CN117436846B - Equipment predictive maintenance method and system based on neural network - Google Patents
Equipment predictive maintenance method and system based on neural network Download PDFInfo
- Publication number
- CN117436846B CN117436846B CN202311736592.9A CN202311736592A CN117436846B CN 117436846 B CN117436846 B CN 117436846B CN 202311736592 A CN202311736592 A CN 202311736592A CN 117436846 B CN117436846 B CN 117436846B
- Authority
- CN
- China
- Prior art keywords
- equipment
- maintenance
- fault
- abnormal
- index
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000012423 maintenance Methods 0.000 title claims abstract description 169
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 48
- 238000000034 method Methods 0.000 title claims abstract description 45
- 230000002159 abnormal effect Effects 0.000 claims abstract description 169
- 238000012549 training Methods 0.000 claims description 16
- 238000004458 analytical method Methods 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 5
- 238000012544 monitoring process Methods 0.000 abstract description 5
- 230000006870 function Effects 0.000 description 6
- 238000004519 manufacturing process Methods 0.000 description 3
- 238000011158 quantitative evaluation Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 238000007796 conventional method Methods 0.000 description 2
- 238000007405 data analysis Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 238000011423 initialization method Methods 0.000 description 1
- 230000000977 initiatory effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011056 performance test Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/20—Administration of product repair or maintenance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Business, Economics & Management (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Biomedical Technology (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- Economics (AREA)
- Health & Medical Sciences (AREA)
- General Business, Economics & Management (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention provides a device predictive maintenance method and a device predictive maintenance system based on a neural network, which relate to the technical field of computers and comprise the following steps: collecting equipment sets of a factory equipment management system, and carrying out data acquisition on each piece of equipment to obtain equipment operation data; performing fault prediction to obtain abnormal equipment in the equipment set and fault types corresponding to the abnormal equipment, wherein the abnormal equipment is equipment with operation characteristic matching degree larger than preset characteristic matching degree; analyzing an instantaneous influence index when a fault occurs, a waiting influence index after the fault occurs and a maintenance influence index used for maintenance; outputting a fault risk index by combining a pre-trained fully-connected neural network; and carrying out maintenance grade identification, obtaining maintenance reminding information and sending the maintenance reminding information to related management personnel. The invention solves the technical problems that the traditional equipment maintenance method is often based on regular maintenance or emergency maintenance after the occurrence of faults, and lacks of real-time monitoring of equipment states and accurate fault prediction.
Description
Technical Field
The invention relates to the technical field of computers, in particular to a device predictive maintenance method and system based on a neural network.
Background
In modern manufacturing industry, realizing efficient equipment management is one of key targets pursued by enterprises, and the traditional equipment maintenance method has a plurality of problems, firstly, the traditional equipment maintenance method is often based on regular maintenance or emergency maintenance after failure occurs, and lacks real-time monitoring and accurate failure prediction on equipment states; second, conventional methods often fail to provide accurate analysis of various key indicators when equipment fails, resulting in unscientific and inefficient maintenance decisions; third, the conventional method lacks quantitative evaluation of the risk of failure of an abnormal device, resulting in difficulty in prioritizing and urgency of device failure.
Disclosure of Invention
The equipment predictive maintenance method based on the neural network aims at solving the technical problems that the traditional equipment maintenance method is often based on regular maintenance or emergency maintenance after the occurrence of faults, and real-time monitoring of equipment states and accurate fault prediction are lacked.
In view of the above, the present application provides a method and a system for device predictive maintenance based on neural networks.
In a first aspect of the present disclosure, a method for device predictive maintenance based on neural networks is provided, the method comprising: collecting a device set of a factory device management system, and collecting data of each device in the device set to obtain device operation data; performing fault prediction according to the equipment operation data to obtain abnormal equipment in the equipment set and fault types corresponding to the abnormal equipment, wherein the abnormal equipment is equipment with an operation characteristic matching degree larger than a preset characteristic matching degree, and the operation characteristic matching degree is the characteristic matching degree between the equipment operation data and the operation data when the equipment breaks down; according to the abnormal equipment and the fault types corresponding to the abnormal equipment, analyzing an instantaneous influence degree index when the abnormal equipment breaks down, a waiting influence degree index after the abnormal equipment breaks down and a maintenance influence degree index when the abnormal equipment is maintained; outputting a fault risk index of the abnormal equipment based on the instantaneous influence index, the waiting influence index and the maintenance influence index and a fully-connected neural network trained in advance; and carrying out maintenance grade identification on the abnormal equipment according to the fault risk index, obtaining maintenance reminding information and sending the maintenance reminding information to related management personnel.
In another aspect of the disclosure, a neural network-based device predictive maintenance system is provided, the system for use in the above method, the system comprising: the device comprises a data acquisition module, a data processing module and a control module, wherein the data acquisition module is used for acquiring a device set of a factory device management system and carrying out data acquisition on each device in the device set to obtain device operation data; the fault prediction module is used for carrying out fault prediction according to the equipment operation data to obtain abnormal equipment in the equipment set and fault types corresponding to the abnormal equipment, wherein the abnormal equipment is equipment with an operation characteristic matching degree larger than a preset characteristic matching degree, and the operation characteristic matching degree is the characteristic matching degree between the equipment operation data and the operation data when the equipment fails; the index analysis module is used for analyzing an instantaneous influence degree index when the abnormal equipment breaks down, a waiting influence degree index after the abnormal equipment breaks down and a maintenance influence degree index when the abnormal equipment is maintained according to the abnormal equipment and the fault types corresponding to the abnormal equipment; the risk index output module is used for outputting the fault risk index of the abnormal equipment based on the instantaneous influence degree index, the waiting influence degree index and the maintenance influence degree index and a fully-connected neural network trained in advance; and the maintenance grade identification module is used for carrying out maintenance grade identification on the abnormal equipment according to the fault risk index, obtaining maintenance reminding information and sending the maintenance reminding information to related management personnel.
One or more technical solutions provided in the present application have at least the following technical effects or advantages:
the equipment state can be monitored in real time by collecting equipment operation data and applying a neural network to conduct fault prediction, equipment faults can be accurately predicted, and the prospective and accuracy of maintenance are improved; by analyzing indexes such as instantaneous influence degree, waiting influence and the like of abnormal equipment, the instantaneous and continuous influence of equipment faults on production can be comprehensively known, and more comprehensive data support is provided for maintenance decision; by utilizing the fully-connected neural network, the fault risk index of the abnormal equipment can be output based on indexes such as instantaneous influence, waiting influence, maintenance influence and the like, so that the quantitative evaluation of the equipment fault priority is realized; according to the fault risk index, a maintenance grade identifier is distributed to the abnormal equipment, and maintenance reminding information is sent to related management staff, so that equipment maintenance can be performed in a targeted manner, and the scientificity and instantaneity of maintenance decision making are improved. In combination, the neural network-based equipment predictive maintenance method improves the efficiency, accuracy and timeliness of equipment maintenance by introducing data analysis and machine learning technologies.
The foregoing description is only an overview of the technical solutions of the present application, and may be implemented according to the content of the specification in order to make the technical means of the present application more clearly understood, and in order to make the above-mentioned and other objects, features and advantages of the present application more clearly understood, the following detailed description of the present application will be given.
Drawings
Fig. 1 is a schematic flow chart of a method for predictive maintenance of a device based on a neural network according to an embodiment of the present application;
fig. 2 is a schematic structural diagram of a device predictive maintenance system based on a neural network according to an embodiment of the present application.
Reference numerals illustrate: the system comprises a data acquisition module 10, a fault prediction module 20, an index analysis module 30, a risk index output module 40 and a maintenance grade identification module 50.
Detailed Description
The embodiment of the application solves the technical problems that the traditional equipment maintenance method is often based on the periodical maintenance or emergency maintenance after the occurrence of faults and lacks of real-time monitoring of equipment states and accurate fault prediction by providing the equipment predictive maintenance method based on the neural network.
Having described the basic principles of the present application, various non-limiting embodiments of the present application will now be described in detail with reference to the accompanying drawings.
Example 1
As shown in fig. 1, an embodiment of the present application provides a method for device predictive maintenance based on a neural network, the method including:
collecting a device set of a factory device management system, and collecting data of each device in the device set to obtain device operation data;
a connection is established with the plant device management system to enable access to the device data, and a device list is obtained to form a device set by connecting to the device management system. For each device in the device set, performing data acquisition operations including device status monitoring to obtain real-time status information of the device, such as temperature, pressure, current, etc.; acquiring a historical operation log of the equipment, wherein the historical operation log comprises past fault records, maintenance records and the like so as to establish the operation history of the equipment; the performance parameters of the device, such as readings of various sensors, operating rate, energy consumption, etc., are collected. And storing the collected equipment operation data, wherein the data provides a basis for subsequent fault prediction and maintenance decision.
Performing fault prediction according to the equipment operation data to obtain abnormal equipment in the equipment set and fault types corresponding to the abnormal equipment, wherein the abnormal equipment is equipment with an operation characteristic matching degree larger than a preset characteristic matching degree, and the operation characteristic matching degree is the characteristic matching degree between the equipment operation data and the operation data when the equipment breaks down;
further, performing fault prediction according to the equipment operation data includes:
establishing a device relation topology network of the device set;
acquiring a fault information base of each device in the device set, wherein the fault information base comprises historical fault types of each device and device operation characteristics corresponding to the historical fault types;
performing feature comparison on the equipment operation data based on the fault information base to obtain feature matching degrees respectively corresponding to each historical fault type;
and outputting the abnormal equipment in the equipment set when the characteristic matching degree is greater than or equal to the preset characteristic matching degree.
Relationships between devices, including connection, dependency, data transfer, etc., are explicitly defined, a device relationship topology model is built based on the relationship definition between devices, which may be a graph model in which devices are represented as nodes, relationships are represented as edges, and connections or dependencies between devices may be represented with directed edges or undirected edges.
The data about the relationships between devices, including connection information, dependencies, data transmission paths, etc., may be obtained from sources such as device manuals, system documents, sensor data, etc. Based on the collected data, a device relationship topology is constructed, which is a graphical structure describing the relationships between devices to provide a clear image of the relationships between devices.
A fault information base structure is established containing unique identifiers of the devices, historical fault types, time of occurrence of the faults, and operating characteristics of the devices associated with each of the historical fault types. Historical fault records for each device in the collection of devices are collected, including past types of faults, time stamps of the occurrence of the faults, and characteristic data of the operation of the device at the time of the occurrence of the faults, by device logging, maintenance records, reports, or other available historical data. And matching the equipment identifier, and associating the fault type in the history fault record with the corresponding equipment operation characteristic so as to establish the corresponding relation between the fault type and the equipment operation characteristic.
Comparing the operation data of the current equipment with the corresponding characteristics of each historical fault type, which can be realized by calculating similarity indexes or distance measures, such as cosine similarity, euclidean distance and the like, calculating the matching degree of the operation data of the equipment and the historical characteristics for each historical fault type, generating a score of the matching degree, reflecting the similarity degree of the current equipment and the historical fault type, and outputting the characteristic matching degree corresponding to each historical fault type.
According to actual conditions and specific requirements, a preset feature matching degree threshold is set, the threshold is a decision criterion and is used for judging whether the feature matching degree is high enough, and when the threshold is reached, the corresponding equipment is marked as abnormal. And comparing the characteristic matching degree with a preset characteristic matching degree threshold value for each device in the device set, if any characteristic matching degree is larger than or equal to the preset characteristic matching degree threshold value, identifying the device as abnormal device, and outputting the device information identified as abnormal for subsequent processing and decision.
Further, the fault information base further includes a historical fault probability corresponding to each historical fault type, including:
judging according to the feature matching degree to obtain the number of fault types greater than or equal to the preset feature matching degree;
if the number of the fault types is more than or equal to 2, identifying according to the fault information base, and acquiring a historical fault probability corresponding to the fault type which is more than or equal to the preset feature matching degree;
and identifying the historical fault probability corresponding to the fault type which is larger than or equal to the preset feature matching degree, and outputting the fault type corresponding to the party with the largest historical fault probability as the fault type corresponding to the abnormal equipment.
And counting the number of fault types which are greater than or equal to a preset characteristic matching degree threshold value, wherein the number reflects the number of fault types possibly existing in the current equipment.
If the number of the obtained fault types is greater than or equal to 2, the fault information base is queried for the 2 or more fault types, and the corresponding historical fault probabilities are obtained, wherein the historical fault probabilities reflect the occurrence frequency of the fault types in the past.
In the obtained historical fault probability, the fault type corresponding to the maximum probability value is identified, and the identified fault type with the maximum historical fault probability is output as the fault type corresponding to the abnormal equipment, so that under the condition that a plurality of fault types are found, the fault type most likely to occur is found through the identification of the historical fault probability, and is output as the fault type corresponding to the abnormal equipment, and guidance is provided for maintenance personnel, so that the maintenance personnel can process the abnormal equipment more specifically.
According to the abnormal equipment and the fault types corresponding to the abnormal equipment, analyzing an instantaneous influence degree index when the abnormal equipment breaks down, a waiting influence degree index after the abnormal equipment breaks down and a maintenance influence degree index when the abnormal equipment is maintained;
further, the method further comprises:
positioning an associated equipment set of the abnormal equipment according to the equipment relation topology network;
analyzing an instantaneous accompanying fault equipment set when the abnormal equipment in the associated equipment set breaks down according to the fault type corresponding to the abnormal equipment;
and obtaining the instantaneous influence degree index when the abnormal equipment fails according to the quantity of the accompanying equipment and the importance of the accompanying equipment of the instantaneous accompanying fault equipment set.
For the devices which are identified as abnormal, traversing from the abnormal devices in the device relation topological network, finding other devices which are directly or indirectly connected with the abnormal devices, and constructing an associated device set according to the traversing result, wherein the set comprises the devices which are directly connected with the abnormal devices and the devices which are indirectly connected with the abnormal devices through the other devices.
For the fault type of the abnormal equipment, analyzing an associated equipment set directly or indirectly connected with the abnormal equipment in the equipment relation topological network, and for each equipment in the associated equipment set, analyzing the possible instantaneous accompanying fault equipment when the abnormal equipment is in fault, for example, analyzing the dependency relationship, propagation paths and the like among the equipment, and constructing an instantaneous accompanying fault equipment set according to the analysis result, wherein the instantaneous accompanying fault equipment set comprises other equipment which is affected simultaneously when the abnormal equipment is in fault.
The number of companion devices in the instantaneous companion failure device set is counted, which indicates how many other devices will be affected when an abnormal device fails. For each companion device in the set of transient companion fault devices, an importance analysis is performed, including taking into account factors such as the function of the device, the status in the system, the impact on overall operation, and the like, to obtain companion device importance. And calculating an instantaneous influence degree index when the abnormal equipment fails by combining the quantity of the accompanying equipment and the importance of the accompanying equipment, for example, carrying out weighted summation on the quantity of the equipment and the importance of the accompanying equipment to obtain a numerical index comprehensively considering the quantity and the importance, and reflecting the instantaneous influence degree of the abnormal event on the system.
Further, according to the fault type corresponding to the abnormal equipment, connecting a maintainer management system, outputting a dispatch waiting time period and dispatch personnel positioning, and obtaining a waiting influence index after the abnormal equipment is in fault;
and acquiring a maintenance outage range, the number of the accompanying outage devices and the accompanying outage time according to the fault type corresponding to the abnormal device, and outputting a maintenance influence index when the abnormal device is maintained.
The method comprises the steps of connecting an application program interface with a maintainer management system, and initiating a query request to the maintainer management system according to the fault type of abnormal equipment to obtain information about the dispatch waiting time and the dispatch personnel positioning, wherein the dispatch waiting time is the time difference from dispatch generation time to the task receiving time of the maintainer, and the positioning information of the maintainer assigned to the abnormal equipment when the dispatch personnel positions the dispatch ticket can be the real-time position of the maintainer.
Analyzing the acquired data such as the waiting time of the dispatch, the positioning of the dispatch personnel and the like, including calculating the waiting time, calculating the arrival time of the dispatch personnel and the like, generating a waiting influence index according to analysis results, for example, summing two periods of time, presetting a waiting influence coefficient according to historical data, representing the influence degree of the waiting time on equipment faults, multiplying the summation result by the waiting influence coefficient, or carrying out exponential operation to obtain the waiting influence index, wherein the index represents the waiting influence degree of abnormal equipment after the faults occur.
Inquiring power-off information to a system according to the fault type of the abnormal equipment, wherein the power-off information comprises data such as maintenance power-off range, the number of the equipment accompanied by power-off, the duration of the accompanying power-off and the like, and the maintenance power-off range can be an area or an equipment list and represents the range influenced by the fault of the abnormal equipment; the companion outage device number is the number of devices affected during a maintenance outage; the accompanying power-off duration is a time difference from the start to the end of power-off.
And comprehensively analyzing the acquired data such as the maintenance outage range, the number of the equipment accompanied by outage, the duration of the accompanying outage and the like, for example, carrying out weighted summation on the data, and generating a maintenance influence degree index according to the result of the combined analysis, thereby reflecting the outage condition and the influence possibly related to the maintenance of the abnormal equipment.
Outputting a fault risk index of the abnormal equipment based on the instantaneous influence index, the waiting influence index and the maintenance influence index and a fully-connected neural network trained in advance;
further, the method further comprises:
collecting an instantaneous influence degree index sample, a waiting influence degree index sample and a maintenance influence degree index sample, and outputting the instantaneous influence degree index sample, the waiting influence degree index sample and the maintenance influence degree index sample as a training data set;
acquiring a pre-trained fully-connected neural network;
training the pre-trained fully-connected neural network according to the instantaneous impact index sample, the waiting impact index sample and the maintenance impact index sample to obtain a fault accompanying risk model;
and inputting the instantaneous influence degree index, the waiting influence degree index and the maintenance influence degree index into the fault accompanying risk model, and outputting a fault risk index of the abnormal equipment.
Sample data is collected from an actual environment, including collecting related data from sources such as a factory equipment management system, maintenance records, fault logs and the like, and an instantaneous influence degree index sample, a waiting influence degree index sample and a maintenance influence degree index sample are obtained. The acquired sample data is constructed into a training dataset, which may be a structured table, where each row represents a sample and each column represents a feature for training the neural network model.
The method comprises the steps of obtaining a pre-trained fully-connected neural network, wherein the pre-trained fully-connected neural network can be obtained from a public model library, a neural network model with other prediction functions is stored in the model library, the fully-connected neural network layer structure can be used after downloading, the fully-connected neural network layer structure is reserved to serve as the pre-trained fully-connected neural network to be output, a training data set (an instantaneous influence index sample, a waiting influence index sample and a maintenance influence index sample) is utilized to re-instantiate the model to realize the function of fault risk prediction, and the fault accompanying risk model is obtained. The weights and biases of the network are initialized, for example, using a random initialization method, and pre-trained weights may also be used. An appropriate loss function is established that will be used to measure the performance of the model on the training data, and an optimization algorithm is selected to minimize the loss function, such as random gradient descent, etc.
Loading a fully-connected neural network which is trained in advance, organizing an instantaneous influence index sample, a waiting influence index sample and a maintenance influence index sample into an expected input format of a model, ensuring that the input dimension is matched with an input layer of the fully-connected neural network, using a training data set with an adjusted format for training the fully-connected neural network, specifically dividing the training data set into a training set and a testing set according to a preset proportion, for example 8:2, inputting the prepared training set into the loaded fully-connected neural network for forward propagation, obtaining the prediction output of the model to the training set, carrying out model performance test by combining the testing set according to the prediction output of the model, carrying out iterative optimization on model parameters according to a test result, and obtaining a fault accompanying risk model when the model converges, for example, reaching the preset iteration times or meeting the preset accuracy.
And inputting the instantaneous influence degree index, the waiting influence degree index and the maintenance influence degree index into the fault accompanying risk model, analyzing the indexes according to the learned characteristics by the model, and outputting a fault risk index of the equipment.
And carrying out maintenance grade identification on the abnormal equipment according to the fault risk index, obtaining maintenance reminding information and sending the maintenance reminding information to related management personnel.
Setting a series of maintenance grades, wherein the grades can be divided according to different ranges of fault risk indexes, for example, three high, medium and low maintenance grades can be defined, abnormal equipment is divided into corresponding maintenance grades according to the set maintenance grade division standards, corresponding maintenance reminding information is generated based on the divided maintenance grades, the corresponding maintenance reminding information comprises equipment identification, fault risk indexes, maintenance grades and the like, and the generated maintenance reminding information is sent to related management staff, so that the management staff can timely take necessary maintenance measures to reduce the risk of equipment faults, and further improve the efficiency and timeliness of equipment maintenance.
Further, the method further comprises:
acquiring the number of abnormal devices in the device set;
judging whether the number of the abnormal devices is larger than or equal to the preset number of the abnormal devices or not;
if the number of the abnormal devices is greater than or equal to the number of the preset abnormal devices, connecting a maintainer management system to obtain a maintenance work order label of the maintainer management system, wherein the maintainer management system is a personnel management system for maintenance when the devices are in failure;
and carrying out maintenance synchronicity identification on the abnormal equipment according to the maintenance work order label, and outputting a priority decision result of the abnormal equipment by using a maintenance grade identifier if the maintenance synchronicity is not satisfied.
Counting the abnormal equipment set, obtaining the number of the abnormal equipment, and presetting a threshold according to actual conditions and specific requirements, namely presetting the number of the abnormal equipment. Comparing the calculated number of the abnormal devices with the preset number of the abnormal devices, and if the number of the abnormal devices is greater than or equal to the preset number of the abnormal devices, indicating that the number of the abnormal devices is more, and needing to take urgent maintenance action; otherwise, it may be selected to wait for more devices to become problematic or to perform regular maintenance.
When the number of the abnormal devices is larger than or equal to the number of the preset abnormal devices, the abnormal devices are connected with a maintainer management system, and the maintainer management system is a system for organizing maintainers to carry out a device maintenance process when the devices are in fault, and comprises functions of task dispatch, scheduling management, real-time positioning and the like.
Analyzing the acquired maintenance work order label comprises judging whether the current maintenance work order is related to abnormal equipment detected by the system, for example, comparing equipment information in the work order label with equipment information of the abnormal equipment, judging whether the equipment information is consistent with the equipment information of the abnormal equipment, judging whether maintenance personnel are processing the equipment information, and judging that the maintenance synchronism is met if the maintenance work order is processing the abnormal equipment.
If the maintenance synchronism is not met, based on information such as failure risk indexes of the abnormal devices, a maintenance grade identification is allocated to each abnormal device according to service requirements and system rules, and the identification can be determined based on factors such as the emergency degree and the influence degree of the device.
And outputting an abnormal equipment list with a maintenance grade identifier, taking the abnormal equipment list as a priority maintenance decision result of the system, making a maintenance plan according to the result, and determining which equipment needs to be maintained first, so as to carry out priority decision on the abnormal equipment.
In summary, the device predictive maintenance method and system based on the neural network provided by the embodiment of the application have the following technical effects:
1. the equipment state can be monitored in real time by collecting equipment operation data and applying a neural network to conduct fault prediction, equipment faults can be accurately predicted, and the prospective and accuracy of maintenance are improved;
2. by analyzing indexes such as instantaneous influence degree, waiting influence and the like of abnormal equipment, the instantaneous and continuous influence of equipment faults on production can be comprehensively known, and more comprehensive data support is provided for maintenance decision;
3. by utilizing the fully-connected neural network, the fault risk index of the abnormal equipment can be output based on indexes such as instantaneous influence, waiting influence, maintenance influence and the like, so that the quantitative evaluation of the equipment fault priority is realized;
4. according to the fault risk index, a maintenance grade identifier is distributed to the abnormal equipment, and maintenance reminding information is sent to related management staff, so that equipment maintenance can be performed in a targeted manner, and the scientificity and instantaneity of maintenance decision making are improved.
In combination, the neural network-based equipment predictive maintenance method improves the efficiency, accuracy and timeliness of equipment maintenance by introducing data analysis and machine learning technologies.
Example two
Based on the same inventive concept as the neural network-based device predictive maintenance method in the foregoing embodiment, as shown in fig. 2, the present application provides a neural network-based device predictive maintenance system, which includes:
the data acquisition module 10 is used for acquiring a device set of a factory device management system, and carrying out data acquisition on each device in the device set to obtain device operation data;
the fault prediction module 20 is configured to perform fault prediction according to the device operation data, so as to obtain an abnormal device in the device set and a fault type corresponding to the abnormal device, where the abnormal device is a device with an operation feature matching degree greater than a preset feature matching degree, and the operation feature matching degree is a feature matching degree between the device operation data and the operation data when the device fails;
the index analysis module 30 is configured to analyze, according to the abnormal device and the fault type corresponding to the abnormal device, an instantaneous impact index when the abnormal device fails, a waiting impact index after the abnormal device fails, and a maintenance impact index when the abnormal device is maintained;
the risk index output module 40 is configured to output a fault risk index of the abnormal device based on the instantaneous impact level index, the waiting impact level index, the maintenance impact level index, and a fully-connected neural network trained in advance;
the maintenance grade identification module 50 is configured to perform maintenance grade identification on the abnormal device according to the fault risk index, and obtain maintenance reminding information and send the maintenance reminding information to related management personnel.
Further, the fault prediction module further comprises the following operation steps:
establishing a device relation topology network of the device set;
acquiring a fault information base of each device in the device set, wherein the fault information base comprises historical fault types of each device and device operation characteristics corresponding to the historical fault types;
performing feature comparison on the equipment operation data based on the fault information base to obtain feature matching degrees respectively corresponding to each historical fault type;
and outputting the abnormal equipment in the equipment set when the characteristic matching degree is greater than or equal to the preset characteristic matching degree.
Further, the system also comprises a historical fault probability identification module for executing the following operation steps:
judging according to the feature matching degree to obtain the number of fault types greater than or equal to the preset feature matching degree;
if the number of the fault types is more than or equal to 2, identifying according to the fault information base, and acquiring a historical fault probability corresponding to the fault type which is more than or equal to the preset feature matching degree;
and identifying the historical fault probability corresponding to the fault type which is larger than or equal to the preset feature matching degree, and outputting the fault type corresponding to the party with the largest historical fault probability as the fault type corresponding to the abnormal equipment.
Further, the system further comprises a decision result acquisition module for executing the following operation steps:
acquiring the number of abnormal devices in the device set;
judging whether the number of the abnormal devices is larger than or equal to the preset number of the abnormal devices or not;
if the number of the abnormal devices is greater than or equal to the number of the preset abnormal devices, connecting a maintainer management system to obtain a maintenance work order label of the maintainer management system, wherein the maintainer management system is a personnel management system for maintenance when the devices are in failure;
and carrying out maintenance synchronicity identification on the abnormal equipment according to the maintenance work order label, and outputting a priority decision result of the abnormal equipment by using a maintenance grade identifier if the maintenance synchronicity is not satisfied.
Further, the system further comprises a fault risk index acquisition module to perform the following operation steps:
collecting an instantaneous influence degree index sample, a waiting influence degree index sample and a maintenance influence degree index sample, and outputting the instantaneous influence degree index sample, the waiting influence degree index sample and the maintenance influence degree index sample as a training data set;
acquiring a pre-trained fully-connected neural network;
training the pre-trained fully-connected neural network according to the instantaneous impact index sample, the waiting impact index sample and the maintenance impact index sample to obtain a fault accompanying risk model;
and inputting the instantaneous influence degree index, the waiting influence degree index and the maintenance influence degree index into the fault accompanying risk model, and outputting a fault risk index of the abnormal equipment.
Further, the system further comprises an instantaneous influence index acquisition module for executing the following operation steps:
positioning an associated equipment set of the abnormal equipment according to the equipment relation topology network;
analyzing an instantaneous accompanying fault equipment set when the abnormal equipment in the associated equipment set breaks down according to the fault type corresponding to the abnormal equipment;
and obtaining the instantaneous influence degree index when the abnormal equipment fails according to the quantity of the accompanying equipment and the importance of the accompanying equipment of the instantaneous accompanying fault equipment set.
Further, the system further comprises a maintenance impact index acquisition module for executing the following operation steps:
according to the fault type corresponding to the abnormal equipment, connecting a maintainer management system, outputting the dispatch waiting time and dispatch personnel positioning, and obtaining waiting influence indexes after the abnormal equipment is in fault;
and acquiring a maintenance outage range, the number of the accompanying outage devices and the accompanying outage time according to the fault type corresponding to the abnormal device, and outputting a maintenance influence index when the abnormal device is maintained.
From the foregoing detailed description of the method for predicting maintenance of a device based on a neural network, it will be apparent to those skilled in the art that the system for predicting maintenance of a device based on a neural network in this embodiment is described more simply for the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, and the relevant points refer to the description of the method section.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (7)
1. A method for predictive maintenance of a device based on a neural network, the method comprising:
collecting a device set of a factory device management system, and collecting data of each device in the device set to obtain device operation data;
performing fault prediction according to the equipment operation data to obtain abnormal equipment in the equipment set and fault types corresponding to the abnormal equipment, wherein the abnormal equipment is equipment with an operation characteristic matching degree larger than a preset characteristic matching degree, and the operation characteristic matching degree is the characteristic matching degree between the equipment operation data and the operation data when the equipment breaks down;
according to the abnormal equipment and the fault types corresponding to the abnormal equipment, analyzing an instantaneous influence degree index when the abnormal equipment breaks down, a waiting influence degree index after the abnormal equipment breaks down and a maintenance influence degree index when the abnormal equipment is maintained;
outputting a fault risk index of the abnormal equipment based on the instantaneous influence index, the waiting influence index and the maintenance influence index and a fully-connected neural network trained in advance;
performing maintenance grade identification on the abnormal equipment according to the fault risk index, obtaining maintenance reminding information and sending the maintenance reminding information to related management personnel;
the method further comprises the steps of:
collecting an instantaneous influence degree index sample, a waiting influence degree index sample and a maintenance influence degree index sample, and outputting the instantaneous influence degree index sample, the waiting influence degree index sample and the maintenance influence degree index sample as a training data set;
acquiring a pre-trained fully-connected neural network;
training the pre-trained fully-connected neural network according to the instantaneous impact index sample, the waiting impact index sample and the maintenance impact index sample to obtain a fault accompanying risk model;
and inputting the instantaneous influence degree index, the waiting influence degree index and the maintenance influence degree index into the fault accompanying risk model, and outputting a fault risk index of the abnormal equipment.
2. The method of claim 1, wherein the fault prediction is based on the plant operational data, the method comprising:
establishing a device relation topology network of the device set;
acquiring a fault information base of each device in the device set, wherein the fault information base comprises historical fault types of each device and device operation characteristics corresponding to the historical fault types;
performing feature comparison on the equipment operation data based on the fault information base to obtain feature matching degrees respectively corresponding to each historical fault type;
and outputting the abnormal equipment in the equipment set when the characteristic matching degree is greater than or equal to the preset characteristic matching degree.
3. The method of claim 2, wherein the fault information base further includes historical fault probabilities for each of the historical fault types, the method comprising:
judging according to the feature matching degree to obtain the number of fault types greater than or equal to the preset feature matching degree;
if the number of the fault types is more than or equal to 2, identifying according to the fault information base, and acquiring a historical fault probability corresponding to the fault type which is more than or equal to the preset feature matching degree;
and identifying the historical fault probability corresponding to the fault type which is larger than or equal to the preset feature matching degree, and outputting the fault type corresponding to the party with the largest historical fault probability as the fault type corresponding to the abnormal equipment.
4. The method of claim 2, wherein the method further comprises:
acquiring the number of abnormal devices in the device set;
judging whether the number of the abnormal devices is larger than or equal to the preset number of the abnormal devices or not;
if the number of the abnormal devices is greater than or equal to the number of the preset abnormal devices, connecting a maintainer management system to obtain a maintenance work order label of the maintainer management system, wherein the maintainer management system is a personnel management system for maintenance when the devices are in failure;
and carrying out maintenance synchronicity identification on the abnormal equipment according to the maintenance work order label, and outputting a priority decision result of the abnormal equipment by using a maintenance grade identifier if the maintenance synchronicity is not satisfied.
5. The method of claim 2, wherein the method further comprises:
positioning an associated equipment set of the abnormal equipment according to the equipment relation topology network;
analyzing an instantaneous accompanying fault equipment set when the abnormal equipment in the associated equipment set breaks down according to the fault type corresponding to the abnormal equipment;
and obtaining the instantaneous influence degree index when the abnormal equipment fails according to the quantity of the accompanying equipment and the importance of the accompanying equipment of the instantaneous accompanying fault equipment set.
6. The method of claim 5, wherein the maintainer management system is connected according to the fault type corresponding to the abnormal equipment, and the waiting duration and the personnel positioning of the dispatch are output to obtain the waiting influence index after the abnormal equipment is in fault;
and acquiring a maintenance outage range, the number of the accompanying outage devices and the accompanying outage time according to the fault type corresponding to the abnormal device, and outputting a maintenance influence index when the abnormal device is maintained.
7. A neural network based device predictive maintenance system for implementing the neural network based device predictive maintenance method of any of claims 1-6, comprising:
the device comprises a data acquisition module, a data processing module and a control module, wherein the data acquisition module is used for acquiring a device set of a factory device management system and carrying out data acquisition on each device in the device set to obtain device operation data;
the fault prediction module is used for carrying out fault prediction according to the equipment operation data to obtain abnormal equipment in the equipment set and fault types corresponding to the abnormal equipment, wherein the abnormal equipment is equipment with an operation characteristic matching degree larger than a preset characteristic matching degree, and the operation characteristic matching degree is the characteristic matching degree between the equipment operation data and the operation data when the equipment fails;
the index analysis module is used for analyzing an instantaneous influence degree index when the abnormal equipment breaks down, a waiting influence degree index after the abnormal equipment breaks down and a maintenance influence degree index when the abnormal equipment is maintained according to the abnormal equipment and the fault types corresponding to the abnormal equipment;
the risk index output module is used for outputting the fault risk index of the abnormal equipment based on the instantaneous influence degree index, the waiting influence degree index and the maintenance influence degree index and a fully-connected neural network trained in advance;
and the maintenance grade identification module is used for carrying out maintenance grade identification on the abnormal equipment according to the fault risk index, obtaining maintenance reminding information and sending the maintenance reminding information to related management personnel.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311736592.9A CN117436846B (en) | 2023-12-18 | 2023-12-18 | Equipment predictive maintenance method and system based on neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311736592.9A CN117436846B (en) | 2023-12-18 | 2023-12-18 | Equipment predictive maintenance method and system based on neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117436846A CN117436846A (en) | 2024-01-23 |
| CN117436846B true CN117436846B (en) | 2024-03-05 |
Family
ID=89551843
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311736592.9A Active CN117436846B (en) | 2023-12-18 | 2023-12-18 | Equipment predictive maintenance method and system based on neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117436846B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117666462B (en) * | 2024-01-31 | 2024-05-17 | 成都苔岑智能设备有限公司 | PLC product standardization control system based on independent allocation of multiple parameters |
| CN118534883B (en) * | 2024-07-29 | 2024-10-01 | 成都秦川物联网科技股份有限公司 | Intelligent equipment fault diagnosis method, system and medium based on industrial Internet of things |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015176565A1 (en) * | 2014-05-22 | 2015-11-26 | 袁志贤 | Method for predicting faults in electrical equipment based on multi-dimension time series |
| JP2017199128A (en) * | 2016-04-26 | 2017-11-02 | 株式会社東芝 | Restoration planning management device, restoration planning program and restoration planning method |
| CN108008332A (en) * | 2017-11-29 | 2018-05-08 | 国网山东省电力公司电力科学研究院 | A kind of new energy Remote testing device method for diagnosing faults based on data mining |
| US11676014B1 (en) * | 2019-02-19 | 2023-06-13 | Viaduct, Inc. | Systems, media, and methods applying machine learning to telematics data to generate vehicle fingerprint |
| WO2023158044A1 (en) * | 2022-02-17 | 2023-08-24 | 주식회사 디엠테크컨설팅 | Method for facility malfunction predictive maintenance using artificial intelligence based on empirical shared knowledge |
-
2023
- 2023-12-18 CN CN202311736592.9A patent/CN117436846B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015176565A1 (en) * | 2014-05-22 | 2015-11-26 | 袁志贤 | Method for predicting faults in electrical equipment based on multi-dimension time series |
| JP2017199128A (en) * | 2016-04-26 | 2017-11-02 | 株式会社東芝 | Restoration planning management device, restoration planning program and restoration planning method |
| CN108008332A (en) * | 2017-11-29 | 2018-05-08 | 国网山东省电力公司电力科学研究院 | A kind of new energy Remote testing device method for diagnosing faults based on data mining |
| US11676014B1 (en) * | 2019-02-19 | 2023-06-13 | Viaduct, Inc. | Systems, media, and methods applying machine learning to telematics data to generate vehicle fingerprint |
| WO2023158044A1 (en) * | 2022-02-17 | 2023-08-24 | 주식회사 디엠테크컨설팅 | Method for facility malfunction predictive maintenance using artificial intelligence based on empirical shared knowledge |
Non-Patent Citations (1)
| Title |
|---|
| 张丽娜.中国优秀硕士学位论文全文数据库.2023,(S2),全文. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117436846A (en) | 2024-01-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117436846B (en) | Equipment predictive maintenance method and system based on neural network | |
| CN112817280A (en) | Implementation method for intelligent monitoring alarm system of thermal power plant | |
| CN105677791B (en) | For analyzing the method and system of the operation data of wind power generating set | |
| CN116932523B (en) | Platform for integrating and supervising third party environment detection mechanism | |
| CN116992346A (en) | Enterprise production data processing system based on artificial intelligence big data analysis | |
| CN117743909A (en) | Heating system fault analysis method and device based on artificial intelligence | |
| CN115204536A (en) | Building equipment fault prediction method, device, equipment and storage medium | |
| CN118244709A (en) | Associated fault equipment and influence factor acquisition method, device and exception handling system | |
| CN113657747B (en) | Intelligent assessment system for enterprise safety production standardization level | |
| CN118211882B (en) | Product quality management system and method based on big data | |
| CN113825162B (en) | Method and device for positioning fault reasons of telecommunication network | |
| CN117572837B (en) | Intelligent power plant AI active operation and maintenance method and system | |
| Lin et al. | Design and implementation of a CPS‐based predictive maintenance and automated management platform | |
| CN116720983A (en) | Power supply equipment abnormality detection method and system based on big data analysis | |
| CN117131364A (en) | Rolling bearing process detection integration method and system | |
| CN114019946B (en) | Method and device for processing monitoring data of industrial control terminal | |
| CN115765182A (en) | Digital twinning method and device based on time sequence database | |
| CN115438093A (en) | Power communication equipment fault judgment method and detection system | |
| CN115598459A (en) | Power failure prediction method for 10kV feeder line fault of power distribution network | |
| Gultekin et al. | Predictive Analytics in IoT and CPS: Enhancing Industrial Machinery Reliability through Sensor Data-Driven Remaining Useful Life Estimation | |
| Lei et al. | Application of distributed machine learning model in fault diagnosis of air preheater | |
| CN118229036B (en) | Operation and maintenance management system and method for smart city | |
| CN117892212B (en) | Distributed heterogeneous energy station situation awareness monitoring method | |
| CN115269740A (en) | Method for monitoring and predicting faults of optical module of server | |
| CN110263811A (en) | A kind of equipment running status monitoring method and system based on data fusion |
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 |