Intelligent operation and maintenance method and system for dual-system evaporative condenser based on fault prediction

Abstract

The present invention belongs to the technical field of intelligent control of condensers. The present invention discloses a dual-system evaporative condenser intelligent operation and maintenance method and system based on fault prediction. The method extracts and integrates detection data with timestamps from multiple label detection devices, performs vertical and horizontal analysis, and obtains single device and process equipment information; extracts abnormal dynamic charts from process equipment information, generates an equipment abnormality matrix and inputs it into a main prediction unit for fault prediction; for results exceeding the co-occurrence threshold, a secondary verification is performed by a backup prediction unit, and if the results are consistent, the fault is confirmed; the dual verification mechanism and dynamic abnormality detection capability significantly improve the accuracy of fault prediction and operation and maintenance efficiency, and is suitable for scenarios with high reliability requirements; finally, based on the fault probability prediction trend, operation and maintenance tasks are generated and intervention is carried out in advance to realize intelligent operation and maintenance.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent control of condensers, and more specifically, to an intelligent operation and maintenance method and system for a dual-system evaporative condenser based on fault prediction. Background Art

[0002] Evaporative condensers are widely used in refrigeration, air conditioning, and industrial cooling systems. Their operating status directly affects the system's energy efficiency and stability. Traditional O&M methods typically rely on scheduled maintenance or post-fault repairs. Scheduled maintenance can be excessive or insufficient, failing to accurately match the equipment's actual status. Post-fault repairs can lead to downtime and production losses. Both methods lack early warning of potential failures.

[0003] For example, Chinese patent publication number CN118860787A discloses an intelligent operation and maintenance fault prediction method, device, equipment, medium, and product. This method preprocesses operation and maintenance data and extracts key target fault features, resolving issues such as the lack of fault prediction and intelligent early warning capabilities. However, the following issues still exist:

[0004] 1. The current fault prediction model has no verification function for missing or delayed data. When there is a problem with the prediction result, it is impossible to determine whether it is the analysis result or the model failure. When the equipment fails, the fault location cannot be obtained in time.

[0005] 2. Directly obtain the operating status of the equipment. The collected data is packaged in the memory as a whole. A large amount of data increases the difficulty and time of analysis. In addition, the equipment itself is not analyzed. When a fault occurs, it is necessary to check the equipment fault one by one, which is inefficient.

[0006] In view of this, the present application proposes a dual-system evaporative condenser intelligent operation and maintenance method and system based on fault prediction. Summary of the Invention

[0007] In order to overcome the above-mentioned defects of the prior art, the present invention proposes a dual-system evaporative condenser intelligent operation and maintenance method and system based on fault prediction, which is used for fault prediction and operation and maintenance of evaporative condensers to ensure the stable operation of the server cooling system.

[0008] In a first aspect, the present invention provides a dual-system evaporative condenser intelligent operation and maintenance method based on fault prediction, which is applied to a server and includes the following steps:

[0009] Extract detection data from multiple independently running tag detection devices, integrate the detection data according to location information to generate detection package data with timestamp;

[0010] Perform vertical analysis on target device nodes based on encapsulated time nodes to obtain single device information; perform horizontal analysis on single device information of all target device nodes based on work operation logic to obtain process device information; extract abnormal dynamic charts from process device information;

[0011] Extract the device anomaly matrix from the anomaly dynamic chart and send it to the fault prediction model corresponding to the primary prediction unit to predict the device failure probability and predicted fault location. Data with a device failure probability exceeding the co-occurrence threshold is then re-verified by the fault prediction model corresponding to the backup prediction unit. If the results of the primary and backup prediction units are consistent, it is confirmed that the device with the location tag information has failed.

[0012] Based on the probability of equipment failure, the equipment failure trend is predicted to obtain pending operation and maintenance tasks, and operation and maintenance intervention is carried out in advance for possible equipment failures based on the pending operation and maintenance tasks.

[0013] As a preferred technical solution of the first aspect of the present invention, the marking logic of the label detection device is:

[0014] Classifying the fault types of the evaporative condenser based on the fault database, and determining a target device node based on the fault type, wherein the target device node includes the location tag information and the detection data;

[0015] Building a local schematic map based on location tag information; the local schematic map includes at least one location tag information, each location tag information corresponds to at least one sensor installation point; each sensor installation point is equipped with at least one sensor, and each sensor collects detection data based on a sampling frequency and a reporting interval;

[0016] The detection data corresponding to the target device node are integrated to obtain the detection package data with a timestamp, the detection package data is stored in the storage module corresponding to the fault database, and the location tag information is used as a retrieval guide to associate the detection package data based on the location tag information; and the detection package data is updated in real time.

[0017] As a preferred technical solution of the first aspect of the present invention, the integration logic of the detection package data is:

[0018] The sampling frequency and reporting interval of all tag detection devices are counted, and the encapsulation time period is determined based on prior knowledge. The last timestamp of the encapsulation time period is marked as the encapsulation time node.

[0019] Acquire corresponding detection data based on the sampling frequency within the encapsulation time period, integrate the detection data of the same type to obtain detection encapsulation data, wherein the detection encapsulation data includes standard detection data and abnormal detection data; during the integration process, determine the detection safety threshold and detection change threshold corresponding to the detection data of adjacent timestamps based on prior knowledge;

[0020] If the test data is within the detection safety threshold and the floating range of adjacent test data is within the detection change threshold, the average value of all test data corresponding to the timestamp is used as the standard test data within the encapsulation time period;

[0021] Otherwise, the detection data within the encapsulation time period is subdivided in turn, and the adjacent detection data whose detection data is within the detection safety threshold and whose floating range is within the detection change threshold are integrated to update the local detection data, and all local monitoring data are integrated in matrix form and marked as abnormal detection data.

[0022] As a preferred technical solution of the first aspect of the present invention, the logic for obtaining the single device information is:

[0023] Using the same location tag information as a retrieval guide, extracting into a plane coordinate system based on the retrieval guide, using the detection package data as the vertical coordinate, the timestamp as the horizontal coordinate, and different package time nodes as the unit time length, using different marking symbols for different types of detection package data;

[0024] Extracting an abnormal feature variable from the abnormality detection data, wherein the abnormal feature variable corresponds to a change curvature per unit time length; and calculating an abnormal frequency of the change curvature per unit time length;

[0025] The abnormal probability is obtained by assigning a probability value based on the abnormal frequency, and the abnormal probability corresponding to all the location tag information of the target device node is counted with a timestamp. The standard detection data is represented by a specific numerical value and represented in a matrix form as a single device information with a timestamp.

[0026] As a preferred technical solution of the first aspect of the present invention, the logic for obtaining the process equipment information is:

[0027] Based on the evaporative condenser, the progress is segmented according to the workflow to form a transition stage. The single device information of all target device nodes corresponding to the transition stage is counted. The process progress time is determined based on the timestamp of the single device information. The process plane coordinate system is established with the workflow as the retrieval guide, with the standard test data as the vertical coordinate and the timestamp as the horizontal coordinate;

[0028] Analyze the transition phase based on the historical database to obtain an estimated progress time interval and an estimated progress information interval; represent the estimated progress time interval and the estimated progress information interval in a process plane coordinate system;

[0029] If the standard test data of the real-time monitoring evaporative condenser is consistent with the estimated progress information interval, it means that there is no abnormality in the current test data. If the time progress process is less than or equal to the expected progress process, it means that there is no abnormality in the current workflow. The current process progress time is updated to the estimated progress time interval and stored in the historical storage database in real time.

[0030] If the time progress process is greater than the expected progress process, it means that the current process progress is abnormal, and the staff's feedback information is obtained in real time; and the process plane coordinate system is marked, and the marked process plane coordinate system is marked as the process equipment information.

[0031] As a preferred technical solution of the first aspect of the present invention, feedback information from staff is obtained in real time, and the feedback information includes remediable defects and irremediable defects;

[0032] If the feedback information indicates that the defect can be eliminated, the staff will be prompted to eliminate the defect and record the elimination timestamp; the standard test data and the cause of the abnormal fluctuation will be stored in the fault database for updating the normal fluctuation detection interval;

[0033] If the feedback information is that the defect cannot be eliminated, the staff will be prompted to check carefully, shut down the equipment in time, and wait for maintenance.

[0034] As a preferred technical solution of the first aspect of the present invention, the acquisition logic of the abnormal dynamic chart is:

[0035] The process plane coordinate system corresponding to the process equipment information is used as the X-axis and Y-axis, the process progress time is used as the X-axis, the vertical coordinates of each inspection package data in the single equipment information are used as the Y-axis, and the inspection type corresponding to the inspection package data in the single equipment information is marked as the Z-axis;

[0036] The execution relationship between the process progress time and the corresponding workflow is used to construct an abnormal dynamic chart corresponding to the process equipment information, and the abnormal dynamic chart is expressed in the form of an AND / OR tree based on the three-dimensional table;

[0037] The AND / OR tree includes a node set, an edge set, and attributes of the edges. The root node in the node set indicates the end of the workflow, and each non-root node represents a refined workflow in the workflow.

[0038] As a preferred technical solution of the first aspect of the present invention, the fault prediction module includes a feature extraction module, a main prediction unit and a backup prediction unit, wherein: the main prediction unit 320 and the backup prediction unit use the same machine learning algorithm for training the fault prediction model, and the feature extraction module extracts the equipment anomaly matrix required by the machine learning algorithm, and outputs the equipment failure probability based on the equipment anomaly matrix.

[0039] As a preferred technical solution of the first aspect of the present invention, the application logic of the fault prediction model is:

[0040] The device anomaly matrix is used as the input factor and the device failure probability is used as the output factor. The device anomaly matrix is a set of abnormal feature variables. The device anomaly matrix is constructed according to the abnormal probability in the single device information, and the rough set theory is used to optimize the device anomaly matrix. The optimized device anomaly matrix is used to calculate the device failure probability through asymmetric importance factors and separation information, and the fault distribution is intuitively displayed, which is convenient for operation and maintenance personnel to quickly locate the fault.

[0041] In a second aspect, the present invention provides a dual-system intelligent operation and maintenance system for evaporative condensers based on fault prediction, which is based on the implementation of the first aspect and is applied to a server. The server includes a data acquisition module, a data analysis module, a fault prediction module, and a maintenance control module. The modules are connected to each other via wired or wireless connections.

[0042] The data acquisition module extracts detection data from multiple independently running tag detection devices, integrates the detection data according to the location information to generate detection package data with a time stamp, and sends the detection package data to the data analysis module;

[0043] The data analysis module performs vertical analysis on target device nodes based on the encapsulated time nodes to obtain single device information; performs horizontal analysis on the single device information of all target device nodes based on the work operation logic to obtain process device information; extracts abnormal dynamic charts from the process device information; and sends the abnormal dynamic charts to the fault prediction module;

[0044] The fault prediction module extracts the device anomaly matrix from the anomaly dynamic chart and sends the device anomaly matrix to the fault prediction model corresponding to the primary prediction unit to predict the device failure probability and predicted failure location. Data with a device failure probability exceeding the co-occurrence threshold is then secondary verified using the fault prediction model corresponding to the backup prediction unit. If the results of the primary and backup prediction units are consistent, the device with the location tag information is confirmed to have failed. The confirmed failed device is then sent to the maintenance control module.

[0045] The maintenance control module predicts equipment failure trends based on the probability of equipment failure to obtain pending operation and maintenance tasks, and performs operation and maintenance interventions on possible equipment failures in advance based on the pending operation and maintenance tasks.

[0046] The technical effects and advantages of the dual-system evaporative condenser intelligent operation and maintenance method and system based on fault prediction of the present invention are as follows:

[0047] The present invention realizes the intelligent operation and maintenance of evaporative condensers in servers through data integration, anomaly detection, fault prediction and proactive operation and maintenance intervention; the dual verification mechanism and dynamic anomaly detection capability significantly improve the accuracy of fault prediction and operation and maintenance efficiency, and is suitable for scenarios with high reliability requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 Schematic diagram of the dual-system evaporative condenser intelligent operation and maintenance system of the present invention;

[0049] Figure 2 This is a partial system framework diagram of the fault prediction module of the present invention;

[0050] Figure 3 This is a flow chart of the intelligent operation and maintenance method for a dual-system evaporative condenser according to the present invention. DETAILED DESCRIPTION

[0051] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0052] Example 1

[0053] See also Figure 1 As shown, the dual-system evaporative condenser intelligent operation and maintenance system based on fault prediction described in this embodiment is applied to a server, which includes a data acquisition module 100, a data analysis module 200, a fault prediction module 300 and a maintenance control module 400; each module is connected via wired or wireless connections;

[0054] The data acquisition module 100 extracts detection data from multiple independently operated tag detection devices, integrates the detection data according to the location information to generate detection package data with a timestamp, and sends the detection package data to the data analysis module 200;

[0055] It should be noted that: this embodiment is for upgrading and using existing equipment, and does not require modification of existing equipment. It only requires the deployment of detection equipment at various key locations of the evaporative condenser. The detection equipment is mainly based on sensors, which have the characteristics of high precision, low power consumption and corrosion resistance to adapt to the operating environment of the evaporative condenser. The sensors that transmit data in real time ensure that the operation and maintenance system can obtain the operating status of the equipment in a timely manner. However, during the real-time transmission process, it cannot be guaranteed that every sensor can transmit data in real time. Once the detection equipment fails or there is an abnormal delay, if there is an error in the detection data itself, even if the analysis result is correct, its analysis logic is based on erroneous data and cannot be popularized. In addition, a large amount of overlapping data will only increase the system data analysis and calculation strength and increase the system burden. Therefore, in this embodiment, we use detection package data to configure the detection data at the detection equipment site according to the sampling frequency, reporting interval and working mode parameters of the sensor; encapsulate the configured data, mark it with a timestamp, and then transmit it through the wireless communication module.

[0056] The purpose of the currently installed sensors is to sound an alarm when an abnormality occurs, prompting maintenance personnel to carry out inspections. This means that the current equipment itself has detection equipment. In addition, when the equipment sounds an alarm, on the one hand, it may be the result of a fault in the detection equipment, a false alarm, and on the other hand, it may be that the equipment cannot work normally, seriously affecting the production process.

[0057] Exemplarily, the tag detection device includes a temperature sensor, a pressure sensor, a flow meter and a vibration sensor for deploying location tag information; the detection data includes temperature values, pressure values, flow values and vibration frequency values.

[0058] Specifically, the label detection device's labeling logic is:

[0059] Classifying the fault types of the evaporative condenser based on the fault database, and determining a target device node based on the fault type, wherein the target device node includes the location tag information and the detection data;

[0060] Building a local schematic map based on location tag information; the local schematic map includes at least one location tag information, each location tag information corresponds to at least one sensor installation point; each sensor installation point is equipped with at least one sensor, and each sensor collects detection data based on a sampling frequency and a reporting interval;

[0061] The detection data corresponding to the target device node are integrated to obtain the detection package data with a timestamp, the detection package data is stored in the storage module corresponding to the fault database, and the location tag information is used as a retrieval guide to associate the detection package data based on the location tag information; and the detection package data is updated in real time.

[0062] Specifically, the integration logic of the detection package data is:

[0063] The sampling frequency and reporting interval of all tag detection devices are counted, and the encapsulation time period is determined based on prior knowledge. The last timestamp of the encapsulation time period is marked as the encapsulation time node.

[0064] Acquire corresponding detection data based on the sampling frequency within the encapsulation time period, integrate the detection data of the same type to obtain detection encapsulation data, wherein the detection encapsulation data includes standard detection data and abnormal detection data; during the integration process, determine the detection safety threshold and detection change threshold corresponding to the detection data of adjacent timestamps based on prior knowledge;

[0065] If the test data is within the detection safety threshold and the floating range of adjacent test data is within the detection change threshold, the average value of all test data corresponding to the timestamp is used as the standard test data within the encapsulation time period;

[0066] Otherwise, the detection data within the encapsulation time period is subdivided in turn, and the adjacent detection data whose detection data is within the detection safety threshold and whose floating range is within the detection change threshold are integrated to update the local detection data, and all local monitoring data are integrated in matrix form and marked as abnormal detection data.

[0067] The data analysis module 200 performs a vertical analysis of the target device node based on the encapsulated time node to obtain single device information; performs a horizontal analysis of the single device information of all target device nodes based on the work operation logic to obtain process device information; extracts abnormal dynamic charts from the process device information; and sends the abnormal dynamic charts to the fault prediction module 300;

[0068] Specifically, the logic for obtaining the single device information is as follows:

[0069] Using the same location tag information as a retrieval guide, extracting into a plane coordinate system based on the retrieval guide, using the detection package data as the vertical coordinate, the timestamp as the horizontal coordinate, and different package time nodes as the unit time length, using different marking symbols for different types of detection package data;

[0070] Extracting an abnormal feature variable from the abnormality detection data, wherein the abnormal feature variable corresponds to a change curvature per unit time length; and calculating an abnormal frequency of the change curvature per unit time length;

[0071] The abnormal probability is obtained by assigning a probability value based on the abnormal frequency, and the abnormal probability corresponding to all the location tag information of the target device node is counted with a timestamp. The standard detection data is represented by a specific numerical value and represented in a matrix form as a single device information with a timestamp.

[0072] It should be noted that a simple comparison is used here to quickly determine whether a device is in an abnormal state, reducing false positives and missed negatives. Status information is then represented in matrix form with a timestamp, providing operators with detailed device operation history and status change trends. This allows operators to focus on addressing confirmed abnormal devices and avoid unnecessary inspection and maintenance of normal equipment. Furthermore, discrete data is sent to the fault prediction module 300 in matrix form, reducing packet loss during transmission.

[0073] To further illustrate, the logic for obtaining the process equipment information is as follows:

[0074] Based on the evaporative condenser, the progress is segmented according to the workflow to form a transition stage. The single device information of all target device nodes corresponding to the transition stage is counted. The process progress time is determined based on the timestamp of the single device information. The process plane coordinate system is established with the workflow as the retrieval guide, with the standard test data as the vertical coordinate and the timestamp as the horizontal coordinate;

[0075] Exemplarily, the transition phase includes a refrigerant condensation phase, a spray water circulation phase, a heat exchange phase, an air flow phase, a heat dissipation phase, and a water circulation phase;

[0076] Analyze the transition phase based on the historical database to obtain an estimated progress time interval and an estimated progress information interval; represent the estimated progress time interval and the estimated progress information interval in a process plane coordinate system;

[0077] If the standard test data of the real-time monitoring evaporative condenser is consistent with the estimated progress information interval, it means that there is no abnormality in the current test data. If the time progress process is less than or equal to the expected progress process, it means that there is no abnormality in the current workflow. The current process progress time is updated to the estimated progress time interval and stored in the historical storage database in real time.

[0078] If the time progress process is greater than the expected progress process, it means that the current process progress is abnormal, and the staff's feedback information is obtained in real time; and the process plane coordinate system is marked, and the marked process plane coordinate system is marked as the process equipment information.

[0079] Further, feedback information from staff is obtained in real time, and the feedback information includes remediable defects and irremediable defects;

[0080] If the feedback information indicates that the defect can be eliminated, the staff will be prompted to eliminate the defect and record the elimination timestamp; the standard test data and the cause of the abnormal fluctuation will be stored in the fault database for updating the detection safety threshold and detection change threshold;

[0081] If the feedback information is that the defect cannot be eliminated, the staff will be prompted to check carefully, shut down the equipment in time, and wait for maintenance.

[0082] Further explanation: the acquisition logic of the abnormal dynamic chart is:

[0083] The process plane coordinate system corresponding to the process equipment information is used as the X-axis and Y-axis, the process progress time is used as the X-axis, the vertical coordinates of each inspection package data in the single equipment information are used as the Y-axis, and the inspection type corresponding to the inspection package data in the single equipment information is marked as the Z-axis;

[0084] The execution relationship between the process progress time and the corresponding workflow is used to construct an abnormal dynamic chart corresponding to the process equipment information, and the abnormal dynamic chart is expressed in the form of an AND / OR tree based on the three-dimensional table;

[0085] The AND / OR tree includes a node set, an edge set, and attributes of the edges. The root node in the node set indicates the end of the workflow, and each non-root node represents a refined workflow in the workflow.

[0086] The fault prediction module 300 extracts the equipment anomaly matrix from the abnormal dynamic chart, sends the equipment anomaly matrix to the fault prediction model corresponding to the main prediction unit to predict the equipment failure probability and the predicted failure location, and performs a secondary verification on the data whose equipment failure probability exceeds the co-occurrence threshold through the fault prediction model corresponding to the backup prediction unit; if the results of the main prediction unit and the backup prediction unit are consistent, it is confirmed that the equipment where the location tag information is located has failed; and the confirmed faulty equipment is sent to the maintenance control module 400.

[0087] Specifically, if Figure 2 As shown, the fault prediction module 300 includes a feature extraction module 310, a main prediction unit 320 and a backup prediction unit 330, wherein: the main prediction unit 320 and the backup prediction unit 330 use the same machine learning algorithm for training the fault prediction model, and the feature extraction module 310 extracts the equipment anomaly matrix required by the machine learning algorithm, and outputs the equipment failure probability based on the equipment anomaly matrix.

[0088] Application logic of the fault prediction model:

[0089] The device anomaly matrix is used as the input factor and the device failure probability is used as the output factor. The device anomaly matrix is a set of abnormal feature variables. The device anomaly matrix is constructed according to the abnormal probability in the single device information, and the rough set theory is used to optimize the device anomaly matrix. The optimized device anomaly matrix is used to calculate the device failure probability through asymmetric importance factors and separation information, and the fault distribution is intuitively displayed, which is convenient for operation and maintenance personnel to quickly locate the fault.

[0090] The maintenance control module 400 predicts the equipment failure trend based on the equipment failure probability to obtain pending operation and maintenance tasks, and performs operation and maintenance intervention on possible equipment failures in advance according to the pending operation and maintenance tasks.

[0091] Example 2

[0092] This embodiment provides an exemplary description based on the first embodiment. The server uses an evaporative condenser for heat dissipation and adopts the above method for intelligent operation and maintenance, wherein:

[0093] Multiple independent temperature sensors, humidity sensors, and pressure sensors are installed on the evaporative condenser, and the temperature, humidity, and pressure data collected by the sensors with time stamps and location tags are sent to the server corresponding to the evaporative condenser.

[0094] The server analyzes the historical data of each sensor and extracts the standard detection data of each sensor. For example, it analyzes the temperature change trend of a temperature sensor over the past week.

[0095] Based on the operating logic of the evaporative condenser, data from all sensors is analyzed horizontally to extract information about the entire condenser's operating status. For example, the relationship between temperature, humidity, and pressure is analyzed to determine whether the condenser is within its normal operating range.

[0096] Based on process equipment information, anomaly dynamic charts are extracted. For example, if a temperature sensor's temperature rises sharply within a short period of time, exceeding the normal range, the anomaly signature is sent to the primary prediction unit, which uses the fault prediction model to predict the probability and location of the equipment failure. For example, the primary prediction unit predicts an 80% probability of a temperature sensor failure. If the primary prediction unit's failure probability exceeds a co-occurrence threshold (e.g., 80%), the data is sent to the backup prediction unit for secondary verification. If the backup prediction unit also predicts an 80% probability of failure for the same sensor, the sensor is confirmed to be faulty.

[0097] Based on the probability of equipment failure, equipment failure trends are predicted. For example, if a temperature sensor is predicted to fail completely within the next 24 hours, operations personnel can be scheduled to replace the temperature sensor in advance based on the prediction, preventing server downtime due to overheating. This allows potential evaporative condenser failures to be detected and addressed in advance, ensuring stable server operation. This not only improves equipment reliability but also reduces downtime and repair costs caused by failures.

[0098] Example 3

[0099] See also Figure 3As shown, for parts not described in detail in this embodiment, please refer to the description of Example 1. A dual-system evaporative condenser intelligent operation and maintenance method based on fault prediction is provided, which is applied to a server and includes the following steps:

[0100] Extract detection data from multiple independently running tag detection devices, integrate the detection data according to location information to generate detection package data with timestamp;

[0101] Perform vertical analysis on target device nodes based on encapsulated time nodes to obtain single device information; perform horizontal analysis on single device information of all target device nodes based on work operation logic to obtain process device information; extract abnormal dynamic charts from process device information;

[0102] Extract the device anomaly matrix from the anomaly dynamic chart and send it to the fault prediction model corresponding to the primary prediction unit to predict the device failure probability and predicted fault location. Data with a device failure probability exceeding the co-occurrence threshold is then re-verified by the fault prediction model corresponding to the backup prediction unit. If the results of the primary and backup prediction units are consistent, it is confirmed that the device with the location tag information has failed.

[0103] Based on the probability of equipment failure, the equipment failure trend is predicted to obtain pending operation and maintenance tasks, and operation and maintenance intervention is carried out in advance for possible equipment failures based on the pending operation and maintenance tasks.

[0104] As a preferred technical solution of the first aspect of the present invention, the marking logic of the label detection device is:

[0105] Classifying the fault types of the evaporative condenser based on the fault database, and determining a target device node based on the fault type, wherein the target device node includes the location tag information and the detection data;

[0106] Building a local schematic map based on location tag information; the local schematic map includes at least one location tag information, each location tag information corresponds to at least one sensor installation point; each sensor installation point is equipped with at least one sensor, and each sensor collects detection data based on a sampling frequency and a reporting interval;

[0107] The detection data corresponding to the target device node are integrated to obtain the detection package data with a timestamp, the detection package data is stored in the storage module corresponding to the fault database, and the location tag information is used as a retrieval guide to associate the detection package data based on the location tag information; and the detection package data is updated in real time.

[0108] The integration logic of the detection package data is:

[0109] The sampling frequency and reporting interval of all tag detection devices are counted, and the encapsulation time period is determined based on prior knowledge. The last timestamp of the encapsulation time period is marked as the encapsulation time node.

[0110] Acquire corresponding detection data based on the sampling frequency within the encapsulation time period, integrate the detection data of the same type to obtain detection encapsulation data, wherein the detection encapsulation data includes standard detection data and abnormal detection data; during the integration process, determine the detection safety threshold and detection change threshold corresponding to the detection data of adjacent timestamps based on prior knowledge;

[0111] If the test data is within the detection safety threshold and the floating range of adjacent test data is within the detection change threshold, the average value of all test data corresponding to the timestamp is used as the standard test data within the encapsulation time period;

[0112] Otherwise, the detection data within the encapsulation time period is subdivided in turn, and the adjacent detection data whose detection data is within the detection safety threshold and whose floating range is within the detection change threshold are integrated to update the local detection data, and all local monitoring data are integrated in matrix form and marked as abnormal detection data.

[0113] The logic for obtaining the single device information is as follows:

[0114] Using the same location tag information as a retrieval guide, extracting into a plane coordinate system based on the retrieval guide, using the detection package data as the vertical coordinate, the timestamp as the horizontal coordinate, and different package time nodes as the unit time length, using different marking symbols for different types of detection package data;

[0115] Extracting an abnormal feature variable from the abnormality detection data, wherein the abnormal feature variable corresponds to a change curvature per unit time length; and calculating an abnormal frequency of the change curvature per unit time length;

[0116] The abnormal probability is obtained by assigning a probability value based on the abnormal frequency, and the abnormal probability corresponding to all the location tag information of the target device node is counted with a timestamp. The standard detection data is represented by a specific numerical value and represented in a matrix form as a single device information with a timestamp.

[0117] The logic for obtaining the process equipment information is as follows:

[0118] Based on the evaporative condenser, the progress is segmented according to the workflow to form a transition stage. The single device information of all target device nodes corresponding to the transition stage is counted. The process progress time is determined based on the timestamp of the single device information. The process plane coordinate system is established with the workflow as the retrieval guide, with the standard test data as the vertical coordinate and the timestamp as the horizontal coordinate;

[0119] Analyze the transition phase based on the historical database to obtain an estimated progress time interval and an estimated progress information interval; represent the estimated progress time interval and the estimated progress information interval in a process plane coordinate system;

[0120] If the standard test data of the real-time monitoring evaporative condenser is consistent with the estimated progress information interval, it means that there is no abnormality in the current test data. If the time progress process is less than or equal to the expected progress process, it means that there is no abnormality in the current workflow. The current process progress time is updated to the estimated progress time interval and stored in the historical storage database in real time.

[0121] If the time progress process is greater than the expected progress process, it means that the current process progress is abnormal, and the staff's feedback information is obtained in real time; and the process plane coordinate system is marked, and the marked process plane coordinate system is marked as the process equipment information.

[0122] Obtaining feedback from staff in real time, including remediable and irremediable defects;

[0123] If the feedback information indicates that the defect can be eliminated, the staff will be prompted to eliminate the defect and record the elimination timestamp; the standard test data and the cause of the abnormal fluctuation will be stored in the fault database for updating the normal fluctuation detection interval;

[0124] If the feedback information is that the defect cannot be eliminated, the staff will be prompted to check carefully, shut down the equipment in time, and wait for maintenance.

[0125] The acquisition logic of the abnormal dynamic chart is:

[0126] The process plane coordinate system corresponding to the process equipment information is used as the X-axis and Y-axis, the process progress time is used as the X-axis, the vertical coordinates of each inspection package data in the single equipment information are used as the Y-axis, and the inspection type corresponding to the inspection package data in the single equipment information is marked as the Z-axis;

[0127] The execution relationship between the process progress time and the corresponding workflow is used to construct an abnormal dynamic chart corresponding to the process equipment information, and the abnormal dynamic chart is expressed in the form of an AND / OR tree based on the three-dimensional table;

[0128] The AND / OR tree includes a node set, an edge set, and attributes of the edges. The root node in the node set indicates the end of the workflow, and each non-root node represents a refined workflow in the workflow.

[0129] The fault prediction module 300 includes a feature extraction module 310, a main prediction unit 320 and a backup prediction unit 330, wherein: the main prediction unit 320 and the backup prediction unit 330 use the same machine learning algorithm to train the fault prediction model, and the feature extraction module 310 extracts the equipment anomaly matrix required by the machine learning algorithm, and outputs the equipment failure probability based on the equipment anomaly matrix.

[0130] Application logic of the fault prediction model:

[0131] The device anomaly matrix is used as the input factor and the device failure probability is used as the output factor. The device anomaly matrix is a set of abnormal feature variables. The device anomaly matrix is constructed according to the abnormal probability in the single device information, and the rough set theory is used to optimize the device anomaly matrix. The optimized device anomaly matrix is used to calculate the device failure probability through asymmetric importance factors and separation information, and the fault distribution is intuitively displayed, which is convenient for operation and maintenance personnel to quickly locate the fault.

[0132] Example 4

[0133] According to an exemplary embodiment, an electronic device includes: a processor and a memory, wherein the memory stores a computer program that can be called by the processor;

[0134] The processor executes the above-mentioned dual-system evaporative condenser intelligent operation and maintenance method based on fault prediction by calling the computer program stored in the memory.

[0135] The electronic devices provided in the embodiments of the present application can vary significantly due to configuration or performance differences. These devices can include one or more processors (Central Processing Units, CPUs) and one or more memories, wherein the memories store at least one computer program, which is loaded and executed by the processors to implement the fault prediction-based intelligent operation and maintenance methods for dual-system evaporative condensers provided in the various method embodiments described above. The electronic devices can also include other components for implementing device functions. For example, the electronic devices can also include components such as wired or wireless network interfaces and input / output interfaces for input and output. The embodiments of the present application are not described in detail here.

[0136] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0137] It should be understood that determining B based on A does not mean determining B based solely on A. B can also be determined based on A and / or other information.

[0138] The above embodiments can be implemented in whole or in part via software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product comprises one or more computer instructions or computer programs. When loaded or executed on a computer, the processes or functions described in accordance with the embodiments of the present invention are fully or partially performed. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another. For example, the computer instructions can be transferred from one website, computer, server, or data center to another website, computer, server, or data center via a wired or wireless network. The computer-readable storage medium can be any available medium accessible by a computer, or a data storage device such as a server or data center that contains a collection of one or more available media. The available medium can be magnetic media (e.g., floppy disks, hard disks, or magnetic tapes), optical media (e.g., DVDs), or semiconductor media. The semiconductor media can be a solid-state drive.

[0139] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the present invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0140] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0141] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only one type. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0142] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0143] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0144] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

[0145] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A dual-system evaporative condenser intelligent operation and maintenance method based on fault prediction, applied to servers, characterized by: The following steps are involved: Extract detection data from multiple independently running tag detection devices, integrate the detection data according to location information to generate detection package data with timestamp; Perform longitudinal analysis on target device nodes based on encapsulation time nodes to obtain single device information; Based on the work operation logic, the single device information of all target device nodes is horizontally analyzed to obtain process device information; Extract abnormal dynamic charts from process equipment information; Extract the device anomaly matrix from the anomaly dynamic chart and send it to the fault prediction model corresponding to the primary prediction unit to predict the device failure probability and predicted fault location. Data with a device failure probability exceeding the co-occurrence threshold is then re-verified by the fault prediction model corresponding to the backup prediction unit. If the results of the primary and backup prediction units are consistent, it is confirmed that the device with the location tag information has failed. Predict equipment failure trends based on equipment failure probability to obtain pending maintenance tasks, and conduct maintenance interventions in advance for potential equipment failures based on pending maintenance tasks; The label detection device's marking logic is as follows: Classifying the fault types of the evaporative condenser based on the fault database, and determining a target device node based on the fault type, wherein the target device node includes the location tag information and the detection data; Building a local schematic map based on location tag information; the local schematic map includes at least one location tag information, each location tag information corresponds to at least one sensor installation point; each sensor installation point is equipped with at least one sensor, and each sensor collects detection data based on a sampling frequency and a reporting interval; Integrate the detection data corresponding to the target device node to obtain detection package data with a timestamp, store the detection package data in a storage module corresponding to the fault database, and use the location tag information as a retrieval guide to associate the detection package data based on the location tag information; And update the detection and packaging data in real time; The integration logic of the detection package data is: The sampling frequency and reporting interval of all tag detection devices are counted, and the encapsulation time period is determined based on prior knowledge. The last timestamp of the encapsulation time period is marked as the encapsulation time node. Acquire corresponding detection data based on the sampling frequency within the packaging time period, and integrate the detection data of the same type to obtain detection packaging data, wherein the detection packaging data includes standard detection data and abnormal detection data; During the integration process, the detection safety threshold and detection change threshold corresponding to the detection data of adjacent timestamps are determined based on prior knowledge; If the test data is within the detection safety threshold and the floating range of adjacent test data is within the detection change threshold, the average value of all test data corresponding to the timestamp is used as the standard test data within the encapsulation time period; Otherwise, the detection data within the encapsulation time period is subdivided in turn, and the adjacent detection data whose detection data is within the detection safety threshold and whose floating range is within the detection change threshold are integrated to update the local detection data, and all local monitoring data are integrated in matrix form and marked as abnormal detection data.

2. The dual-system evaporative condenser intelligent operation and maintenance method based on fault prediction according to claim 1 is characterized in that: The logic for obtaining the single device information is as follows: Using the same location tag information as a retrieval guide, extracting into a plane coordinate system based on the retrieval guide, using the detection package data as the vertical coordinate, the timestamp as the horizontal coordinate, and different package time nodes as the unit time length, using different marking symbols for different types of detection package data; Extracting an abnormal feature variable from the abnormality detection data, wherein the abnormal feature variable corresponds to a curvature of change per unit time length; The abnormal frequency of the curvature of statistical change per unit time length; The abnormal probability is obtained by assigning a probability value based on the abnormal frequency, and the abnormal probability corresponding to all the location tag information of the target device node is counted with a timestamp. The standard detection data is represented by a specific numerical value and represented in a matrix form as a single device information with a timestamp.

3. The dual-system evaporative condenser intelligent operation and maintenance method based on fault prediction according to claim 2 is characterized in that: The logic for obtaining the process equipment information is as follows: Based on the evaporative condenser, the progress is segmented according to the workflow to form a transition stage. The single device information of all target device nodes corresponding to the transition stage is counted. The process progress time is determined based on the timestamp of the single device information. The process plane coordinate system is established with the workflow as the retrieval guide, with the standard test data as the vertical coordinate and the timestamp as the horizontal coordinate; Analyze the transition phase based on the historical database to obtain the estimated progress time interval and estimated progress information interval; Represent the estimated progress time interval and the estimated progress information interval in the process plane coordinate system; If the standard test data of the real-time monitoring evaporative condenser is consistent with the estimated progress information interval, it means that there is no abnormality in the current test data. If the time progress process is less than or equal to the expected progress process, it means that there is no abnormality in the current workflow. The current process progress time is updated to the estimated progress time interval and stored in the historical storage database in real time. If the time progress process is greater than the expected progress process, it means that the current process progress is abnormal, and the staff's feedback information is obtained in real time; and the process plane coordinate system is marked, and the marked process plane coordinate system is marked as the process equipment information.

4. The dual-system evaporative condenser intelligent operation and maintenance method based on fault prediction according to claim 3 is characterized in that: Obtaining feedback from staff in real time, including remediable and irremediable defects; If the feedback information indicates that the defect can be eliminated, the staff will be prompted to eliminate the defect and record the defect elimination timestamp; Store standard test data and abnormal fluctuation causes in the fault database for updating the normal fluctuation detection interval; If the feedback information is that the defect cannot be eliminated, the staff will be prompted to check carefully, shut down the equipment in time, and wait for maintenance.

5. The dual-system evaporative condenser intelligent operation and maintenance method based on fault prediction according to claim 4 is characterized in that: The acquisition logic of the abnormal dynamic chart is: The process plane coordinate system corresponding to the process equipment information is used as the X-axis and Y-axis, the process progress time is used as the X-axis, the vertical coordinates of each inspection package data in the single equipment information are used as the Y-axis, and the inspection type corresponding to the inspection package data in the single equipment information is marked as the Z-axis; The execution relationship between the process progress time and the corresponding workflow is used to construct an abnormal dynamic chart corresponding to the process equipment information, and the abnormal dynamic chart is expressed in the form of an AND / OR tree based on the three-dimensional table; The AND / OR tree includes a node set, an edge set, and attributes of the edges. The root node in the node set indicates the end of the workflow, and each non-root node represents a refined workflow in the workflow.

6. The dual-system evaporative condenser intelligent operation and maintenance method based on fault prediction according to claim 5 is characterized in that: The fault prediction module (300) includes a feature extraction module (310), a main prediction unit (320) and a backup prediction unit (330), wherein the main prediction unit (320) and the backup prediction unit (330) use the same machine learning algorithm to train the fault prediction model, extract the device anomaly matrix required by the machine learning algorithm through the feature extraction module (310), and output the device failure probability based on the device anomaly matrix.

7. The dual-system evaporative condenser intelligent operation and maintenance method based on fault prediction according to claim 6 is characterized in that: Application logic of the fault prediction model: The device anomaly matrix is used as the input factor and the device failure probability is used as the output factor. The device anomaly matrix is a set of abnormal feature variables. The device anomaly matrix is constructed according to the abnormal probability in the single device information, and the rough set theory is used to optimize the device anomaly matrix. The optimized device anomaly matrix is used to calculate the device failure probability through asymmetric importance factors and separation information, and the fault distribution is intuitively displayed, which is convenient for operation and maintenance personnel to quickly locate the fault.

8. A dual-system evaporative condenser intelligent operation and maintenance system based on fault prediction, based on the implementation of the dual-system evaporative condenser intelligent operation and maintenance method based on fault prediction according to any one of claims 1 to 7, characterized in that: Applied to a server, the server comprises a data acquisition module (100), a data analysis module (200), a fault prediction module (300) and a maintenance control module (400); the modules are connected to each other via wired or wireless connections; A data acquisition module (100) extracts detection data from a plurality of independently operated tag detection devices, integrates the detection data according to the location information to generate detection package data with a timestamp, and sends the detection package data to a data analysis module (200); A data analysis module (200) performs longitudinal analysis on target device nodes based on encapsulation time nodes to obtain single device information; Performing horizontal analysis on the single device information of all target device nodes based on the work operation logic to obtain process device information; extracting abnormal dynamic graphs from the process device information; and sending the abnormal dynamic graphs to the fault prediction module (300); The fault prediction module (300) extracts the device anomaly matrix from the abnormal dynamic graph, sends the device anomaly matrix to the fault prediction model corresponding to the main prediction unit to predict the device failure probability and the predicted failure location, and performs secondary verification on the data of the device failure probability exceeding the co-occurrence threshold through the fault prediction model corresponding to the backup prediction unit; if the results of the main prediction unit and the backup prediction unit are consistent, it is confirmed that the device with the location tag information has failed; and the confirmed failed device is sent to the maintenance control module (400); The maintenance control module (400) predicts the equipment failure trend based on the equipment failure probability to obtain pending operation and maintenance tasks, and performs operation and maintenance intervention on possible equipment failures in advance according to the pending operation and maintenance tasks.

Images