is that the turnout switch machine is faulty, the detection module sends maintenance information to maintenance personnel, and the maintenance personnel carries out timely inspection and maintenance on the turnout switch machine.
2. The system of claim 1, wherein the evaluation module evaluates and analyzes the environmental information in the detected data information and the detected fault information to obtain main influencing factors, and the signal preprocessing module preprocesses the acoustic signals of different acquired time periods according to the main influencing factors to obtain preprocessing results, and the specific analysis process is as follows:
step 1, an evaluation analysis module evaluates and analyzes environment information in detection data information and detected fault information to obtain main influence factors corresponding to each acquisition time period;
step 2, using the first acquisition time period t 1 The method comprises the steps of performing iteration for a starting point, dividing n acquisition time periods into n-1 acquisition tuples, wherein two adjacent acquisition time periods are one acquisition tuple;
step 3, when main influencing factors corresponding to one acquisition tuple are the same, a signal preprocessing module invokes different noise reduction processing methods of acoustic signals from a storage module, and processes the acoustic signals in two acquisition time periods in one acquisition tuple by using different noise reduction processing methods to obtain two noise reduction processing results, wherein the two noise reduction processing results are noise reduction acoustic signals after noise reduction processing, and correlation coefficients of the two noise reduction acoustic signals are calculated;
step 4, when the main influencing factors corresponding to one acquisition tuple are different, adopting the same noise reduction processing method to obtain two noise reduction processing results, calculating correlation coefficients of the two noise reduction processing results, and obtaining a correlation vector R= (R) according to the correlation coefficients 1 ，r 2 ，r 3 ...r n-1 ) The calculation equation of the correlation coefficient is as follows:
wherein N is the number of features in the feature set of the noise-reduced acoustic signal after noise reduction, X and Y are feature vectors of the signal, X (j), Y (j) is a parameter value of the j-th feature, and the preprocessing result obtained by the signal preprocessing module through analysis comprises a correlation vector, filtered noise and a filtered acoustic signal and is sent to the data analysis module.
3. The system of claim 2, wherein the evaluation analysis module evaluates and analyzes the environmental information in the detected data information and the detected fault information to obtain main influencing factors corresponding to each acquisition time period, the fault information includes acoustic signals corresponding to the fault, the acquisition time period, a noise reduction processing method, the environmental information and the fault type, analyzes the environmental information in the fault information of the same fault type, calculates the influence degree of all the environmental influencing factors on the fault, calculates a fixed ratio of the influence degree of different environmental influencing factors on the fault, and enables the fixed ratio to correspond to the fault type, and when the evaluation module analyzes the different acquisition time periods, analyzes the ratio of each acquisition time period, compares the ratio with the fixed ratio, determines the environmental influencing factor with the largest ratio in the ratio as the main influencing factor, and sends the main influencing factor in each acquisition time period to the signal preprocessing module.
4. The system of claim 1, wherein the data analysis module analyzes the filtered noise and the correlation vector in the preprocessing result to obtain noise influence parameters, first calculates the difference between the signal to noise ratios of the acoustic signals in each acquisition time period before and after the filtering, and then weights each element in the correlation vector according to the difference between the signal to noise ratios to obtain the noise influence parameters, and the signal acquisition module monitors each part of the switch machine and acquires the acoustic signals of the switch machine running every other time period.
5. A switch machine fault acoustic diagnostic system according to claim 1, wherein,
the acoustic detection module performs joint detection on the acoustic signals after the noise reduction processing of the signal preprocessing module according to the analysis result of the data analysis module, selects different acoustic signal feature subsets according to the threshold value analyzed and processed by the data analysis module, and then performs detection analysis on the acoustic signals by utilizing the acoustic signal feature subsets and an analysis method in machine learning to judge the acoustic signals.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110944630.4A CN113672859B (en) | 2021-08-17 | 2021-08-17 | Fault acoustic diagnosis system for switch machine |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110944630.4A CN113672859B (en) | 2021-08-17 | 2021-08-17 | Fault acoustic diagnosis system for switch machine |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113672859A CN113672859A (en) | 2021-11-19 |
| CN113672859B true CN113672859B (en) | 2024-03-29 |
Family
ID=78543433
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110944630.4A Active CN113672859B (en) | 2021-08-17 | 2021-08-17 | Fault acoustic diagnosis system for switch machine |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113672859B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114820544B (en) * | 2022-05-09 | 2023-06-20 | 西安得眠堂健康管理工作部 | Production monitoring management system of throat-moistening candy |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106017954A (en) * | 2016-05-13 | 2016-10-12 | 南京雅信科技集团有限公司 | Turnout point machine fault early warning system and method based on audio analysis |
| CN106596123A (en) * | 2016-11-22 | 2017-04-26 | 东软集团股份有限公司 | Device fault diagnosis method, device fault diagnosis device, and device fault diagnosis system |
| CN108540330A (en) * | 2018-04-24 | 2018-09-14 | 南京邮电大学 | A kind of network fault diagnosis method based on deep learning under heterogeneous network environment |
| CN110132598A (en) * | 2019-05-13 | 2019-08-16 | 中国矿业大学 | Slewing rolling bearing fault noise diagnostics algorithm |
| CN110596490A (en) * | 2019-09-11 | 2019-12-20 | 天津科技大学 | Intelligent detection method for railway turnout fault |
| CN112697272A (en) * | 2021-02-09 | 2021-04-23 | 郑州铁路职业技术学院 | Point switch fault detection device based on acoustics |
| CN112801367A (en) * | 2021-01-27 | 2021-05-14 | 长沙理工大学 | Fault prediction method based on ARMret model considering rare variables |
| CN113155504A (en) * | 2021-04-30 | 2021-07-23 | 苏州通汇轨道交通技术有限公司 | Intelligent testing system for point switch for rail transit |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20200012944A1 (en) * | 2018-07-05 | 2020-01-09 | Alstom Transport Technologies | Method and system for detecting switch degradation and failures |
-
2021
- 2021-08-17 CN CN202110944630.4A patent/CN113672859B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106017954A (en) * | 2016-05-13 | 2016-10-12 | 南京雅信科技集团有限公司 | Turnout point machine fault early warning system and method based on audio analysis |
| CN106596123A (en) * | 2016-11-22 | 2017-04-26 | 东软集团股份有限公司 | Device fault diagnosis method, device fault diagnosis device, and device fault diagnosis system |
| CN108540330A (en) * | 2018-04-24 | 2018-09-14 | 南京邮电大学 | A kind of network fault diagnosis method based on deep learning under heterogeneous network environment |
| CN110132598A (en) * | 2019-05-13 | 2019-08-16 | 中国矿业大学 | Slewing rolling bearing fault noise diagnostics algorithm |
| CN110596490A (en) * | 2019-09-11 | 2019-12-20 | 天津科技大学 | Intelligent detection method for railway turnout fault |
| CN112801367A (en) * | 2021-01-27 | 2021-05-14 | 长沙理工大学 | Fault prediction method based on ARMret model considering rare variables |
| CN112697272A (en) * | 2021-02-09 | 2021-04-23 | 郑州铁路职业技术学院 | Point switch fault detection device based on acoustics |
| CN113155504A (en) * | 2021-04-30 | 2021-07-23 | 苏州通汇轨道交通技术有限公司 | Intelligent testing system for point switch for rail transit |
Non-Patent Citations (1)
| Title |
|---|
| 深度学习在轴承故障诊断领域的应用研究;洪腾蛟等;《科学技术与工程》;20210808;第21卷(第22期);9203-9211 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113672859A (en) | 2021-11-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109324604B (en) | Intelligent train comprehensive fault analysis method based on multi-source signals | |
| CN110704801B (en) | Bridge cluster structure operation safety intelligent monitoring and rapid detection complete method | |
| CN112504673B (en) | Carrier roller fault diagnosis method, system and storage medium based on machine learning | |
| CN110405537B (en) | Method for establishing guide rail precision prediction model based on deep learning | |
| CN112660745A (en) | Intelligent diagnosis method and system for carrier roller fault and readable storage medium | |
| CN108304348B (en) | Bearing residual life prediction method based on binary wiener process | |
| CN113077005B (en) | Anomaly detection system and method based on LSTM self-encoder and normal signal data | |
| CN110259648B (en) | Fan blade fault diagnosis method based on optimized K-means clustering | |
| CN112393906B (en) | Method for diagnosing, classifying and evaluating health of weak signal fault of bogie bearing of metro vehicle | |
| CN107844067B (en) | A kind of gate of hydropower station on-line condition monitoring control method and monitoring system | |
| CN113672859B (en) | Fault acoustic diagnosis system for switch machine | |
| CN111964909A (en) | Rolling bearing operation state detection method, fault diagnosis method and system | |
| CN108592812B (en) | Method for extracting load strain characteristics and monitoring cracks of optical fiber of fan blade | |
| CN112487890A (en) | Bearing acoustic signal fault diagnosis method based on parallel sparse filtering | |
| CN116383633A (en) | Method and system for detecting faults of machine-made sand vibrating screen through multi-factor comprehensive analysis | |
| CN115060491A (en) | Fan bearing operation health degree assessment method and system based on multi-source data | |
| CN114739671A (en) | Bearing fault diagnosis method based on improved generalized S transformation | |
| CN110956112B (en) | Novel high-reliability slewing bearing service life assessment method | |
| CN207992717U (en) | A kind of gate of hydropower station on-line condition monitoring system | |
| CN112362214A (en) | Method and system for online identification of belt tension | |
| CN115238753B (en) | Self-adaptive SHM data cleaning method based on local outlier factor | |
| CN112660746B (en) | Roller fault diagnosis method and system based on big data technology and storage medium | |
| CN114660378A (en) | Multi-source detection parameter-based contact network comprehensive diagnosis method | |
| CN114004268A (en) | Online diagnosis method and device for traction system fault | |
| CN110451395B (en) | Escalator fault monitoring method |
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 |