CN118445106B - On-line state monitoring method and health management system for coal machine equipment - Google Patents
On-line state monitoring method and health management system for coal machine equipment Download PDFInfo
- Publication number
- CN118445106B CN118445106B CN202410606443.9A CN202410606443A CN118445106B CN 118445106 B CN118445106 B CN 118445106B CN 202410606443 A CN202410606443 A CN 202410606443A CN 118445106 B CN118445106 B CN 118445106B
- Authority
- CN
- China
- Prior art keywords
- fault
- health management
- data
- coal machine
- equipment
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
- G06F11/0703—Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation
- G06F11/079—Root cause analysis, i.e. error or fault diagnosis
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
- G06F11/0703—Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation
- G06F11/0706—Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation the processing taking place on a specific hardware platform or in a specific software environment
- G06F11/0736—Error or fault processing not based on redundancy, i.e. by taking additional measures to deal with the error or fault not making use of redundancy in operation, in hardware, or in data representation the processing taking place on a specific hardware platform or in a specific software environment in functional embedded systems, i.e. in a data processing system designed as a combination of hardware and software dedicated to performing a certain function
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/3003—Monitoring arrangements specially adapted to the computing system or computing system component being monitored
- G06F11/3013—Monitoring arrangements specially adapted to the computing system or computing system component being monitored where the computing system is an embedded system, i.e. a combination of hardware and software dedicated to perform a certain function in mobile devices, printers, automotive or aircraft systems
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/3055—Monitoring arrangements for monitoring the status of the computing system or of the computing system component, e.g. monitoring if the computing system is on, off, available, not available
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/34—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
- G06F11/3457—Performance evaluation by simulation
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/02—Knowledge representation; Symbolic representation
- G06N5/022—Knowledge engineering; Knowledge acquisition
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Quality & Reliability (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Computer Hardware Design (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the technical field of equipment online monitoring, in particular to an online state monitoring method and a health management system of coal machine equipment, comprising the following steps: connecting a visual simulation platform to simulate the coal machine fault and constructing a fault knowledge graph; on-line monitoring and data feedback recognition faults, a single health management column is determined, the health management system is combined to carry out coal machine health management, the technical problems that short delay exists in the process of detecting and collecting data to carry out fault diagnosis analysis, the risks of damage and production interruption of coal machine equipment caused by delay diagnosis are increased are solved, visual simulation is introduced to carry out simulation on the coal machine equipment, meanwhile, fault occurrence simulation analysis is carried out, the coal machine equipment management and fault diagnosis are combined, a deep-learning misplacement full-connection network is combined, a fault knowledge map is constructed, the subtle changes of the running state of equipment are comprehensively captured, and therefore the fault sources are rapidly located, and the technical effects of damage and production interruption risks of the coal machine equipment caused by delay diagnosis are reduced.
Description
Technical Field
The invention relates to the technical field of equipment online monitoring, in particular to an online state monitoring method and a health management system of coal machine equipment.
Background
The monitoring and maintenance of the coal machine equipment depend on the periodic inspection and emergency treatment after the fault occurs, but the fault discovery corresponding to the periodic inspection is not timely. Along with the development of large-scale continuous operation of coal machine equipment, the reliability and safety of the equipment become key factors for determining the production efficiency.
Due to the technical conditions, on-line state monitoring of coal equipment only provides basic operation parameters, the identification precision of a single-mapping fault monitoring management mode is limited, and in addition, data are commonly transmitted from detection acquisition to carry out fault diagnosis analysis, short delay exists in the process, and the fault early warning and processing lag is caused by short delay diagnosis, so that the risks of equipment damage and production interruption are increased.
In summary, the prior art has the technical problems that the process from detection and acquisition to fault diagnosis and analysis of data has short delay, and the risk of damage and production interruption of coal machine equipment caused by delay diagnosis is increased.
Disclosure of Invention
The application provides an on-line state monitoring method and a health management system for coal machine equipment, and aims to solve the technical problems that short delay exists in the process of detecting and collecting data to diagnosing and analyzing faults in the prior art, and the risk of damage and production interruption of the coal machine equipment caused by delay diagnosis is increased.
In view of the above problems, the technical scheme for realizing the application is as follows:
In one aspect of the application, an on-line state monitoring method of coal machine equipment is provided, wherein the method comprises the following steps: connecting a visual simulation platform, and constructing a fault simulation module based on the coal machine equipment, wherein the fault simulation module is provided with a coal machine operation model; introducing a coal machine fault factor, performing fault occurrence simulation analysis and system fault tolerance analysis based on the fault simulation module, and integrating and constructing a fault knowledge graph; constructing a fault decision model based on the fault knowledge graph, wherein the fault decision model is of a staggered full-connection network structure; performing on-line state monitoring of the coal machine, performing fuzzy decision and data feedback of the equipment end of multidimensional detection sensing data, performing equipment fault identification by combining the fault decision model, and determining equipment fault distribution, wherein the equipment fault distribution comprises real-time faults and predicted faults; performing hierarchical binarization processing on the equipment fault distribution by taking a characteristic threshold value and occurrence probability as references, and determining a health management list; based on the health management list, the health management of the coal machine equipment is carried out by combining a health management system.
In another aspect of the present application, a health management system is provided, where the health management system is applied to a coal machine device, the health management system is a distributed management system, and the system includes: constructing N micro-managers based on fault type management, constructing a first upper computer layer based on management cooperativity, constructing a second upper computer layer based on management dimension, and connecting the N micro-managers, the first upper computer layer and the second upper computer layer based on a quick interaction channel to obtain the health management system; and transmitting the health management list to the health management system to perform health management of the coal machine equipment.
In summary, the one or more technical schemes provided by the application realize that visual simulation is introduced to simulate the coal machine equipment, meanwhile, fault generation simulation analysis is carried out, the coal machine equipment management and fault diagnosis are combined, a deep learning dislocation full-connection network is combined, a fault knowledge graph is constructed, and the subtle change of the running state of the equipment is comprehensively captured, so that the fault source is rapidly positioned, the risks of damage and production interruption of the coal machine equipment caused by delay diagnosis are reduced, real-time faults and potential faults are accurately distinguished, and the technical effect of fault diagnosis accuracy is improved.
Drawings
FIG. 1 is a schematic flow chart of a method for monitoring the on-line state of coal machine equipment;
FIG. 2 is a schematic flow chart of a fault decision model in a dislocated fully-connected network structure in an on-line state monitoring method of coal machine equipment;
fig. 3 is a schematic structural diagram of a health management system according to the present application.
Detailed Description
Example 1
The application is specifically described below with reference to the accompanying drawings, and as shown in fig. 1, the application provides an on-line state monitoring method of coal machine equipment, wherein the method comprises the following steps:
S1: connecting a visual simulation platform, and constructing a fault simulation module based on the coal machine equipment, wherein the fault simulation module is provided with a coal machine operation model; s2: and introducing a coal machine fault factor, performing fault occurrence simulation analysis and system fault tolerance analysis based on the fault simulation module, and integrating and constructing a fault knowledge graph.
The visual simulation platform is connected, has strong data processing capability and graphic display function, can support modeling and simulation of complex equipment, further develops necessary software interfaces, ensures that actual running data of coal machine equipment can be smoothly transmitted to the simulation platform, and control signals of the visual simulation platform can be fed back to the equipment (if closed-loop test is required); physical parameters, running conditions and environmental factors related to the coal machine equipment are configured on the visual simulation platform, so that a foundation is laid for subsequent simulation.
Constructing a fault simulation module, specifically, constructing a coal machine operation model, wherein a visual simulation platform is utilized to construct a mathematical model according to a mechanical structure, an electrical system, a control system and the like of coal machine equipment, and the mathematical model can reflect the working state of the coal machine equipment under normal and abnormal conditions; integrating fault elements, namely embedding various potential fault factors such as abrasion, overload, electrical short circuit and the like into a coal machine operation model, wherein each fault factor corresponds to a specific model parameter change or logic judgment rule; visual interface design, including, design interactive interface, observe the running state of the model of the coal machine intuitively, including normal workflow and trouble simulation scene.
Based on the fault simulation module, performing fault occurrence simulation analysis and system fault tolerance analysis, specifically setting fault situations, such as damage of specific components, sensor failure and the like, and simulating the influence of different fault situations under different working conditions; starting a simulation program, enabling a coal machine model to run simulation under different fault conditions, and collecting simulation data, wherein the simulation data comprises equipment response, performance index change and the like; and (3) analyzing simulation results, and evaluating the recovery capability and stability of the coal machine system in the face of various faults and the influence on the production efficiency, including the effectiveness of redundant design, fault isolation capability and the like.
Integrating and constructing a fault knowledge graph, and specifically, arranging all data obtained from fault simulation, including fault types, trigger conditions, performance characteristics, influence ranges and the like; classifying the fault simulation data by using a data analysis method such as a clustering algorithm, classifying similar faults, and determining a hierarchical structure of fault type-fault source-fault characteristics; creating a fault knowledge graph comprising a fault type, a fault source, a correlation between fault features and a decision level of feature threshold-occurrence probability; and carrying out interlayer association by using a graph theory method, generating an initial graph, and carrying out co-pointing disambiguation treatment to ensure the accuracy and consistency of the graph.
In the practical application process, the method further comprises the step of designing a structure of the extensible fault knowledge graph, so that the fault knowledge graph can be dynamically updated along with the accumulation of new data, and comprises the steps of screening invalid identification data, generating a new sub-graph, and positioning and updating based on the correlation between spectrums.
In summary, the efficient simulation and diagnosis tool for the faults of the coal feeder equipment is provided, a continuous learning and optimizing mechanism is established, and the continuity and effectiveness of the on-line state monitoring management of the coal feeder equipment are ensured.
S3: constructing a fault decision model based on the fault knowledge graph, wherein the fault decision model is of a staggered full-connection network structure; s4: and carrying out on-line state monitoring of the coal machine, carrying out fuzzy decision and data feedback of the equipment end of the multidimensional detection sensing data, carrying out equipment fault identification by combining the fault decision model, and determining equipment fault distribution, wherein the equipment fault distribution comprises real-time faults and predicted faults.
Constructing a fault decision model of a staggered full-connection network structure, and specifically, sorting out historical fault data and normal operation data required by a training model based on fault characteristics, types and related thresholds in a fault knowledge graph; designing a staggered full-connection network structure, wherein the staggered full-connection network structure comprises a real-time decision unit and a trend prediction unit, and the real-time decision unit is responsible for processing the current state data of the equipment and performing instant fault judgment; the trend prediction unit predicts possible faults in the future according to the historical data.
The dislocated fully-connected network structure comprises a real-time decision unit and a trend prediction unit, and further comprises the real-time decision unit: the internal construction feature extraction block is used for extracting key features of the multidimensional detection sensing data, and the fault decision block performs instant fault identification based on the extracted key features; trend prediction unit: the data storage block is used for storing historical data, and the trend prediction block is used for carrying out fault trend analysis based on the data; and a lateral interaction channel is arranged between the real-time decision unit and the trend prediction unit, so that bidirectional flow of data and information is allowed.
The model is trained by utilizing the existing fault data set, and the network weight and bias are adjusted, so that the model can accurately identify the known fault type and has a certain generalization capability to predict the unknown fault.
Carrying out on-line state monitoring of the coal machine, carrying out equipment end fuzzy decision and data feedback of multidimensional detection sensing data, and particularly, deploying multidimensional sensors such as vibration, temperature, pressure and the like on the coal machine equipment, and monitoring the running state of the equipment in real time; the equipment end is provided with a fuzzy logic processor, the collected original data is subjected to preliminary processing and fuzzy decision analysis, and possible fuzzy modes in the data are identified according to preset fuzzy rules, so that the false alarm rate is reduced; and returning the key data and the decision result after fuzzy processing in a wireless or wired mode.
Performing equipment fault identification by combining the fault decision model, determining equipment fault distribution, specifically, performing instant fault judgment by combining a real-time decision unit of the fault decision model after receiving the fuzzy key data and the decision result, and determining whether a real-time fault and the type thereof occur according to the output of the real-time decision unit of the fault decision model; the trend prediction unit of the fault decision model utilizes long-term data sequence analysis to monitor corresponding real-time data in combination with the online state of the coal machine, and predicts future fault trend and time window; and (3) integrating real-time decision and prediction results, determining fault distribution of equipment, including faults existing in current real time and faults expected to occur in the future, and classifying according to severity and fault probability.
Through the steps, closed-loop management from data acquisition and analysis to decision making is realized, and the operation efficiency and maintenance management level of the coal machine equipment are improved.
S5: performing hierarchical binarization processing on the equipment fault distribution by taking a characteristic threshold value and occurrence probability as references, and determining a health management list; s6: based on the health management list, the health management of the coal machine equipment is carried out by combining a health management system.
Determining a characteristic threshold and occurrence probability, specifically, collecting multidimensional monitoring data (such as vibration, temperature, current and the like) of coal machine equipment in normal operation and fault states, and carrying out statistical analysis; setting a feature threshold value for each monitoring feature based on the data analysis result, wherein the feature threshold value is a critical value for measuring feature abnormality, and if the feature threshold value exceeds the critical value, the feature is considered to have potential faults; and calculating the probability of faults of different fault types under specific conditions by utilizing the historical fault data and combining a machine learning algorithm.
Performing hierarchical binarization processing on the equipment fault distribution by taking a characteristic threshold and occurrence probability as references, specifically mapping equipment monitoring data to corresponding characteristic thresholds, and judging whether each characteristic exceeds the threshold; and (3) one layer of binarization: for each feature, if the measured value exceeds a set threshold value, marking as 1 (abnormal), otherwise, marking as 0 (normal), and forming a layer of binary data of the fault feature; binarization of two layers: and adding a layer of binary data according to the attribution relation of the fault types, and determining which fault types have the probability exceeding a preset probability threshold value by combining the probability of the fault types, wherein the probability is marked as 1, and otherwise, the probability is marked as 0.
In the practical application process, the method further comprises one-layer binarization, two-layer binarization, three-layer binarization, four-layer binarization and … …, and the method is used for carrying out hierarchical binarization processing to realize comprehensive evaluation of fault types.
Determining a health management list, and integrating the effective fault data determined after hierarchical binarization processing to form the health management list, wherein each item represents a fault type or fault combination and the emergency and urgency of the fault type or fault combination; determining a management time limit, i.e. a time range in which action must be taken, for each fault according to the degree of influence of the fault and the deterioration efficiency; and according to the management time limit and the fault severity, the faults in the health management single column are prioritized, so that the high-risk faults are ensured to be processed preferentially.
Based on the health management list, the health management of the coal machine equipment is carried out by combining the health management system, and specifically, the health management list is transmitted to the health management system, and the health management system automatically schedules resources according to the content of the list, schedules maintenance tasks, realizes distributed management on different faults, and ensures efficient response.
In the actual application process, the method further comprises the steps of continuously updating the fault knowledge graph and the decision model according to the actual operation and maintenance condition and the newly-appearing fault data, and dynamically adjusting the health management strategy to realize self-optimization and learning of the system.
Through the steps, real-time faults of the coal machine equipment are timely identified and dealt with, future faults and occurrence probability thereof are predicted and prevented in advance, and stable operation of the equipment is effectively guaranteed.
Further, the integration construction fault knowledge graph comprises the following steps:
Clustering and integrating the fault occurrence simulation data, and determining a one-dimensional hierarchical map for fault identification, wherein the one-dimensional hierarchical map at least comprises a map layer of fault type-fault source-fault characteristics; determining a two-dimensional hierarchical map of fault occurrence decision based on the system fault tolerance, wherein the two-dimensional hierarchical map at least comprises a feature threshold-map layer of occurrence probability; fitting the one-dimensional hierarchical map and the two-dimensional hierarchical map, performing interlayer correlation based on fault correlation, and determining an initialization map; and performing co-fingering disambiguation on the initialized map to generate the fault knowledge map.
Acquiring fault occurrence simulation data, specifically, generating coal machine equipment operation data under various fault conditions through a fault simulation module; noise and inconsistent data points are removed, ensuring data quality.
Clustering and integrating the simulated data of the faults, determining a one-dimensional hierarchical map of fault identification, specifically classifying the simulated data of the faults by using a clustering algorithm (such as K-means, DBSCAN and the like), and classifying similar faults into one class according to fault expression; and determining the association relation among various faults, and establishing a hierarchical structure of fault type-fault source-fault characteristics. For example, bearing overheating is classified as a type of failure, the source of which may be poor lubrication, corresponding features including increased temperature and increased vibration.
Based on the system fault tolerance, determining a two-dimensional hierarchical map for fault occurrence decision, specifically, performing system fault tolerance analysis, evaluating the tolerance of coal machine equipment under different fault conditions, and determining the maximum bearing capacity of the system; setting a characteristic threshold value for each key monitoring characteristic according to the historical data, and distinguishing normal states from abnormal states; and analyzing the fault occurrence probability when the specific fault characteristics exceed the threshold by using a statistical analysis or machine learning model, and constructing a two-dimensional map of the characteristic threshold-occurrence probability.
Fitting the one-dimensional hierarchical map and the two-dimensional hierarchical map, carrying out interlayer association based on fault correlation, determining an initialization map, specifically, carrying out interlayer association on the one-dimensional fault characteristic map and the two-dimensional decision map through fault types to form a multi-dimensional view, reflecting the relation between the fault characteristics and the fault occurrence probability, and constructing a preliminary fault knowledge map frame which comprises all known fault types, sources, characteristics and occurrence probability and threshold values thereof.
Performing co-fingering disambiguation on the initialized map to generate the fault knowledge map, specifically, checking and solving redundant or conflict information possibly existing in the map, and ensuring the accuracy of the map as in the case that the same fault feature is erroneously associated with different fault types; and finally, formally generating a fault knowledge graph, and clearly displaying key information such as association, characteristic threshold value, occurrence probability and the like among faults.
Through the steps, the constructed fault knowledge graph becomes the basis of follow-up on-line state monitoring and health management, and the support system rapidly and accurately identifies faults, predicts fault trend and guides maintenance decision.
Further, as shown in fig. 2, the fault decision model is a dislocated fully-connected network structure, and the method of the present application includes:
Constructing a real-time decision unit and a trend prediction unit which are distributed in parallel based on the fault knowledge graph; the real-time decision unit comprises a feature extraction block and a fault decision block, the trend prediction unit comprises a data storage block and a trend prediction block, and a lateral interaction channel is established between the fault decision block and the data storage block; the real-time decision unit and the trend prediction unit operate relatively independently, and data can be interacted.
Based on the fault knowledge graph, constructing a parallel distributed real-time decision unit, specifically, planning the structure and function of the real-time decision unit according to the information of the fault type, the feature, the threshold value, the occurrence probability and the like which are arranged in the fault knowledge graph, wherein the real-time decision unit comprises a feature extraction block and a fault decision block, and further comprises a feature extraction block: developing an algorithm or automatically extracting meaningful features from sensor data by utilizing the prior art (such as convolutional neural network CNN in deep learning), wherein the extracted features can directly reflect the current state of coal machine equipment; and (3) constructing a fault decision block: based on the extracted features, comprehensive analysis is performed on the features by using technologies such as a full connection layer or a multi-layer perceptron (MLP), whether faults exist currently or not is judged, and the fault type is identified.
Based on the fault knowledge graph, a trend prediction unit in parallel distribution is constructed, specifically, the structure and the function of the trend prediction unit are planned according to the information of the fault type, the feature, the threshold, the occurrence probability and the like which are arranged in the fault knowledge graph, a lateral interaction channel is established between the fault decision block and the data storage block, and further, the data storage block: an efficient data storage structure is designed for storing decision results and historical monitoring data generated by the real-time decision unit, so that long-term trend analysis is facilitated; trend prediction block: based on the stored data, the development trend of faults is analyzed by using methods such as time sequence analysis, a cyclic neural network (RNN) or a long and short time memory network (LSTM) and the like, and the probability of faults occurring in the future is predicted.
The fault decision block and the data storage block are provided with lateral interaction channels, and the lateral interaction channels are specific: creating a data exchange channel between a fault decision block of a real-time decision unit and a data storage block of a trend prediction unit, allowing specific information between the fault decision block and the data storage block to circulate, and further, laterally interacting the interactive flow of the channel: the real-time result of the real-time decision unit is stored in the data storage block of the trend prediction unit and fed back to the trend prediction block of the trend prediction unit as input to more accurately predict the future fault trend, and meanwhile, the analysis result of trend prediction can be used for correcting the threshold value or strategy of the real-time decision unit.
The real-time decision unit and the trend prediction unit operate relatively independently, and data can be interacted, and specifically, the real-time decision unit and the trend prediction unit operate relatively independently: the real-time decision unit and the trend prediction unit are ensured to operate independently, the real-time unit is focused on diagnosis of the current state, and the trend unit focuses on identification of a long-term mode; the data may interact: through the designed interactive interface, the data flow between the two units is dynamic and efficient, the analysis result of the real-time data can influence the input of trend prediction in time, and the prediction analysis can also back feed the accuracy of real-time decision.
In the actual application process, the method further comprises the steps of constructing a training and verification data set based on the historical fault record and the simulation data; the whole network is trained by utilizing algorithms such as gradient descent, back propagation and the like, so that the model can accurately perform fault identification and trend prediction; and evaluating the performance of the model by the methods of cross validation, AUC-ROC curve and the like, and optimizing the model parameters according to the evaluation result.
Through the steps, the built dislocation full-connection network structure can efficiently integrate real-time monitoring and trend prediction, and provides comprehensive health management support for coal machine equipment.
Further, after the integration construction of the fault knowledge graph, the fault knowledge graph has expandability, and the method comprises the following steps:
Screening invalid identification data based on the data storage block, wherein the invalid identification data meets a frequency threshold; generating a new sub-map based on the invalid identification data; based on the correlation between spectrums, determining the distribution position of the new-added sub-spectrum, and updating the fault knowledge spectrum; and updating and learning the fault decision model based on the updated fault knowledge graph.
Collecting data based on the data storage block, specifically, monitoring coal machine equipment on line to obtain multidimensional detection sensing data; and storing the multidimensional detection sensing data into a data storage block, performing preliminary fault identification attempt, and recording the case of incorrect identification or incorrect identification.
Screening invalid identification data based on the data storage block, specifically, setting a frequency threshold (failure part without identification) according to historical data analysis to distinguish occasional identification errors from frequent unidentified failure modes; within the data-holding block, those frequently incorrectly identified fault data are identified and marked as invalid identification data.
Generating a new sub-map based on the invalid identification data, and specifically, deeply analyzing the characteristics of the invalid identification data, including fault types, characteristic expressions, occurrence frequencies and the like; then, a new sub-graph, namely a new sub-graph, is created, and the new sub-graph is designed for a special failure mode which is not identified or is newly identified, and comprises a new failure type-failure source-failure characteristic relation.
Based on the updated fault knowledge graph, updating and learning the fault decision model, and specifically, analyzing the relevance between the new sub-graph and the existing fault knowledge graph, including causal relationship, similarity and the like between faults; based on the relation between spectrums, determining the optimal position of the new sub-spectrum in the overall fault knowledge spectrum so as to ensure the logic and integrity of the spectrum; integrating the new sub-graph into the original fault knowledge graph to form an updated fault knowledge graph; then, based on the updated fault knowledge graph, adjusting the structure or parameters of the fault decision model to ensure that the model can understand and utilize newly added knowledge; retraining the fault decision model by utilizing a training set containing new and old fault data so as to learn and master new fault recognition rules; and evaluating the updated performance of the model through the test set, so that the model can still maintain or improve the recognition accuracy and the prediction capability after expansion.
Through the steps, the fault knowledge graph can continuously absorb newly-appearing fault information, so that the timeliness and the accuracy of the fault decision model are maintained.
Further, the device fault distribution is subjected to hierarchical binarization processing, the method comprises the following steps:
Identifying the equipment fault distribution, mapping fault characteristics and the characteristic threshold value, and determining a plurality of mapping groups; traversing the mapping groups, and carrying out threshold positive and negative metering to determine a layer of binary data, wherein the negative direction of the threshold is characteristic abnormality, the binary value is 1, and the positive direction of the threshold is 0; traversing the first-layer binary data, carrying out addition of the first-layer binary data and binary processing based on occurrence probability of the fault type based on attribution relation of the fault type, and determining effective fault data; and integrating the effective fault data to generate the health management list.
Identifying the equipment fault distribution, mapping fault characteristics and the characteristic threshold value, determining a plurality of mapping groups, and specifically, finishing all known equipment fault characteristics, wherein the equipment fault characteristics relate to various parameters such as temperature, vibration, noise and the like; setting a feature threshold for each fault feature, wherein the feature threshold is a critical value for judging and measuring feature abnormality, and different fault features correspond to different thresholds because of different influence degrees and abnormal manifestations on equipment health; and matching the fault characteristics with the corresponding characteristic threshold values, constructing a plurality of mapping groups, and preparing a foundation for subsequent binarization processing.
Traversing the plurality of mapping groups, performing positive and negative metering on a threshold value, determining a layer of binary data, specifically checking each mapping group one by one, and comparing an actually measured fault characteristic value with a preset characteristic threshold value; if the feature value exceeds the threshold (i.e., the feature is abnormal), the flag is 1; if the feature is within the threshold (normal), marked as 0, a layer of binary data is generated, intuitively reflecting which features are in an abnormal state.
Traversing the first layer of binary data, carrying out addition of the first layer of binary data and two layers of binarization processing based on occurrence probability of the fault type based on the attribution relation of the fault type, and determining effective fault data, specifically, organizing the first layer of binary data based on the fault type, namely determining a plurality of characteristics under the same fault type to carry out attribution of the fault type in an aggregation mode; for each fault type, adding binary data of all the features belonging to the type to obtain an abnormal indication value; according to the addition result and the positive and negative indexes (1 represents abnormality and 0 represents normal), the comprehensive occurrence probability of each fault type is calculated, specifically, the probability can be estimated through the proportion of the abnormal feature quantity to the total feature quantity, the fault type meeting a certain probability threshold is considered as a valid fault, namely, the fault type meeting the certain probability threshold is identified as 1 (represents abnormality).
Integrating the effective fault data to generate the health management single column, and specifically, identifying the effective fault data with occurrence probability reaching or exceeding a set threshold value based on the result of two-layer binarization processing; integrating the valid fault data into a single health management column, wherein each item of data represents an identified fault type and occurrence probability thereof, or is directly identified as 1 or 0; the health management list is then used to guide the formulation of maintenance strategies, such as prioritizing high probability faults, or scheduling maintenance based on the degree of fault impact and deterioration rate.
Through the steps, the fault distribution of the equipment is converted into binary data which is easy to understand and operate, so that further health management decisions and actions are facilitated.
Further, after determining the health management list, the method of the application comprises the following steps:
Identifying the health management list, and determining management time limit, wherein the management time limit is determined based on fault influence degree and fault deterioration efficiency; based on the management time limit, carrying out priority ordering of fault management, and determining a management sequence, wherein the management sequence is provided with a timestamp mark; and carrying out coal machine equipment health management based on the health management system by combining the management sequences.
Identifying the health management list, determining a management time limit, and specifically identifying each fault entry including a fault type, a severity of a characteristic abnormality (fault influence degree) and a speed of fault deterioration (fault deterioration efficiency) based on the health management list; evaluating the influence degree of each fault, and considering the potential influence of the influence on the overall performance, the production efficiency and the safety of the equipment; analyzing the speed of fault deterioration, judging the time frame that the fault can quickly lead to more serious consequences, and setting a reasonable management time limit for each fault, namely taking measures in the time to prevent the fault deterioration.
Based on the management time limit, carrying out priority ordering of fault management, determining a management sequence, and particularly constructing a priority index according to the fault influence degree and the management time limit, wherein repair cost and required resources may also need to be considered; using this index to order the faults in the health management list, ensuring that the faults with the greatest threat to the operation of the equipment and urgent are ranked in front; a time stamp is allocated to each fault management task to indicate ideal starting or finishing time, so that resources are helped to be scheduled and progress is monitored; forming a management sequence based on the sequencing result, wherein the management sequence comprises a fault repair task, a responsible person, expected starting and ending time and the like; according to the management sequence, the technical team, spare part supply and downtime are coordinated, ensuring planned implementation.
Carrying out coal machine equipment health management based on the health management system by combining the management sequence, specifically, monitoring equipment states in real time by using the health management system, and ensuring that fault prevention and repair work is executed according to the management sequence; during execution, feedback information is collected, management effects are evaluated, and management sequences and policies are adjusted as necessary to cope with newly occurring problems or changes.
In the actual application process, the method also comprises the step of periodically reviewing the health management process and the result, and evaluating the effect of the management measures, including the fault solving speed, the resource utilization rate and the like; according to the evaluation result, continuously iterating and optimizing the flow of the health management system, and improving the fault prediction accuracy and management efficiency; and by combining new fault cases and solutions, updating a fault knowledge graph and a decision model, and keeping the advancement and adaptability of the system.
Through the steps, the coal machine equipment is effectively and healthily managed in time, the failure rate is reduced, and the running stability of the coal machine equipment is improved.
In summary, the beneficial effects of the embodiment of the application are as follows:
1. By constructing a fault simulation module and combining the coal machine operation model to perform fault occurrence simulation analysis and system fault tolerance analysis, a fault knowledge graph is formed, deep understanding and accurate recognition of the fault mode of the coal machine equipment are realized, and false alarm and missing report are reduced.
2. By utilizing an on-line monitoring technology and multidimensional sensing data, the state of equipment is tracked in real time, fault distribution is processed through hierarchical binarization, the data processing flow is simplified, potential faults are found in time, and the response speed of fault decision is improved.
3. And a fault decision model of a staggered full-connection network structure is introduced, and a single health management column is combined, so that the key faults are ensured to be processed preferentially, intelligent evaluation and maintenance strategy planning of the equipment state are realized, faults are prevented in advance, and the maintenance cost is reduced.
4. Due to the fact that the health management list is identified, management time limit is determined, and the management time limit is determined based on fault influence degree and fault deterioration efficiency; based on management time limit, carrying out priority ordering of fault management, and determining a management sequence, wherein the management sequence is provided with a timestamp mark; and carrying out coal machine equipment health management based on the health management system by combining the management sequences. The coal machine equipment is effectively and healthily managed in time, the failure rate is reduced, and the running stability of the coal machine equipment is improved.
Example two
As shown in fig. 3, the present application provides a health management system, wherein the health management system is applied to a coal machine equipment, the health management system is a distributed management system, and the system includes:
Constructing N micro-managers based on fault type management, constructing a first upper computer layer based on management cooperativity, constructing a second upper computer layer based on management dimension, and connecting the N micro-managers, the first upper computer layer and the second upper computer layer based on a quick interaction channel to obtain the health management system; and transmitting the health management list to the health management system to perform health management of the coal machine equipment.
The health management system architecture is designed, and preferably N micro-managers are designed and constructed according to different fault types possibly occurring in coal machine equipment. Each micro-manager focuses on monitoring, diagnosis and treatment advice for one or a class of specific faults, including setting corresponding fault identification algorithms, threshold settings and response policies.
In a feasible implementation manner, a first upper computer layer is constructed and is mainly responsible for coordinating information communication and task allocation among micro-managers, so that management cooperativity is ensured, and the first upper computer layer needs to have an efficient task scheduling algorithm and can dynamically adjust the workload of the micro-manager according to the current equipment state and fault priority.
In one possible implementation, a second upper computer layer is constructed, and based on management dimensions such as geographical distribution, equipment types or division of maintenance team, the second upper computer layer is constructed, and is mainly responsible for resource management and policy formulation at a macroscopic level, so that efficient operation and resource optimization of the whole health management system are ensured.
Establishing a rapid interaction channel, preferably, establishing a set of efficient and stable communication protocol for data exchange among the micro-manager, the first upper computer layer and the second upper computer layer; the deployment network infrastructure, including servers, routers and necessary network security equipment, ensures the stability and security of the fast interaction channel.
Based on a rapid interaction channel, the N micro-managers, the first upper computer layer and the second upper computer layer are connected to obtain the health management system, preferably, based on an on-line state monitoring method of coal machine equipment, a health management list is determined, and the health management list comprises information such as fault identification results, management time limit, priority and the like and is packaged into a data format suitable for transmission; transmitting the health management list to a health management system (the integrity, timeliness and safety of data need to be ensured) through a quick interaction channel; after the health management system receives the health management list, the second upper computer layer firstly analyzes the data and distributes tasks to the corresponding first upper computer layer according to the fault type and the management dimension; the first upper computer layer further refines the task and transmits specific operation instructions to the corresponding micropipe managers, and the micropipe managers execute fault diagnosis or guide on-site maintenance personnel to intervene; and continuously collecting equipment state data by the micro-manager in the execution process, and feeding back the equipment state data to the upper computer layer through the quick interaction channel so as to adjust the management strategy or optimize follow-up actions in real time.
In a possible implementation manner, the method further comprises the step of periodically performing performance evaluation on the health management system, wherein the performance evaluation comprises indexes such as response time, processing efficiency, fault processing accuracy and the like; the fault model is continuously learned and optimized by utilizing the data in the fault processing process, the fault knowledge graph is updated, and the intelligent level of the system is improved; according to the business requirement and the technical development, the micro-manager and the upper computer layer are subjected to software upgrading, or a new micro-manager is added to cope with the newly-occurring fault type or management requirement.
Further, the system of the present application comprises:
The health management system has expandability, and performs optimization learning based on a preset management period on the health management system based on updating and new addition of the micro manager and the management upper computer.
In a feasible implementation mode, the health management system has expandability, is optimized, and analyzes whether the aspects of processing capacity, fault coverage rate, response speed and the like of the existing health management system meet the current and expanded requirements at regular intervals or according to actual operation conditions, and identifies functions of a micro-manager and an upper computer layer which need to be newly added or updated; based on the expansion demand analysis, a detailed expansion plan is formulated, including the types of micro-managers needing to be newly added, the optimization direction of the management upper computer, the expected performance improvement targets and the like.
Based on the updating and the new addition of the micro-manager and the management upper computer, preferably, according to an expansion scheme, the micro-manager is used for covering the newly-appearing fault type or optimizing the existing functions, and meanwhile, the management upper computer layer is designed or upgraded to enhance the collaborative management capability and the data processing capability, and it is required that all newly-developed functions are subjected to strict tests to ensure the stability and the compatibility; and on the premise of ensuring that the normal operation of the existing system is not influenced, integrating the newly developed micropipe manager and the updated upper computer layer into the health management system step by step, and configuring corresponding communication protocols and interfaces.
The health management system is subjected to optimization learning based on a preset management period, and preferably, the preset management period (such as weekly, monthly or quarterly) is set as a trigger point of the optimization learning of the health management system according to the running characteristics of equipment, the occurrence frequency of faults and the condition of system resources; when each management period is finished, key indexes such as equipment operation data, fault processing records, response time and the like in the period of time are automatically collected; and synchronously evaluating the performance of the health management system, and adjusting fault threshold values, decision model parameters, resource allocation strategies and the like based on evaluation results so as to improve the overall efficiency of the system. The performance of the coal machine is continuously improved through optimization learning based on the preset management period, the change of the equipment operation environment and the increase of the management requirement are adapted, and the long-term stable and efficient operation of the coal machine equipment is ensured.
In summary, any of the steps may be stored as computer instructions or programs in a non-limiting computer memory and may be called for recognition by a non-limiting computer processor, where unnecessary limitations are not made.
Further, the foregoing technical solutions only represent preferred technical solutions of the embodiments of the present application, and some modifications may be made by those skilled in the art to some parts of them to represent novel principles of the embodiments of the present application, and it is obvious that those skilled in the art may make various changes and modifications to the present application without departing from the scope of the application.
Claims (4)
1. An on-line state monitoring method of coal machine equipment, which is characterized by comprising the following steps:
Connecting a visual simulation platform, and constructing a fault simulation module based on the coal machine equipment, wherein the fault simulation module is provided with a coal machine operation model;
introducing a coal machine fault factor, performing fault occurrence simulation analysis and system fault tolerance analysis based on the fault simulation module, and integrating and constructing a fault knowledge graph;
constructing a fault decision model based on the fault knowledge graph, wherein the fault decision model is of a staggered full-connection network structure;
performing on-line state monitoring of the coal machine, performing fuzzy decision and data feedback of the equipment end of multidimensional detection sensing data, performing equipment fault identification by combining the fault decision model, and determining equipment fault distribution, wherein the equipment fault distribution comprises real-time faults and predicted faults;
performing hierarchical binarization processing on the equipment fault distribution by taking a characteristic threshold value and occurrence probability as references, and determining a health management list;
based on the health management list, carrying out health management on the coal machine equipment by combining a health management system;
Wherein, the integration construction fault knowledge graph comprises:
Clustering and integrating the fault occurrence simulation data, and determining a one-dimensional hierarchical map for fault identification, wherein the one-dimensional hierarchical map at least comprises a map layer of fault type-fault source-fault characteristics;
determining a two-dimensional hierarchical map of fault occurrence decision based on the system fault tolerance, wherein the two-dimensional hierarchical map at least comprises a feature threshold-map layer of occurrence probability;
fitting the one-dimensional hierarchical map and the two-dimensional hierarchical map, performing interlayer correlation based on fault correlation, and determining an initialization map;
Performing co-fingering disambiguation on the initialized map to generate the fault knowledge map, wherein the co-fingering disambiguation refers to checking and solving redundant or conflict information possibly existing in the map, and comprises the situation that the same fault characteristic is erroneously associated with different fault types;
Wherein, the fault decision model is a staggered full-connection network structure, and comprises:
constructing a real-time decision unit and a trend prediction unit which are distributed in parallel based on the fault knowledge graph;
The real-time decision unit comprises a feature extraction block and a fault decision block, the trend prediction unit comprises a data storage block and a trend prediction block, a lateral interaction channel is established between the fault decision block and the data storage block, the data storage block is used for storing decision results generated by the real-time decision unit and historical monitoring data, and the trend prediction block is used for predicting the probability of faults occurring in the future based on the stored data to analyze the development trend of the faults;
The lateral interaction flow corresponding to the lateral interaction channel comprises that an instant result of a real-time decision unit is stored in a data storage block of the trend prediction unit and fed back to the trend prediction block of the trend prediction unit as input, and meanwhile, an analysis result of trend prediction is also used for correcting a threshold value or strategy of the real-time decision unit, and the real-time decision unit and the trend prediction unit operate relatively independently and can interact with each other;
after the integration construction of the fault knowledge graph, the fault knowledge graph has expandability, and comprises the following steps:
screening invalid identification data based on the data storage block, wherein the invalid identification data meets a frequency threshold;
generating a new sub-map based on the invalid identification data;
based on the correlation between spectrums, determining the distribution position of the new-added sub-spectrum, and updating the fault knowledge spectrum;
Based on the updated fault knowledge graph, updating and learning the fault decision model;
Wherein after determining the health management list, the method comprises the following steps:
Identifying the health management list, and determining management time limit, wherein the management time limit is determined based on fault influence degree and fault deterioration efficiency;
Based on the management time limit, carrying out priority ordering of fault management, and determining a management sequence, wherein the management sequence is provided with a timestamp mark;
and carrying out coal machine equipment health management based on the health management system by combining the management sequences.
2. The method of claim 1, wherein said performing hierarchical binarization processing on said equipment failure distribution comprises:
identifying the equipment fault distribution, mapping fault characteristics and the characteristic threshold value, and determining a plurality of mapping groups;
traversing the mapping groups, and carrying out threshold positive and negative metering to determine a layer of binary data, wherein the negative direction of the threshold is characteristic abnormality, the binary value is 1, and the positive direction of the threshold is 0;
Traversing the first-layer binary data, carrying out addition of the first-layer binary data and binary processing based on occurrence probability of the fault type based on attribution relation of the fault type, and determining effective fault data;
and integrating the effective fault data to generate the health management list.
3. A health management system for performing the method of any one of claims 1-2 for health management of coal equipment, the health management system being a distributed management system, the system comprising:
Constructing N micro-managers based on fault type management, constructing a first upper computer layer based on management cooperativity, constructing a second upper computer layer based on management dimension, and connecting the N micro-managers, the first upper computer layer and the second upper computer layer based on a quick interaction channel to obtain the health management system;
and transmitting the health management list to the health management system to perform health management of the coal machine equipment.
4. The system of claim 3, wherein the health management system has extensibility, and performs optimization learning based on a preset management period on the basis of updating and adding of the micro manager and the management host computer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410606443.9A CN118445106B (en) | 2024-05-16 | 2024-05-16 | On-line state monitoring method and health management system for coal machine equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410606443.9A CN118445106B (en) | 2024-05-16 | 2024-05-16 | On-line state monitoring method and health management system for coal machine equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118445106A CN118445106A (en) | 2024-08-06 |
| CN118445106B true CN118445106B (en) | 2024-10-18 |
Family
ID=92333163
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410606443.9A Active CN118445106B (en) | 2024-05-16 | 2024-05-16 | On-line state monitoring method and health management system for coal machine equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118445106B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN119199527B (en) * | 2024-11-28 | 2025-03-07 | 比博斯特(江苏)汽车科技有限公司 | Fault diagnosis method for drive motor |
| CN119283048B (en) * | 2024-12-13 | 2025-05-20 | 天津赛威工业技术有限公司 | Industrial robot safety monitoring method and system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114265955A (en) * | 2021-12-23 | 2022-04-01 | 彩虹无线（北京）新技术有限公司 | Fault prediction method and device based on diagnosis fault code evolution map |
| CN115759258A (en) * | 2022-11-16 | 2023-03-07 | 广东电网有限责任公司 | Relation adjusting method of power grid equipment entity based on knowledge graph |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101545215B1 (en) * | 2013-10-30 | 2015-08-18 | 삼성에스디에스 주식회사 | system and method for automatically manageing fault events of data center |
| CN116069874A (en) * | 2023-01-06 | 2023-05-05 | 重庆长安汽车软件科技有限公司 | Fault location method, device, equipment and storage medium based on knowledge graph |
-
2024
- 2024-05-16 CN CN202410606443.9A patent/CN118445106B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114265955A (en) * | 2021-12-23 | 2022-04-01 | 彩虹无线（北京）新技术有限公司 | Fault prediction method and device based on diagnosis fault code evolution map |
| CN115759258A (en) * | 2022-11-16 | 2023-03-07 | 广东电网有限责任公司 | Relation adjusting method of power grid equipment entity based on knowledge graph |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118445106A (en) | 2024-08-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR102730303B1 (en) | AI-based railway signal system abnormality detection and diagnosis method | |
| CN118096131B (en) | Operation and maintenance inspection method based on electric power scene model | |
| CN118445106B (en) | On-line state monitoring method and health management system for coal machine equipment | |
| CN117853078B (en) | Unmanned aerial vehicle-based dynamic safety alarm system and method for construction site of building engineering | |
| CN118915566A (en) | Heating ventilation equipment abnormity on-line monitoring system based on Internet of things | |
| US10902368B2 (en) | Intelligent decision synchronization in real time for both discrete and continuous process industries | |
| CN117930815A (en) | Wind turbine generator remote fault diagnosis method and system based on cloud platform | |
| CN117787711A (en) | Arch bridge construction single-end management system based on big data | |
| CN117055502A (en) | Intelligent control system based on Internet of things and big data analysis | |
| CN112926257A (en) | Reciprocating natural gas compressor fault diagnosis system and diagnosis method | |
| CN118011990B (en) | Industrial data quality monitoring and improving system based on artificial intelligence | |
| CN117666367A (en) | Remote control method and system for underground coal mine machinery | |
| CN112328703A (en) | A device and method for diagnosing health status of numerically controlled equipment based on incremental learning | |
| CN118016268A (en) | Intelligent medical system based on artificial intelligence | |
| CN113395182B (en) | Intelligent network equipment management system and method with fault prediction | |
| CN118446671B (en) | Integrated intelligent operation and maintenance control method and system | |
| CN118896427B (en) | Block chain-based air source heat pump monitoring and regulating method and system | |
| CN118732645B (en) | A fault early warning method and system for giant hydropower units | |
| CN118396595A (en) | Vehicle maintenance method and device, electronic equipment and storage medium | |
| CN120087774A (en) | Comprehensive service system and method for intelligent security protection | |
| CN118966796A (en) | A mine safety risk investigation method and system based on intelligent Internet of Things devices | |
| CN119204412A (en) | An automated data collection and intelligent analysis system | |
| CN120067941A (en) | Early warning method and system for power protection system abnormality | |
| CN119849987A (en) | Frame production information management method and system | |
| CN119828611A (en) | Robot data processing method and system based on industrial Internet of things |
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 |