CN112660746A - Roller fault diagnosis method and system based on big data technology and storage medium - Google Patents
Roller fault diagnosis method and system based on big data technology and storage medium Download PDFInfo
- Publication number
- CN112660746A CN112660746A CN202011437844.4A CN202011437844A CN112660746A CN 112660746 A CN112660746 A CN 112660746A CN 202011437844 A CN202011437844 A CN 202011437844A CN 112660746 A CN112660746 A CN 112660746A
- Authority
- CN
- China
- Prior art keywords
- logistic regression
- regression model
- roller
- audio data
- model
- 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.)
- Granted
Links
Images
Landscapes
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention discloses a roller fault diagnosis method and system based on a big data technology and a storage medium. The roller fault diagnosis method based on the big data technology comprises the following steps: s1, acquiring carrier roller audio data; s2, extracting the characteristics of the audio data; the characteristics of the audio data specifically include one or any combination of sharpness, noise annoyance and speech interference level; s3, inputting the characteristics of the audio data into a trained logistic regression model, and identifying the running state of the carrier roller by the logistic regression model; s4, if the roller runs abnormally, executing alarm, monitoring or control operation; if the carrier roller runs normally, the fault diagnosis of the carrier roller at the current moment is completed, and the step S5 is executed; and S5, updating the time, repeatedly executing the steps S1 to S4, and carrying out roller fault diagnosis at the next time. The invention can realize the real-time diagnosis of the roller fault, and has the advantages of easy realization, low cost and low algorithm complexity.
Description
Technical Field
The invention relates to the field of fault diagnosis of conveyor carrier rollers, in particular to a carrier roller fault diagnosis method and system based on a big data technology and a storage medium.
Background
The belt conveyor is used for conveying materials and is an important part of industrial production process flow. The belt conveyor can form an efficient transportation assembly line, improves the industrial production efficiency, reduces the labor intensity of workers, and is widely applied to industries such as mines, electric power, wharfs and the like. The belt conveyor runs in a long-time load mode, and various faults are easy to occur, such as: damaged carrier roller, torn belt, etc. Among them, the failure of the idler is one of the main causes of the stoppage of the belt conveyor. The carrier roller is the important operation part of belt conveyor, and numerous (about 1 ~ 3 meters arrange a set of), mainly play support sticky tape and bear and reduce the operational resistance effect. The common carrier roller faults are poor operation (blocking, eccentricity and damage), if an abnormal carrier roller is not found in the early stage and cannot be replaced in time, the friction, fire, tearing and the like of an adhesive tape are easily caused, even personal casualty accidents are caused, and the major economic losses of equipment damage, production halt and the like are caused. Therefore, the roller of the belt conveyor is monitored abnormally, fault precursors are found in time and an alarm is given, and the roller monitoring device has great significance for safe and efficient production. The belt conveyor idler trouble mainly manifests the form as abnormal sound, noise, for example: high-frequency hoarseness, sand abnormal sound and the like, so that the time-frequency domain characteristics of the sound signal can be extracted and analyzed, and early detection and alarm of the abnormal carrier roller are realized.
At present, the fault detection of the carrier roller of the belt conveyor mainly adopts manual inspection, detection workers complete maintenance work by means of beating, observing, overhearing and the like according to years of working experience, the mode is low in efficiency, detection leakage exists, and early faults of the carrier roller cannot be found in time. In addition, research on partial automatic detection methods of the carrier roller is available, wherein most automatic detection methods adopt a method of fusing multiple sensors to judge carrier roller faults, but the method has complex hardware and complex data and is inconvenient to apply; a few automatic detection methods only collect audio data for judgment, although the collected data is simple, the fault is judged only by judging the decibel of the audio data, the audio data is only subjected to simple processing judgment, the characteristics of the audio are not deeply mined, the judgment result is inaccurate, and the interference is easy to occur. For example, patent application CN201910168680.0 discloses an intelligent unmanned inspection system for a coal transportation system, which realizes fault diagnosis by fusion of various information such as vision and noise, but has high hardware cost and large calculation amount. For example, patent application No. CN201810532489.5 discloses a fault identification method, device and system for a conveyor, the device identifies the fault of a carrier roller by extracting a periodic audio signal, but the scheme has high sampling frequency, the feature extraction of the audio signal is single, and the audio signal is still susceptible to interference.
Disclosure of Invention
The invention aims to overcome the defects that the carrier roller fault cannot be judged quickly and efficiently due to insufficient utilization of audio signal characteristics in the prior art, and provides a carrier roller fault diagnosis method, a carrier roller fault diagnosis system and a storage medium which are convenient to implement, high in operation efficiency and based on a big data technology.
In order to achieve the above purpose, the invention provides the following technical scheme:
a roller fault diagnosis method based on big data technology comprises the following steps:
s1, acquiring carrier roller audio data;
s2, extracting the characteristics of the audio data; the characteristics of the audio data specifically include one or any combination of sharpness, noise annoyance and speech interference level;
s3, inputting the characteristics of the audio data into a trained logistic regression model, and identifying the running state of the carrier roller by the logistic regression model;
s4, if the roller runs abnormally, executing alarm, monitoring or control operation; if the carrier roller runs normally, the fault diagnosis of the carrier roller at the current moment is completed, and the step S5 is executed;
and S5, updating the time, repeatedly executing the steps S1 to S4, and carrying out roller fault diagnosis at the next time.
Preferably, the characteristics of the audio data collected in step S2 include sharpness, noise annoyance, and speech interference level; and performing frequency domain transformation on the audio data to obtain frequency domain information of the audio data, and then calculating the sharpness, the noise annoyance degree and the speech interference level.
Preferably, the calculation formula of the sharpness and the speech interference level is as follows:
sharpness calculation formula:
where S is sharpness, k is a weighting factor, 24Bark represents 24 characteristic bands of the sharpness model, z is a critical band, N' (z) is the characteristic loudness in the critical band z, g (z) is a loudness weighting function set according to different critical bands:
the speech interference level calculation formula is as follows:
where LST is speech interference level, LP1、LP2、LP3Respectively, the noise sound pressure levels of the triple frequency bands centered at 500Hz, 1000Hz, and 2000 Hz.
Preferably, the logistic regression model is trained by:
a1, acquiring the marked audio data characteristics;
a2, constructing a logistic regression model, inputting the marked audio data features into the logistic regression model, and adjusting parameters in the logistic regression model by adopting a gradient descent method to obtain the trained logistic regression model.
Preferably, the logistic regression model mathematical expression is as shown in formula (1):
wherein X represents a training sample set; x represents an independent variable, namely, an audio feature extracted from each sample data; h (x) is a model prediction function and represents the fault probability of the carrier roller; beta represents a parameter of the logistic regression model; y' is a model prediction result and is used for representing the identified running state of the carrier roller; h, (x) is not less than 0.5, and y' is 1, which indicates that the carrier roller runs abnormally; h, (x) is less than 0.5, and y' is 0, which indicates that the carrier roller runs normally;
the loss function is shown in equation (2):
where m represents the number of data points in the data set, i.e., the number of samples; x is a training sample; i represents the ith set of samples or the ith prediction; h (x) represents a prediction result obtained by inputting the training sample into the model, and y represents a real result labeled by the training sample;
solving the loss function by adopting a gradient descent method, wherein the solving result is shown as the formula (3):
where j represents the number of gradient iterations.
Preferably, the logistic regression model is trained by:
step B1, acquiring the labeled audio data features, and dividing the labeled audio data features into a training data set and a testing data set;
step B2, constructing a logistic regression model, inputting the training data set into the logistic regression model, and adjusting parameters in the logistic regression model by adopting a gradient descent method to obtain the trained logistic regression model;
and step B3, inputting the test data set into the trained logistic regression model for testing to obtain a prediction result, and performing statistical analysis on the prediction result of the test data set and the actual result of the test data to obtain the prediction accuracy of the model.
Preferably, the audio data collected in the real-time identification process of the model in the step S1 and the corresponding carrier roller operating state identified in the step S3 are also used as training samples for training the logistic regression model.
A big data technology based idler failure diagnostic system operating as any one of the above methods, comprising the following modules connected in series: the device comprises a data acquisition module, a feature extraction module, a logistic regression model module and a monitoring module.
Preferably, the system further comprises a model training module; the output end of the characteristic extraction module is connected with the input end of the model training module, and the output end of the model training module is connected with the input end of the logistic regression model module.
A readable storage medium having stored thereon a computer program which, when executed by a processor, carries out the steps of any of the big data technology based idler failure diagnosis methods as described above.
Compared with the prior art, the invention has the beneficial effects that:
1. the method comprises the steps of carrying out deep analysis on audio data collected in real time to obtain characteristic indexes such as sharpness, noise annoyance degree and speech interference level, inputting extracted data characteristics into a trained logistic regression model, monitoring the running state of a roller of the rubber belt conveyor according to a model prediction result, and realizing real-time diagnosis of roller abnormity.
2. Through the state of real-time supervision bearing roller, can realize that the bearing roller in time discovers the anomaly of bearing roller when not breaking down, carries out the precursor discernment of trouble, improves the security, reduces the influence to production.
3. In the identification process, the collected audio data is also used for training the logistic regression model, and the accuracy of the model is improved.
Description of the drawings:
fig. 1 is a flowchart of a idler fault diagnosis method based on big data technology according to an exemplary embodiment 1 of the present invention;
FIG. 2 is a flowchart of model training in exemplary embodiment 1 of the present invention;
FIG. 3 is a flowchart of model training with test validation in exemplary embodiment 1 of the present invention;
fig. 4 is a system block diagram of a roller fault diagnosis system based on big data technology in exemplary embodiment 2 of the invention.
Detailed Description
The present invention will be described in further detail with reference to test examples and specific embodiments. It should be understood that the scope of the above-described subject matter is not limited to the following examples, and any techniques implemented based on the disclosure of the present invention are within the scope of the present invention.
Example 1
As shown in fig. 1, the present embodiment provides a method for diagnosing a carrier roller fault based on big data technology, which includes the following steps:
s1, acquiring carrier roller audio data;
s2, extracting the characteristics of the audio data; the characteristics of the audio data specifically include one or any combination of sharpness, noise annoyance and speech interference level;
s3, inputting the characteristics of the audio data into a trained logistic regression model, and identifying the running state of the carrier roller by the logistic regression model;
s4, if the roller runs abnormally, executing alarm, monitoring or control operation; if the carrier roller runs normally, the fault diagnosis of the carrier roller at the current moment is completed, and the step S5 is executed;
and S5, updating the time, repeatedly executing the steps S1 to S4, and carrying out roller fault diagnosis at the next time.
In the embodiment, a large amount of audio data are collected for analysis and classification, and particularly, sound quality analysis is performed on the audio data in carrier roller operation through a voice signal processing technology, so that deep feature analysis is performed on the audio data, and feature indexes such as sharpness, noise annoyance degree or speech interference level are obtained; then, classifying the characteristic indexes of the audio signals by adopting a logistic regression model, accurately identifying the running state of a carrier roller of the belt conveyor, and judging whether the carrier roller runs normally; when the carrier roller operates abnormally, alarm operation and the like are timely executed to remind of carrying out related maintenance and repair work. When the abnormal phenomena such as looseness of the carrier roller or overhigh temperature of the belt occur, the carrier roller can still run, and the conventional fault diagnosis method cannot find the abnormal phenomena in advance; the method of the embodiment can early discover the abnormal phenomena by analyzing the characteristics of the collected audio data, thereby realizing the precursor monitoring of the fault; and the fault of the belt conveyor is effectively prevented by combining the identification and the alarm of the logistic regression model, the defect of manual inspection based on experience maintenance can be completely overcome, the purpose of automatic unmanned inspection is really realized, the working efficiency is greatly improved, and the early warning is realized, the omission detection is avoided and the like.
Illustratively, step S1 specifically includes the following steps: and arranging microphones near the carrier roller according to the pickup distance of the microphones, and collecting audio data of the carrier roller through the microphones. The microphone collects data in units of frames, usually 10ms samples are taken as one frame, and the frame length array N is determined according to a sampling frequency and a frame length, for example, the sampling frequency is 16kHz, 10ms represents 160 samples, i.e., the frame length array N is 160. For smooth transition of Data in subsequent processing, Data of a previous frame and Data of a current frame are spliced to form an array Data with the length of 2N. I.e. the audio Data of the idlers, is recorded as Data set for subsequent Data analysis.
In the embodiment, only the audio data is collected for subsequent analysis, and only related hardware equipment such as a microphone, a DSP and the like needs to be added during implementation, so that the newly added equipment is simple and low in cost; and the subsequent data analysis and processing process can be realized only by a software algorithm without damaging the original production line, and the method is not influenced by environmental factors and has strong anti-interference performance.
Illustratively, the characteristics of the audio data collected at step S2 include sharpness, noise annoyance, and speech disturbance level. The audio features are divided into time domain features and frequency domain features, and the variety is wide. In the embodiment, three characteristic indexes of sharpness, noise annoyance degree and speech interference level are selected for characteristic analysis, the accuracy of model classification and identification and occupied resources are considered, and a model with higher accuracy can be obtained by occupying fewer resources.
Step S2 extracts features of the audio data by the following steps. And (4) carrying out sound quality analysis on the audio data acquired in the step (S1) and extracting technical indexes such as sharpness, noise annoyance degree and speech interference level. The specific operation is as follows: carrying out frequency domain transformation on the audio data to obtain frequency domain information of the audio data, and then calculating the sharpness, the noise annoyance degree and the speech interference level by using the audio data before and after transformation; the calculation formulas of the sharpness S, the noise annoyance PA and the speech interference level LST are shown as follows;
sharpness S: the sharpness calculation method has no standard method at present, and the invention adopts DIN45692-2009 standard, and the prototype expression thereof is as follows:
k is a weighting coefficient, k is more than or equal to 0.105 and less than 0.115, and the value k is 0.11 in the embodiment; 24Bark represents the 24 characteristic bands of the sharpness model; z is a critical frequency band, N' (z) is the specific loudness on the critical frequency band z, the relation index between the z and the 24Bark characteristic frequency band is given in DIN45692-2009, and the loudness calculation is carried out according to ISO 532-2-2017; g (z) loudness weighting functions set according to different critical bands for the DIN45692 standard model:
noise annoyance PA: the index describes the boredom degree of sound, comprehensively considers the influence of loudness, sharpness, roughness and fluctuation degree, and belongs to a subjective psychological test index. According to the invention, an ISO standard Noise database Noise 92 is played back through an audio test system, subjective evaluation is carried out based on ISO/TS 15666-.
Speech-interference level LST: according to ISO/TR 3352-As a speech interference level evaluation index. Is calculated by the formula
Wherein L isP1、LP2、LP3Respectively representing the noise sound pressure levels of three-time frequency bands with 500Hz, 1000Hz and 2000Hz as the centers, and the calculation formula of the sound pressure level is
p0For reference sound pressure, the value of 2 × 10 is taken in this embodiment-5Pa，peFor effective sound pressure value, the calculation formula is
x represents the sampled data of the speech signal and T represents the number of audio data points.
In the embodiment, the audio data is subjected to deep feature analysis by selecting feature indexes such as sharpness, noise annoyance degree and speech interference level. The sharpness index may be used to describe the timbre characteristics; the noise annoyance degree can be used to exclude the influence of part of non-interesting noise, such as natural noise, noise in other directions, etc.; the speech disturbance level may be focused on sound pressure levels of multiple frequencies. Three characteristic indexes of sharpness, noise annoyance degree and speech interference level are selected for characteristic analysis, accuracy of model classification and recognition and occupied resources are considered, and a model with high accuracy can be obtained by occupying less resources.
In the embodiment, a logistic regression model is selected to classify and identify the extracted audio features. The logistic regression model belongs to a classification algorithm in machine learning, and can be trained by using the existing labeled data to obtain an exact prediction result. The logistic regression model is simple in structure, convenient to implement, small in calculation amount during classification, low in calculation cost, high in speed, low in storage resource, not prone to being influenced by noise data, and convenient for observation of sample probability scores. In the embodiment, deep feature analysis is performed on audio data through a voice signal processing technology, and then the extracted features are identified and classified based on a logistic regression model, so that the roller fault diagnosis method based on the big data technology, which is convenient to implement and efficient in operation, is obtained.
Illustratively, as shown in FIG. 2, the logistic regression model of step S3 is trained by:
a1, acquiring the marked audio data characteristics;
and collecting a large amount of carrier roller audio data, extracting the characteristics of the audio data according to the step S2, and labeling the running state of the carrier roller.
In this embodiment, a large amount of audio data is acquired, the audio data is sorted according to the acquisition time sequence, characteristic indexes such as sharpness, noise annoyance degree and speech interference level of each data are calculated, carrier roller operation state labeling is performed on each group of data, and a data characteristic analysis result is mapped with the carrier roller operation state. The marking information includes that the carrier roller normally operates and the carrier roller abnormally operates, wherein the carrier roller abnormally operates to indicate that the carrier roller is loosened or abnormal phenomena such as too high temperature of a belt occur, the carrier roller possibly still operates at the moment, but the carrier roller fails and stops operating due to continuous use and untimely overhaul, so that safety problems and production influence are caused. In the embodiment, the phenomena of looseness and the like of the carrier roller are classified as abnormal phenomena, the logistic regression model is trained, data support is provided for fault diagnosis of the carrier roller of the belt conveyor in the next step, and the abnormal phenomena can be conveniently detected and identified in time in the follow-up process. The method can realize the carrier roller fault precursor detection, improves the safety and ensures the smooth production. The audio data is labeled in the manner shown in table 1, and table 1 shows an array in which the operation states of five groups of carrier rollers are normal and the operation states of five groups of carrier rollers are abnormal;
table 1 carrier roller running state marking example table
| Sharpness degree | Degree of noise annoyance | Speech interference stage | Carrier roller condition |
| 1.355 | 267.241 | 92.32 | Is normal |
| 1.356 | 266.025 | 90.96 | Is normal |
| 1.392 | 245.282 | 89.21 | Is normal |
| 1.405 | 242.927 | 88.59 | Is normal |
| 1.283 | 212.157 | 92.27 | Is normal |
| 1.162 | 217.609 | 86.11 | Abnormality (S) |
| 1.224 | 223.427 | 87.52 | Abnormality (S) |
| 1.221 | 234.667 | 85.05 | Abnormality (S) |
| 1.256 | 249.289 | 85.24 | Abnormality (S) |
| 1.208 | 249.380 | 89.68 | Abnormality (S) |
A2, constructing a logistic regression model, inputting the marked audio data features into the logistic regression model, and adjusting parameters in the logistic regression model by adopting a gradient descent method to obtain the trained logistic regression model.
Illustratively, the mathematical expression of the logistic regression model is shown in formula (1).
Wherein X represents a training sample set; x represents an independent variable, namely, an audio feature extracted from each sample data; h (x) is a model prediction function and represents the fault probability of the carrier roller; beta represents a parameter of the logistic regression model; y' represents a model prediction result and is used for representing the identified carrier roller running state; h, (x) is not less than 0.5, and y' is 1, which indicates that the carrier roller runs abnormally; h (x) <0.5 and y ═ 0, which indicates that the idler is operating normally.
Illustratively, the loss function is shown in equation (2), where m represents the number of data points in the data set, i.e., the number of samples; x is a training sample; i represents the ith set of samples or the ith prediction; h (x) represents a prediction result obtained by inputting the training sample into the model, and y represents a real result labeled by the training sample;
the algorithm for solving the minimum loss function adopts a gradient descent method, namely, the minimum value of beta derivation of a solution formula (2) is solved, and the solution result is shown as a formula (3);
in the formula, j represents the number of gradient iterations. In the model training process, the model parameter beta is adjusted through multiple iterations, and when the iteration times reach a preset threshold value or the training error reaches the preset threshold value, the model training is finished.
In this embodiment, the obtained labeled audio data features are input into the logistic regression model as training samples, the logistic regression model calculates a training error according to a loss function, and then parameters in the logistic regression model are adjusted by a gradient descent method until the number of iterations reaches a preset threshold or the training error reaches an expected value through iteration, and the training of the model is finished to obtain a trained logistic regression model.
Illustratively, as shown in fig. 3, when the logistic regression model is trained, the method further includes a step of testing and verifying the trained model, and the specific steps are as follows:
step B1, dividing the audio data features labeled in the step A1 into a training data set and a testing data set;
step B2, training a logistic regression model by using a training data set;
and step B3, inputting the test data set into the trained logistic regression model for testing to obtain a prediction result, and performing statistical analysis on the prediction result of the test data set and the actual result of the test data to obtain the prediction accuracy of the model.
In this embodiment, the labeled sample data is divided into a training data set and a testing data set, which respectively account for 80% and 20%, where the training data set is used to train the model, and the testing data set is used to test and verify the trained model to determine the validity of the model. And carrying out statistical analysis on the prediction result of the test data set and the actual result of the test data to obtain the prediction accuracy of the model. The prediction accuracy here represents the current training effect of the model, and is used for testing and verifying the trained model, and if the prediction accuracy is low, the model needs to be further adjusted. For example, the prediction accuracy may be low due to insufficient model training, uneven data distribution, or the algorithm model not meeting the current application scenario. By timely finding out the problem of the model during training, the training efficiency of the model and the accuracy of the model can be improved.
Illustratively, the audio data collected during the real-time identification process of the model in step S1 and the corresponding idler operating state identified in step S3 are also used as training samples for training the logistic regression model. The audio data collected in real time are also used for training the model, the number of model training samples is increased, meanwhile, the data collected in the real-time identification process are more in line with the application scene, the logistic regression model is trained through the audio data, the logistic regression model more in line with the current application scene can be obtained, and the accuracy of the model is improved.
The present example trains a logistic regression model through steps A1-A2, or steps B1-B3; and then, carrying out feature extraction on the audio data acquired in real time, inputting the extracted data features into a trained logistic regression model, and realizing real-time diagnosis on the carrier roller according to the running state of the carrier roller of the monitoring rubber belt conveyor. If the operation of the roller of the belt conveyor is judged to be abnormal, alarm operation is executed so as to take monitoring prompt measures, the roller is overhauled in time and the fault of the belt conveyor is effectively prevented.
Example 2
As shown in fig. 4, the present embodiment provides a roller fault diagnosis system based on big data technology, and the roller fault diagnosis system based on big data technology described below and the roller fault diagnosis method based on big data technology described above may be referred to correspondingly.
Referring to fig. 4, the system includes the following modules connected in series: the device comprises a data acquisition module, a feature extraction module, a logistic regression model module and a monitoring module;
in this embodiment, the data acquisition module is used for acquiring carrier roller audio data; the characteristic extraction module is used for extracting the characteristics of the audio data; the logistic regression model module is used for identifying the running state of the carrier roller according to the characteristics of the audio data; the monitoring module is used for executing alarming, monitoring or controlling operation according to the identification and classification result of the logistic regression model module. The logistic regression model module stores a trained logistic regression model, and corresponding parameters of the trained logistic regression model are deployed according to the model training process. After the logistic regression model identifies that the running state of the carrier roller is abnormal, the warning device can be driven to correspondingly remind, and a display screen, a camera and the like are adopted to collect the current running condition for further judgment and observation, or the automatic power off and the like are adopted to ensure the safety of production.
The carrier roller fault diagnosis system based on the big data technology further comprises a model training module; the output end of the characteristic extraction module is connected with the input end of the model training module, and the output end of the model training module is connected with the input end of the logistic regression model module. The model training module is used for adjusting parameters of the logistic regression model according to the marked audio data characteristics, and after model training is completed, the trained parameters of the model are deployed into the logistic regression model module so as to recognize the carrier roller running state of the audio data collected in real time.
Illustratively, the model training module is further configured to train the logistic regression model according to the audio data collected in real time and the recognition result output by the logistic regression model module. The audio data collected in real time are also used for training the model, the number of model training samples can be increased, meanwhile, the data collected in the real-time identification process are more in line with the application scene, the logistic regression model is trained through the audio data, the logistic regression model more in line with the current application scene can be obtained, and the accuracy of the model is improved.
By applying the system provided by the embodiment of the invention, the audio data is obtained, the characteristics of the audio data collected in real time are extracted, the extracted data characteristics are input into the trained logistic regression model, the running state of the roller of the rubber belt conveyor can be monitored according to the model prediction result, and the real-time diagnosis and the precursor diagnosis of the roller are realized.
Example 3
Corresponding to the above method embodiment, the present embodiment further provides a readable storage medium, and a readable storage medium described below and a roller fault diagnosis method based on big data technology described above may be referred to correspondingly.
A readable storage medium, on which a computer program is stored, and when being executed by a processor, the computer program implements the steps of the idler failure diagnosis method based on big data technology of the above method embodiment.
The readable storage medium may be a usb disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and various other readable storage media capable of storing program codes.
The foregoing is merely a detailed description of specific embodiments of the invention and is not intended to limit the invention. Various alterations, modifications and improvements will occur to those skilled in the art without departing from the spirit and scope of the invention.
Claims (10)
1. A roller fault diagnosis method based on big data technology is characterized by comprising the following steps:
s1, acquiring carrier roller audio data;
s2, extracting the characteristics of the audio data; the characteristics of the audio data specifically include one or any combination of sharpness, noise annoyance and speech interference level;
s3, inputting the characteristics of the audio data into a trained logistic regression model, and identifying the running state of the carrier roller by the logistic regression model;
s4, if the roller runs abnormally, executing alarm, monitoring or control operation; if the carrier roller runs normally, the fault diagnosis of the carrier roller at the current moment is completed, and the step S5 is executed;
and S5, updating the time, repeatedly executing the steps S1 to S4, and carrying out roller fault diagnosis at the next time.
2. A roller fault diagnosis method based on big data technology as claimed in claim 1, wherein the characteristics of the audio data collected in step S2 include sharpness, noise annoyance and speech interference level; and performing frequency domain transformation on the audio data to obtain frequency domain information of the audio data, and then calculating the sharpness, the noise annoyance degree and the speech interference level.
3. A roller fault diagnosis method based on big data technology as claimed in claim 2, characterized in that the calculation formula of the sharpness and speech disturbance level is as follows:
sharpness calculation formula:
where S is sharpness, k is a weighting factor, 24Bark represents 24 characteristic bands of the sharpness model, z is a critical band, N' (z) is the characteristic loudness in the critical band z, g (z) is a loudness weighting function set according to different critical bands:
the speech interference level calculation formula is as follows:
where LST is speech interference level, LP1、LP2、LP3Respectively represent 500Hz,Noise sound pressure levels of three-fold frequency band centered at 1000Hz and 2000 Hz.
4. A roller fault diagnosis method based on big data technology as claimed in claim 1, characterised in that the logistic regression model is trained by the following steps:
a1, acquiring the marked audio data characteristics;
a2, constructing a logistic regression model, inputting the marked audio data features into the logistic regression model, and adjusting parameters in the logistic regression model by adopting a gradient descent method to obtain the trained logistic regression model.
5. A roller fault diagnosis method based on big data technology as claimed in claim 4, wherein the logistic regression model mathematical expression is as shown in equation (1):
wherein X represents a training sample set; x represents an independent variable, namely, an audio feature extracted from each sample data; h (x) is a model prediction function and represents the fault probability of the carrier roller; beta represents a parameter of the logistic regression model; y' is a model prediction result and is used for representing the identified running state of the carrier roller; h, (x) is not less than 0.5, and y' is 1, which indicates that the carrier roller runs abnormally; h, (x) is less than 0.5, and y' is 0, which indicates that the carrier roller runs normally;
the loss function is shown in equation (2):
where m represents the number of data points in the data set, i.e., the number of samples; x is a training sample; i represents the ith set of samples or the ith prediction; h (x) represents a prediction result obtained by inputting the training sample into the model, and y represents a real result labeled by the training sample;
solving the loss function by adopting a gradient descent method, wherein the solving result is shown as the formula (3):
where j represents the number of gradient iterations.
6. A roller fault diagnosis method based on big data technology as claimed in claim 4, wherein the logistic regression model is trained by the following steps:
step B1, acquiring the labeled audio data features, and dividing the labeled audio data features into a training data set and a testing data set;
step B2, constructing a logistic regression model, inputting the training data set into the logistic regression model, and adjusting parameters in the logistic regression model by adopting a gradient descent method to obtain the trained logistic regression model;
and step B3, inputting the test data set into the trained logistic regression model for testing to obtain a prediction result, and performing statistical analysis on the prediction result of the test data set and the actual result of the test data to obtain the prediction accuracy of the model.
7. A roller fault diagnosis method based on big data technology as claimed in claim 4, wherein the audio data collected during the real-time identification of the model in step S1 and the corresponding roller operating state identified in step S3 are also used as training samples for training the logistic regression model.
8. A large data technology based idler failure diagnosis system operating in accordance with the method of claims 1-7 including the following modules connected in series: the device comprises a data acquisition module, a feature extraction module, a logistic regression model module and a monitoring module.
9. A idler fault diagnosis system based on big data technology as claimed in claim 7 further including a model training module; the output end of the characteristic extraction module is connected with the input end of the model training module, and the output end of the model training module is connected with the input end of the logistic regression model module.
10. A readable storage medium, characterized in that the readable storage medium has stored thereon a computer program which, when being executed by a processor, realizes the steps of the idler failure diagnosis method based on big data technology according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011437844.4A CN112660746B (en) | 2020-12-07 | 2020-12-07 | Roller fault diagnosis method and system based on big data technology and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011437844.4A CN112660746B (en) | 2020-12-07 | 2020-12-07 | Roller fault diagnosis method and system based on big data technology and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112660746A true CN112660746A (en) | 2021-04-16 |
| CN112660746B CN112660746B (en) | 2022-04-08 |
Family
ID=75401875
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011437844.4A Active CN112660746B (en) | 2020-12-07 | 2020-12-07 | Roller fault diagnosis method and system based on big data technology and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112660746B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115096375A (en) * | 2022-08-22 | 2022-09-23 | 启东亦大通自动化设备有限公司 | Carrier roller running state monitoring method and device based on carrier roller carrying trolley detection |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017136489A1 (en) * | 2016-02-03 | 2017-08-10 | Caspo, Llc | Smart cooking system that produces and uses hydrogen fuel |
| CN108706308A (en) * | 2018-05-29 | 2018-10-26 | 华电重工股份有限公司 | A kind of fault recognition method of conveyer, apparatus and system |
| CN110589405A (en) * | 2019-09-23 | 2019-12-20 | 北京华方智汇科技有限公司 | Roller abnormity detection system and method for belt conveyor inspection robot |
| CN111517092A (en) * | 2020-06-03 | 2020-08-11 | 太原理工大学 | Transmission belt tearing detection method |
| CN111573197A (en) * | 2020-04-22 | 2020-08-25 | 北京华能新锐控制技术有限公司 | Abnormal sound detection method for distributed belt conveyor |
| CN111661590A (en) * | 2020-06-08 | 2020-09-15 | 天地（常州）自动化股份有限公司 | Method for detecting tearing damage of conveying belt of mining belt conveyor |
-
2020
- 2020-12-07 CN CN202011437844.4A patent/CN112660746B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017136489A1 (en) * | 2016-02-03 | 2017-08-10 | Caspo, Llc | Smart cooking system that produces and uses hydrogen fuel |
| CN108706308A (en) * | 2018-05-29 | 2018-10-26 | 华电重工股份有限公司 | A kind of fault recognition method of conveyer, apparatus and system |
| CN110589405A (en) * | 2019-09-23 | 2019-12-20 | 北京华方智汇科技有限公司 | Roller abnormity detection system and method for belt conveyor inspection robot |
| CN111573197A (en) * | 2020-04-22 | 2020-08-25 | 北京华能新锐控制技术有限公司 | Abnormal sound detection method for distributed belt conveyor |
| CN111517092A (en) * | 2020-06-03 | 2020-08-11 | 太原理工大学 | Transmission belt tearing detection method |
| CN111661590A (en) * | 2020-06-08 | 2020-09-15 | 天地（常州）自动化股份有限公司 | Method for detecting tearing damage of conveying belt of mining belt conveyor |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115096375A (en) * | 2022-08-22 | 2022-09-23 | 启东亦大通自动化设备有限公司 | Carrier roller running state monitoring method and device based on carrier roller carrying trolley detection |
| CN115096375B (en) * | 2022-08-22 | 2022-11-04 | 启东亦大通自动化设备有限公司 | Carrier roller running state monitoring method and device based on carrier roller carrying trolley detection |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112660746B (en) | 2022-04-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112660745B (en) | Intelligent diagnosis method and system for carrier roller fault and readable storage medium | |
| CN112504673B (en) | Carrier roller fault diagnosis method, system and storage medium based on machine learning | |
| CN110867196B (en) | Machine equipment state monitoring system based on deep learning and voice recognition | |
| CN110405537B (en) | Method for establishing guide rail precision prediction model based on deep learning | |
| US20170161963A1 (en) | Method of identifying anomalies | |
| CN107844067B (en) | A kind of gate of hydropower station on-line condition monitoring control method and monitoring system | |
| CN111507376A (en) | Single index abnormality detection method based on fusion of multiple unsupervised methods | |
| CN102623009A (en) | Abnormal emotion automatic detection and extraction method and system on basis of short-time analysis | |
| CN112173636B (en) | Method for detecting faults of belt conveyor carrier roller by inspection robot | |
| CN113982850A (en) | Fan comprehensive health analysis method and system fusing high-frequency and low-frequency signals | |
| CN114462475A (en) | Unsupervised machine abnormal sound detection method and unsupervised machine abnormal sound detection device based on single classification algorithm | |
| CN110580492A (en) | Track circuit fault precursor discovery method based on small fluctuation detection | |
| CN112660746B (en) | Roller fault diagnosis method and system based on big data technology and storage medium | |
| CN114563150A (en) | Bridge health online detection module generation method, detection method, tool box and device | |
| CN116778964A (en) | Power transformation equipment fault monitoring system and method based on voiceprint recognition | |
| CN207992717U (en) | A kind of gate of hydropower station on-line condition monitoring system | |
| CN117150418B (en) | Transformer operation detection period formulation method and system based on state characteristic fault tree | |
| CN110879820A (en) | Industrial data processing method and device | |
| CN115424635B (en) | Cement plant equipment fault diagnosis method based on sound characteristics | |
| CN117435908A (en) | Multi-fault feature extraction method for rotary machine | |
| CN117391443A (en) | Dust removal equipment state monitoring and early warning method and system | |
| CN116720073A (en) | Abnormality detection extraction method and system based on classifier | |
| Olsson et al. | Case-based reasoning combined with statistics for diagnostics and prognosis | |
| CN113672859A (en) | Switch point machine fault acoustic diagnosis system | |
| CN108627242A (en) | Ring cold machine air leak rate of air curtain detection method based on the unilateral detection method of blind source separating |
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 |