Data uploading fault monitoring system based on intelligent early warning

Abstract

The invention relates to the field of computer science and technology, in particular to a data uploading fault monitoring system based on intelligent early warning, which comprises a target data acquisition module, an early warning feature extraction module, an alarm rule setting module, a linkage alarm module, a network construction module, an alarm analysis module and a management database, the system carries out data acquisition by defining an early warning target, extracts each early warning feature of the early warning target data, determines alarm points according to historical early warning target data of each device, obtains an association relationship between the alarm points of each device, constructs a single alarm point rule and a linkage alarm rule, and generates alarm content. According to the method, the alarm features and the inherent features of the early-warning target are subjected to correlation analysis, the alarm early-warning network is constructed, the risk condition of the specified early-warning target is further evaluated, the essence and the influence range of the alarm event can be known more comprehensively, and thus the cause of the fault and the possible consequence can be judged more accurately.

Description

Technical Field

[0001] The present invention relates to the field of computer science and technology, and more specifically, to a data upload fault monitoring system based on intelligent early warning. Background Art

[0002] Mine exploitation operations play a crucial role in the energy supply field, continuously delivering important energy for industrial development and social life. However, the mine production environment is harsh, filled with various complex risk factors. The operating environment is complex and challenging, with limited underground mining space, variable geological conditions, and safety hazards such as gas, coal dust, and water hazards constantly threatening production safety and the lives and health of personnel.

[0003] With the increase in the depth of mine exploitation and the continuous expansion of production scale, the number of equipment increases and the system complexity improves. Traditional manual inspection and simple monitoring methods are no longer sufficient to meet the requirements of modern mine safety production. According to incomplete statistics, mine accidents caused by equipment failures and untimely safety warnings still occur frequently in recent years, resulting in huge casualties and economic losses.

[0004] Regarding the intelligent early warning and fault monitoring of mines, there are still some deficiencies in the existing technologies. For example, the Chinese patent with the application number 202410379441.0 discloses a method for fault diagnosis and prevention of mine hoists based on a knowledge graph. This solution uses advanced knowledge graph technology, combines the historical fault data and maintenance records of mine hoists, and constructs a knowledge graph including fault diagnosis, fault prevention, and maintenance recommendations. This graph can analyze the abnormal conditions of mine hoists in real time, identify potential fault signs, and provide targeted maintenance suggestions, which not only improves the operation efficiency and safety of mine hoists but also provides an innovative solution for the intelligent maintenance of mine hoists.

[0005] However, this solution has the following deficiencies: The alarm rule setting of this solution is simple, with less description of the alarm rule settings. It does not set different levels of data thresholds as alarm rules from historical data, considering multiple factors such as parameter types and equipment types, making it difficult to comprehensively and accurately judge fault situations. At the same time, it basically does not involve the content related to linkage alarms. The knowledge graph solution only relies on single knowledge graph queries and maintenance measure recommendations, which may have limitations when dealing with situations where multiple devices or multiple fault factors are interrelated. Summary of the Invention

[0006] In order to overcome the deficiencies in the background art, the embodiments of the present invention provide a data upload fault monitoring system based on intelligent early warning, which can effectively solve the problems involved in the above background art.

[0007] The object of the present invention can be achieved by the following technical solutions: The present invention provides a data upload fault monitoring system based on intelligent early warning, including: a target data acquisition module, which is used to define an early warning target and collect data for the early warning target.

[0008] An early warning feature extraction module, which is used to preprocess the acquired early warning target data and perform dimensionality reduction operations, and extract each early warning feature of the early warning target data.

[0009] An alarm rule setting module, which is used to determine alarm points according to the historical early warning target data of each device, obtain the correlation relationship between the alarm points of each device, and construct a single alarm point rule.

[0010] A linkage alarm module, which is used to draw a network topology diagram and set linkage alarm rules, priorities and trigger conditions based on various correlations.

[0011] A network construction module, which is used to generate alarm content, perform correlation analysis on the alarm features and the inherent features of the early warning target, and construct an alarm early warning network.

[0012] An alarm analysis module, which is used to evaluate the risk situation of a specified early warning target according to the alarm early warning network.

[0013] A management database, which is used to store early warning target data, historical early warning target data, alarm points and threshold information, network topology diagram information, and alarm early warning network data.

[0014] Preferably, the target data acquisition module is connected to each device in the mine through a sensor network. The sensor network uses a communication protocol to transmit data and can automatically identify and screen data interfaces of different types of devices, so as to collect early warning target data of various devices.

[0015] At the same time, the target data acquisition module supports multiple acquisition methods. For the acquisition of the third-party data upload status, Party A coordinates with the third party to provide relevant interfaces.

[0016] Preferably, the specific operation method of the early warning feature extraction module is: calculate the covariance matrix between features according to the preprocessed early warning target data, perform eigenvalue decomposition on the covariance matrix to obtain each eigenvalue and the corresponding eigenvector.

[0017] Arrange the eigenvalues in descending order, and select the eigenvectors corresponding to the eigenvalues with the cumulative contribution rate reaching the set value as the main components.

[0018] Project the early warning target data onto the main components to obtain the dimensionality-reduced early warning target data, and extract the corresponding eigenvector and record it as the early warning feature.

[0019] Preferably, the specific operation method of the alarm rule setting module is as follows: S1. Collect the equipment type, parameter category, and installation location of each device in the mine, and obtain the historical warning target data and corresponding timestamps of each device in the mine, where the historical warning target data includes normal data and fault data.

[0020] S2. Select alarm points by analyzing historical warning target data, sort out the correlation relationships between alarm points, and classify alarm points according to parameter types.

[0021] S3. Set data thresholds for each level as alarm rules according to the historical warning target data of each device in the mine. The data thresholds for each level include a first-level threshold, a second-level threshold, and a third-level threshold.

[0022] Preferably, the specific analysis method of the linkage alarm module is as follows: Sort out the physical connection and data transmission relationships between each device in the mine, draw a network topology diagram of the mine equipment in combination with the correlation relationships between alarm points, and divide the equipment importance levels according to the association conditions of the devices in the network topology diagram. The equipment importance levels include first-level equipment, second-level equipment, and third-level equipment, so as to obtain the equipment importance levels of the corresponding devices for each alarm point.

[0023] Based on the correlation relationships between alarm points in the network topology diagram, set linkage alarm rules based on multiple correlations respectively, and set priorities and trigger conditions for the linkage alarm rules.

[0024] The priority of the linkage alarm rule is based on the equipment importance level and the number of device associations. The number of device associations is determined by the number of associations of the device with other devices in the network topology diagram.

[0025] The trigger condition adopts a multi-parameter combination form, requiring multiple parameters to simultaneously meet their respective alarm rules to trigger the linkage alarm. For each parameter, set the corresponding abnormal fluctuation duration condition. After the main device triggers an alarm, within a specified time, when it is detected that the relevant parameters of the devices functionally associated with it simultaneously meet the corresponding alarm rules and the abnormal fluctuation duration reaches the set duration, the linkage alarm is triggered.

[0026] Preferably, the specific analysis method for the correlation relationship between alarm points of each device is as follows: Compare each warning feature value with the data thresholds of each level to obtain the alarm level corresponding to each warning feature value. When a certain warning feature value exceeds the first-level threshold, it is initially determined that there is an alarm situation. Mark the warning feature as an alarm point, trigger a separate alarm, and determine the device corresponding to the warning feature value.

[0027] Starting from the device, trace along the physical connections and data transmission paths between devices in the network topology diagram to check whether there are alarm points in the upstream devices connected to it. If there are alarm points in the upstream device and there is a direct physical connection and data association with the device in the network topology, record the alarm points of the upstream device as associated alarm points, thereby obtaining the association relationship between the alarm points of each device.

[0028] Preferably, the specific content of the linkage alarm rule is as follows: The linkage alarm rule is constructed based on multiple associations of devices, and the multiple associations of devices include functional association, spatial location association, and time series association.

[0029] Under the linkage alarm rule constructed based on multiple associations of devices, when an alarm is triggered by a certain device, record its alarm time. Within a subsequent specific time window, as long as other devices meet the alarm trigger conditions set by any one of the functional association, spatial location association, and time series association, a linkage alarm is triggered.

[0030] Preferably, the specific analysis method for generating the alarm content is as follows: Record the time when the alarm is detected, query the parameter category to which the source warning feature triggering the alarm belongs, and determine the device type of the device corresponding to the alarm according to the classification setting of the alarm points in the system.

[0031] Obtain the specific values of the warning features corresponding to the alarm when the alarm is triggered, the alarm level, the device corresponding to the alarm, and the device type of the device to which it belongs, and combine them in a set format to generate the alarm content.

[0032] Preferably, the specific operation method of the network construction module is as follows: Perform correlation analysis on the warning features corresponding to each alarm and the inherent features of the corresponding warning target to obtain the correlation relationship between the warning features corresponding to the alarm and the warning target features. Taking the warning target as a node and the number of alarms and the correlation relationship between the warning features corresponding to the alarm and the warning target features as edges, construct an alarm warning network.

[0033] Preferably, the specific analysis method of the alarm analysis module is as follows: Select a specified warning target, select each time period at a fixed time length, read the number of alarms of the specified warning target and the device association situation of the specified warning target from the alarm warning network, and comprehensively evaluate to obtain the risk level of the specified warning target in each time period.

[0034] Formulate a risk level standard, compare the risk levels of the specified warning target in each time period with the risk level standard, and determine the risk level to which the specified warning target belongs in each time period. The risk level standard includes low risk, medium risk, and high risk.

[0035] Feedback on the risk levels of the specified warning targets for each time period.

[0036] Compared with the prior art, the present invention has the following beneficial effects: First, by defining warning targets, collecting data for the warning targets, and extracting each warning feature of the warning target data, the present invention can make the data collection work more targeted, avoid collecting a large amount of irrelevant data, and improve the efficiency and quality of data collection.

[0037] Second, the present invention determines the alarm points based on the historical warning target data of each device, obtains the correlation relationship between the alarm points of each device, constructs a single alarm point rule, and sets up a linkage alarm rule based on multiple correlations, which helps to discover the potential connections and mutual influences between devices. By setting up the linkage alarm rule, these chain reactions can be captured in time, and countermeasures can be taken in advance to avoid the spread and expansion of faults.

[0038] Third, by generating alarm content, the present invention conducts a correlation analysis between the alarm features and the inherent features of the warning target, constructs an alarm warning network, and then evaluates the risk situation of the specified warning target, enabling a more comprehensive understanding of the essence and scope of influence of the alarm event, and thus more accurately judging the cause of the fault and the possible consequences. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0040] Figure 1 It is the system module connection diagram of the present invention.

[0041] Figure 2 is Figure 1 the flowchart of the alarm rule setting module in

[0042] Figure 3 is Figure 2 the flowchart of the correlation relationship between the alarm points of each device in DETAILED DESCRIPTION OF THE EMBODIMENTS

[0043] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0044] Please refer toFigure 1 As shown, a data upload fault monitoring system based on intelligent early warning includes a target data acquisition module, an early warning feature extraction module, an alarm rule setting module, a linkage alarm module, a network construction module, an alarm analysis module, and a management database.

[0045] The management database is connected to the early warning feature extraction module, the alarm rule setting module, the linkage alarm module, the network construction module, and the alarm analysis module; the linkage alarm module is connected to the alarm rule setting module and the network construction module; the early warning feature extraction module is connected to the label data acquisition module and the alarm rule setting module.

[0046] The target data acquisition module is used to define the warning target and collect data for the warning target.

[0047] The target data acquisition module is connected to various devices in the mine through a sensor network. The sensor network uses a communication protocol to transmit data and can automatically identify and screen data interfaces of different types of equipment to collect early warning target data of various types of equipment. It can adapt to various complex and diverse equipment interface types in the mine to ensure that the early warning target data of various types of equipment are collected without omission, provide sufficient and accurate data basis for subsequent early warning analysis, and improve the effectiveness and stability of the early warning system.

[0048] At the same time, the target data acquisition module supports multiple collection methods. For the collection of third-party data upload status, Party A coordinates the third party to provide relevant interfaces; it solves the problem of cross-system data acquisition, enriches data sources, and improves the system's adaptability to data collection in complex environments.

[0049] It should be noted that the warning targets are the operating data of various equipment in the mine and the environmental data collected by sensors for collecting various types of data. The operating data include ventilation equipment data, lifting equipment data, and mining equipment data. The environmental data include gas concentration, temperature, humidity, carbon monoxide concentration, wind speed, and dust concentration.

[0050] The warning feature extraction module is used to preprocess the acquired warning target data and perform dimensionality reduction operations to extract various warning features of the warning target data.

[0051] The specific operation method of the warning feature extraction module is as follows: the covariance matrix between the features is calculated based on the preprocessed warning target data, and each eigenvalue and the corresponding eigenvector are obtained by performing eigenvalue decomposition on the covariance matrix; the principal components are extracted as warning features by calculating the covariance matrix, performing eigenvalue decomposition and other steps, so as to mine the potential key features in the data, highlight the factors that have an important indicative role in the warning, provide strong support for the subsequent modules to identify abnormal situations more efficiently and accurately, and significantly improve the sensitivity and accuracy of the warning.

[0052] It should be noted that the preprocessing is to normalize the early warning target data and convert it to the same numerical range.

[0053] Arrange the eigenvalues in descending order, and select the eigenvectors corresponding to the eigenvalues with the cumulative contribution rate reaching the set value as the principal components; the eigenvalue reflects the importance of the corresponding eigenvector in the data. The larger the eigenvalue, the greater the variance of the data in the direction represented by the eigenvector, that is, the more information it contains.

[0054] Exemplarily, the reference set value of the contribution rate is 80%.

[0055] Project the early warning target data onto the principal components to obtain the early warning target data after dimensionality reduction, and extract the corresponding eigenvector as the early warning feature; through the dimensionality reduction operation, while retaining the key information of the data, the amount of data and the complexity of analysis are greatly reduced.

[0056] The alarm rule setting module is used to determine the alarm points according to the historical early warning target data of each device, obtain the correlation relationship between the alarm points of each device, and construct a single alarm point rule.

[0057] Please refer to Figure 2 As shown, the specific operation method of the alarm rule setting module is as follows: S1. Collect the device type, parameter category, and installation location of each device in the mine, obtain the historical early warning target data of each device in the mine and the corresponding timestamps, where the historical early warning target data includes normal data and fault data; the normal data reflects the stable operation state of the device, the fault data shows the situation when the device is abnormal, and the timestamp records the generation time of the data, which is convenient for analyzing the change law of the device state over time.

[0058] S2. Select the alarm points by analyzing the historical early warning target data, sort out the correlation relationship between the alarm points, and classify the alarm points according to the parameter type; there are various correlations between the related alarm points, such as functional correlation, spatial position correlation, and time series correlation. Sorting out the correlation relationship can construct a device fault propagation model to comprehensively understand the occurrence and development process of the fault.

[0059] Exemplarily, there is a large amount of gas concentration data. If the device fails or is abnormal later when the gas concentration reaches 1.0% multiple times, set 1.0% as an alarm point.

[0060] Exemplarily, after the gas sensor in a certain area alarms, if the ventilation equipment in the adjacent area alarms, combined with the spatial position correlation, it can be inferred that gas accumulation may be caused by ventilation problems.

[0061] It should be noted that the classification of alarm points according to parameter types includes gas concentration alarm points, temperature alarm points, humidity alarm points, wind speed alarm points, dust concentration alarm points, and carbon monoxide concentration alarm points.

[0062] S3. Set data thresholds for each level as alarm rules respectively according to the historical warning target data of each device in the mine. The data thresholds for each level include a first-level threshold, a second-level threshold, and a third-level threshold. Different levels of alarms can be made according to the degree of data deviation from the normal range, and corresponding measures can be taken according to the alarm level, improving the efficiency and effectiveness of fault response.

[0063] It should be noted that the first-level threshold is the earliest warning limit and is a sign that the device operation state begins to deviate from normal. It refers to the normal data fluctuation range in history and the minimum requirements for the safe operation of the device. The second-level threshold indicates that the device operation anomaly intensifies and the potential fault increases, and is set based on the parameter critical values in historical fault data and the impact on device performance. The third-level threshold is the most serious situation, where the device or environment is in a dangerous state, and is set based on safety regulations, major accident cases, and the ultimate bearing capacity of the device.

[0064] Exemplarily, for carbon monoxide concentration, when the carbon monoxide concentration reaches 0.0024%, it may be an early sign of slight leakage of the device or extremely small amount of incomplete combustion of coal underground. At this time, 0.0024% is set as the first-level threshold. When it reaches 0.024%, it has obvious harm to human health and means that there may be relatively serious device failures or abnormal coal combustion underground. 0.024% is set as the second-level threshold. When it reaches 0.1%, it seriously endangers the life safety of personnel, and 0.1% is set as the third-level threshold.

[0065] The linkage alarm module is used to draw a network topology diagram and set linkage alarm rules, priorities, and trigger conditions based on various correlations.

[0066] The specific analysis method of the linkage alarm module is as follows: Sort out the physical connections and data transmission relationships between each device in the mine, draw a network topology diagram of the mine devices in combination with the correlation relationships between alarm points, and divide the device importance levels according to the device association situations in the network topology diagram. The device importance levels include first-level devices, second-level devices, and third-level devices, so as to obtain the device importance levels of the corresponding devices for each alarm point. It can clearly show the overall architecture of the mine devices and the importance of each device, providing a basis for formulating reasonable linkage alarm rules.

[0067] The data transmission relationships between each device in the mine include wired networks and wireless networks.

[0068] Based on the association relationships between alarm points in the network topology diagram, multiple linkage alarm rules are set respectively based on various associations, and priorities and trigger conditions are set for the linkage alarm rules; it can avoid the limitations of single alarms, take into account the mutual influence between devices, more comprehensively monitor the operating status of mine equipment, and reduce the situations of missed alarms and false alarms.

[0069] The priority of the linkage alarm rule is based on the equipment importance level and the equipment association quantity. The equipment association quantity is determined by the number of associations of this equipment with other equipment in the network topology diagram; it increases the reliability of the alarm. Only when multiple relevant parameters simultaneously meet the alarm rule and the abnormal fluctuation lasts for a certain period of time, the linkage alarm is triggered, which can effectively exclude short-term and non-substantive abnormal situations and improve the accuracy of the alarm.

[0070] In a preferred embodiment of the present invention, the priority of the linkage alarm rule is as follows: extract the equipment importance level and the equipment association quantity, and then perform a summation calculation according to the weights to obtain the priority coefficient of the linkage alarm rule, and compare it with the priority of the corresponding linkage alarm rule of the preset priority coefficients of each linkage alarm rule to obtain the priority coefficient of the linkage alarm rule.

[0071] Exemplarily, the weights corresponding to the equipment importance level and the equipment association quantity are 0.5 and 0.5.

[0072] The trigger condition adopts a multi-parameter combination form, requiring that multiple parameters simultaneously meet their respective alarm rules to trigger the linkage alarm. For each parameter, a corresponding abnormal fluctuation duration condition is set. After the main equipment triggers an alarm, within a specified time, it is detected that the relevant parameters of the equipment functionally associated with it simultaneously meet the corresponding alarm rules, and the abnormal fluctuation duration reaches the set duration, then the linkage alarm is triggered.

[0073] Please refer to Figure 3 As shown, the specific analysis method for the association relationships between the alarm points of each equipment is: compare each warning characteristic value with each level data threshold to obtain the alarm level corresponding to each warning characteristic value. When a certain warning characteristic value exceeds the first-level threshold, it is initially determined that there is an alarm situation, and this warning characteristic is recorded as an alarm point to trigger a single alarm, and determine the equipment corresponding to the warning characteristic value.

[0074] Starting from the said device, trace along the physical connections and data transmission paths between devices in the network topology diagram to check whether there are alarm points in the upstream devices connected thereto. If there are alarm points in the upstream devices and there are direct physical connections and data associations with the said device in the network topology, then record the alarm points of the upstream devices as associated alarm points, thereby obtaining the association relationships between the alarm points of each device; it is possible to find out the potential connections between devices, comprehensively grasp the scope of possible faults and related devices, which helps to more accurately judge the cause and impact degree of the faults and take more effective countermeasures.

[0075] It should be noted that in a specific embodiment, there are multiple devices in a mine, and they form a specific network topology through physical connections and data transmissions. The specific devices include: main ventilator, underground local ventilator, gas sensor, data collector, and monitoring center server. The relationships between the devices are as follows: The main ventilator provides the main ventilation for the entire mine and is connected to each area underground through ventilation ducts; the underground local ventilator is responsible for the ventilation of local areas and is connected to the ventilation ducts of the main ventilator; gas sensors are installed at various positions underground to monitor the gas concentration in real time and transmit data to the data collector through cables; the data collector collects the data of each gas sensor and then transmits the data to the monitoring center server through the network.

[0076] The gas sensor at a certain position underground detects that the gas concentration exceeds the set threshold and triggers an alarm, and trace starts from this gas sensor.

[0077] Physical connection trace: This gas sensor is connected to the data collector through a cable. Check the data collector and find that the communication status indicator light of the data collector shows abnormality. After further inspection, its data transmission interface fails and cannot receive and transmit the data of the gas sensor normally. Moreover, there is a direct physical connection (cable connection) and data association (responsible for collecting the data of this gas sensor) between this data collector and this gas sensor. Therefore, record the communication fault alarm point of the data collector as the associated alarm point of the alarm point of this gas sensor.

[0078] Functional association trace: Continue the analysis. The abnormal gas concentration may be related to the ventilation situation. Trace along the ventilation duct to the underground local ventilator that provides ventilation for this area and find that the motor current of this local ventilator increases abnormally, resulting in insufficient ventilation volume and making the gas unable to be discharged in time. There is a direct functional association between the underground local ventilator and this gas sensor (ventilation affects the gas concentration). Therefore, record the abnormal motor current alarm point of the underground local ventilator as the associated alarm point of the alarm point of this gas sensor. Further trace to the main ventilator. If the main ventilator also has a fault that affects the overall ventilation effect and causes abnormal pressure of the local ventilator, then the fault alarm point of the main ventilator can also be used as an associated alarm point.

[0079] The specific content of the linked alarm rule is as follows: The linked alarm rule is constructed based on multiple correlations of devices. The multiple correlations of devices include functional correlation, spatial location correlation, and time series correlation. Considering the relationships between devices from multiple perspectives makes the alarm system more flexible and comprehensive, and can adapt to different types of fault situations.

[0080] It should be noted that for the functional correlation of devices, when the key operating parameters of a certain device are lower than or exceed the set threshold, and at the same time, specific environmental parameters or its own operating parameters within the operating area of other devices that are closely related to it functionally also show abnormalities, the linked alarm is immediately triggered.

[0081] It should be noted that in a specific embodiment, the shearer, scraper conveyor, and ventilator are functionally interrelated. The shearer is responsible for coal mining, the scraper conveyor is responsible for transporting the mined coal out, and the ventilator provides fresh air for the coal mining face and discharges harmful gases.

[0082] Rule setting: When the motor current of the shearer exceeds 120% of the rated value and lasts for 3 minutes, at the same time, the conveying speed of the scraper conveyor is lower than 80% of the normal speed and lasts for 2 minutes, and the air volume of the ventilator is lower than 90% of the normal air volume and lasts for 1 minute, the linked alarm is triggered.

[0083] Trigger situation: One day, the motor current of the shearer suddenly increased to 130% of the rated value and lasted for 3 minutes. At the same time, the conveying speed of the scraper conveyor dropped to 70% of the normal speed and lasted for 2 minutes, and the air volume of the ventilator also dropped to 85% of the normal air volume and lasted for 1 minute.

[0084] At this time, the triggering conditions of the functional correlation linked alarm rule are met, and the system immediately issues a linked alarm to remind the staff to check the operating status of the shearer, scraper conveyor, and ventilator, as there may be equipment failures or overload situations.

[0085] For the spatial location correlation, if the operating parameters of the devices in a certain area show abnormalities, and the relevant parameters of the devices in the adjacent area or the parameters of the upstream and downstream devices also show abnormal changes accordingly, the linked alarm will be triggered.

[0086] It should be noted that in a specific embodiment, in the same transportation roadway, multiple devices are installed, such as belt conveyors, transfer machines, lighting devices, etc. These devices are adjacent in space and may be affected by each other's failures.

[0087] Rule setting: When the belt deviation sensor of the belt conveyor detects belt deviation and the duration reaches 2 minutes, at the same time, the motor temperature of the adjacent transfer machine exceeds 110% of the normal temperature and the duration reaches 1 minute, and more than 50% of the lighting equipment in this roadway goes out and the duration reaches 30 seconds, a linkage alarm is triggered.

[0088] Trigger situation: At a certain moment, the belt of the belt conveyor deviates, the belt deviation sensor alarms and lasts for 2 minutes. At the same time, the motor temperature of the adjacent transfer machine rises to 115% of the normal temperature and lasts for 1 minute, and 60% of the lighting equipment in this roadway goes out and lasts for 30 seconds.

[0089] At this time, the trigger condition of the linkage alarm rule for spatial position correlation is met, and the system issues a linkage alarm, prompting the staff that there may be mechanical failures, electrical failures or other abnormal conditions in this roadway, and it is necessary to conduct a timely investigation.

[0090] Regarding time series correlation, when the operating parameters of other associated devices show abnormal fluctuations within a specific time period after a specific device is started, or when a device failure lasts for a certain period of time and abnormal operating signs appear in other devices in the relevant area, a linkage alarm is triggered.

[0091] It should be noted that in a specific embodiment, there is a correlation between the gas monitoring system and the ventilation system in terms of time series. When the gas concentration increases, it is usually necessary for the ventilation system to adjust the air volume in a timely manner to reduce the gas concentration.

[0092] Rule setting: When the gas sensor detects that the gas concentration exceeds 0.8% and the duration reaches 1 minute, within the next 2 minutes, if the air volume of the ventilator does not increase correspondingly to the specified ventilation volume (set according to the corresponding relationship between gas concentration and ventilation volume), a linkage alarm is triggered.

[0093] Trigger situation: The gas sensor detects that the gas concentration suddenly rises to 0.9% and lasts for 1 minute. The system automatically monitors the operation of the ventilator and finds that the air volume of the ventilator does not increase as required within the next 2 minutes and remains at a low level.

[0094] At this time, the trigger condition of the linkage alarm rule for time series correlation is met, and the system issues a linkage alarm, reminding the staff to adjust the operating parameters of the ventilator in a timely manner to prevent gas accumulation from causing safety accidents.

[0095] Under the above linkage alarm rules constructed based on various correlations of devices, when an alarm is triggered by a certain device, record its alarm time. Within a subsequent specific time window, as long as other devices meet the alarm trigger conditions set by any one of the functional correlation, spatial location correlation, and time series correlation, a linkage alarm is triggered; within the specific time window, triggering the linkage alarm according to whether other devices meet the alarm trigger conditions set by the correlation can timely detect potential problems of other devices associated with the faulty device, make preparations for prevention and response in advance, and avoid the further expansion of the fault.

[0096] The specific analysis method for generating the alarm content is as follows: Record the time when the alarm is detected, query the parameter category to which the source warning feature triggering the alarm belongs, and judge the device type of the device corresponding to the alarm according to the classification setting of the alarm point by the system.

[0097] Obtain the specific value of the warning feature corresponding to the alarm when the alarm is triggered, the alarm level, the device corresponding to the alarm, and the device type of the device to which it belongs, and combine them in the set format to generate the alarm content; make the alarm information more standardized and clear, facilitate reading and understanding, improve the efficiency and accuracy of information transmission, and is conducive to taking corresponding treatment measures in a timely manner.

[0098] It should be noted that the set format is the rule for combining the obtained information and consists of the following parts:

[0099] Start identifier: Used to identify the start of the alarm content. For example, it can be a specific character or string, such as "

Alarm start

[0100] Description of warning feature value: Clearly describe the name and specific value of the warning feature in words. For example, "Temperature warning feature value: 35°C" or "Pressure warning feature value: 20MPa".

[0101] Description of alarm level: Clearly write out the name of the alarm level, such as "Alarm level: Level 2".

[0102] Information of device corresponding to alarm: Include the identification information of the device, such as "Device corresponding to alarm: Device number 001" or "Device corresponding to alarm: Device name is XX model machine".

[0103] Description of device type of device to which it belongs: Write out the type of the device, such as "Device type of device to which it belongs: Electronic device".

[0104] End identifier: Used to identify the end of the alarm content, such as "

Alarm end

[0105] Exemplarily, a complete alarm content:

Alarm start

Alarm end

[0106] The network construction module is used to generate alarm content, perform correlation analysis on alarm characteristics and inherent characteristics of warning targets, and construct an alarm warning network.

[0107] The specific operation method of the network construction module is as follows: Perform correlation analysis on the warning characteristics of each alarm corresponding to the alarm and the inherent characteristics of the corresponding warning target to obtain the correlation relationship between the warning characteristics of the alarm corresponding to the alarm and the warning target characteristics. Taking the warning target as a node and the number of alarms and the correlation relationship between the warning characteristics of the alarm corresponding to the alarm and the warning target characteristics as edges, construct an alarm warning network; it can intuitively display the relationships between different warning targets and their connections with alarm characteristics, providing a clear network structure for subsequent alarm analysis and facilitating the exploration of potential failure modes and rules.

[0108] It should be noted that for the correlation relationship between the warning characteristics of the alarm corresponding to the alarm and the warning target characteristics, please refer to Table 1 specifically. In Table 1, some representative data are listed.

[0109] Table 1. Correlation relationship between warning characteristics of the alarm corresponding to the alarm and warning target characteristics

[0110]

[0111]

[0112] The alarm analysis module is used to evaluate the risk situation of a specified warning target according to the alarm warning network.

[0113] The specific analysis method of the alarm analysis module is as follows: Select a specified warning target and select each time period at a fixed time length. Read the number of alarms of the specified warning target and the device association situation of the specified warning target from the alarm warning network, and comprehensively evaluate to obtain the risk degree of the specified warning target in each time period; it can dynamically monitor the risk status of the specified warning target in different time periods, taking into account the time factor and the associated influence between devices, making the risk assessment more comprehensive and accurate.

[0114] In a preferred embodiment of the present invention, the risk degree of the specified warning target in each time period is as follows: Extract the number of alarms of the specified warning target and the number of device associations of the specified warning target in each time period, and then perform a weighted summation calculation to obtain the risk degree of the specified warning target in each time period.

[0115] Exemplarily, the weights corresponding to the number of alarms of the specified warning target and the number of device associations of the specified warning target in each time period are 0.5 and 0.5.

[0116] Formulate a risk level standard, compare the risk levels of the specified warning targets in each time period with the risk level standard, and determine the risk levels to which the specified warning targets in each time period belong. The risk level standard includes low risk, medium risk, and high risk; it is convenient to take corresponding management and maintenance measures according to different risk levels, improving the safety and reliability of the operation of mine equipment.

[0117] It should be noted that low risk indicates that the specified warning target operates relatively stably in the current time period, and the possibility of failure or causing serious problems is relatively low. In terms of device associations, the devices strongly associated with it are in good operating condition, and the association relationship has not shown abnormal fluctuations. At the same time, even if there are alarms, their severity levels are mostly minor, having little impact on the overall system operation; medium risk means that some signs that need attention have emerged for the specified warning target, and there are certain risk hazards. In terms of device associations, the association relationship with some key devices may become unstable, such as data transmission delays and slightly decreased collaborative work efficiency; high risk means that the specified warning target faces a greater risk of failure or has had an obvious impact on the system. In terms of device associations, it may have triggered a series of chain reactions, causing abnormalities in multiple devices associated with it, and even affecting the entire production process or system function.

[0118] Provide feedback on the risk levels to which the specified warning targets in each time period belong.

[0119] It should be noted that the form of the feedback can display the changes in the risk levels of the specified warning targets in each time period in the form of a chart, with different risk levels identified by different colors, such as green for low risk, yellow for medium risk, and red for high risk. At the same time, generate a detailed risk report, which includes the risk levels in each time period, analysis of the main risk factors (such as alarm types, problems with associated devices, etc.), and corresponding recommended measures.

[0120] A management database is used to store warning target data, historical warning target data, alarm points and threshold information, network topology map information, and alarm warning network data.

[0121] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention, and still be covered by the protection scope of the present invention.

Claims

1. A data upload fault monitoring system based on intelligent early warning, characterized in that: The system specifically includes the following modules: The target data acquisition module is used to define the warning target and collect data for the warning target; The warning feature extraction module is used to pre-process the acquired warning target data and perform dimensionality reduction operations to extract various warning features of the warning target data; The alarm rule setting module is used to determine the alarm point based on the historical warning target data of each device, obtain the correlation between the alarm points of each device, and construct a single alarm point rule; The linkage alarm module is used to draw the network topology map and set linkage alarm rules, priorities and trigger conditions based on various associations; The network construction module is used to generate alarm content, associate and analyze the alarm features with the inherent features of the warning target, and build an alarm warning network; Alarm analysis module, used to evaluate the risk of designated warning targets based on the alarm warning network; Management database, used to store warning target data, historical warning target data, alarm point and threshold information, network topology information, and alarm warning network data.

2. According to the data upload fault monitoring system based on intelligent early warning according to claim 1, it is characterized in that: The target data acquisition module is connected to various devices in the mine through a sensor network. The sensor network uses a communication protocol to transmit data and can automatically identify and filter data interfaces of different types of devices, thereby collecting early warning target data of various types of equipment; At the same time, the target data acquisition module supports multiple collection methods. For the collection of third-party data upload status, Party A coordinates the third party to provide relevant interfaces.

3. According to the data upload fault monitoring system based on intelligent early warning according to claim 1, it is characterized in that: The specific operation method of the early warning feature extraction module is: The covariance matrix between the features is calculated based on the preprocessed warning target data, and each eigenvalue and the corresponding eigenvector are obtained by performing eigenvalue decomposition on the covariance matrix; Arrange the eigenvalues ​​in descending order, and select the eigenvectors corresponding to the eigenvalues ​​whose cumulative contribution rate reaches the set value as the principal components; The warning target data is projected onto the principal component to obtain the warning target data after dimensionality reduction, and the corresponding features are extracted and recorded as warning features.

4. According to the data upload fault monitoring system based on intelligent early warning according to claim 1, it is characterized in that: The specific operation method of the alarm rule setting module is as follows: S1. Collect the equipment type, technical parameters, and installation location of each device in the mine, and obtain the historical warning target data of each device in the mine and the corresponding timestamp, where the historical warning target data includes normal data and fault data; S2. Select alarm points by analyzing historical warning target data and classify alarm points according to the key parameter types; S3. According to the historical warning target data of each device in the mine, each level of data threshold is set as an alarm rule, and the data threshold of each level includes a first-level threshold, a second-level threshold, and a third-level threshold.

5. The data upload fault monitoring system based on intelligent early warning according to claim 1 is characterized in that: The specific analysis method of the linkage alarm module is: Sorting out the physical connection and data transmission relationship between the various devices in the mine, thereby drawing a network topology diagram of the mine equipment, and dividing the equipment importance levels according to the association of the devices in the network topology diagram, and the equipment importance levels include primary equipment, secondary equipment, and tertiary equipment; Based on the network topology, obtain the association relationship between equipment alarm points, and set linkage alarm rules, priorities and trigger conditions based on equipment function association, spatial location association and time series association; The priority of the linkage alarm rule is determined based on the importance level of the equipment and the number of associated equipment. The trigger condition adopts a multi-parameter combination, and a corresponding duration condition is set for the abnormal fluctuation of each parameter. When all relevant parameters meet the alarm rules at the same time and the duration of the abnormal fluctuation reaches the set duration, the linkage alarm is triggered.

6. The data upload fault monitoring system based on intelligent early warning according to claim 5 is characterized in that: The specific analysis method of the correlation relationship between the alarm points of each device is as follows: Compare each warning feature value with each level of data threshold. When a warning feature value exceeds the alarm threshold range, it is preliminarily determined that an alarm situation exists, and the warning feature is recorded as an alarm point, and the device corresponding to the warning feature value is determined; Starting from the device, trace back along the physical connection and data transmission path between devices in the network topology diagram to find out whether the early warning characteristic value of the upstream device connected to it exceeds the alarm threshold range. If the early warning characteristic value of the upstream device exceeds the alarm threshold range and there is a direct physical connection and data association with the device in the network topology, the upstream device is recorded as an associated alarm point, thereby obtaining the association relationship between the alarm points of each device.

7. The data upload fault monitoring system based on intelligent early warning according to claim 5 is characterized in that: The specific content of the linkage alarm rule is: In terms of equipment functional relevance, when a key operating parameter of a device is lower than or exceeds a set threshold, and at the same time, specific environmental parameters in the operating area of ​​other functionally closely related devices or their own operating parameters also show abnormalities, a linkage alarm is immediately triggered; In terms of spatial location correlation, if the operating parameters of equipment in a certain area are abnormal, and the related parameters of equipment in adjacent areas or the parameters of upstream and downstream equipment also change abnormally, a linkage alarm will be triggered; Regarding time series correlation, when the operating parameters of other associated devices fluctuate abnormally within a specific time period after a specific device is started, or when other devices in the related area show signs of abnormal operation after a device failure lasts for a certain period of time, the linkage alarm is triggered.

8. The data upload fault monitoring system based on intelligent early warning according to claim 1 is characterized in that: The specific analysis method for generating alarm content is: Record the time when the alarm triggering conditions are met, query the category of the source data that triggers the alarm, and determine the type of mine equipment corresponding to the alarm based on the system's classification settings for early warning targets; The specific value and alarm level of the parameter when the alarm is triggered are obtained, and the obtained alarm time, warning target, and equipment type are combined according to the set format to generate the alarm content.

9. The data upload fault monitoring system based on intelligent early warning according to claim 8 is characterized in that: The specific operation method of the network construction module is as follows: The alarm characteristics of each alarm are correlated with the inherent characteristics of the corresponding warning target to obtain the correlation between the alarm characteristics and the warning target characteristics. An alarm warning network is constructed with the warning target as the node and the correlation between the number of alarms, alarm characteristics and warning target characteristics as the edge.

10. The data upload fault monitoring system based on intelligent early warning according to claim 1 is characterized in that: The specific analysis method of the alarm analysis module is: Select the designated warning target, and select each time period according to a fixed duration, read the alarm count of the designated warning target in each time period and the equipment association of the designated warning target from the alarm warning network, and comprehensively evaluate the risk level of the designated warning target in each time period; Formulate risk level standards, compare the risk levels of the designated warning targets in each period with the risk level standards, and determine the risk levels of the designated warning targets in each period. The risk level standards include low risk, medium risk, and high risk; Provide feedback on the risk level of the designated warning targets in each time period.

Images