Intelligent switch cabinet anti-explosion abnormity monitoring method and system

Abstract

The invention relates to the technical field of switch cabinet state monitoring, in particular to an intelligent switch cabinet anti-explosion abnormity monitoring method and system, and the system comprises a data collection module, a data storage and management module, a correlation analysis module, a time sequence prediction module, a risk assessment and early warning module, a communication interface module and a man-machine interaction module. According to the invention, by setting a multi-source sensing and intelligent analysis integrated monitoring architecture, early prediction and accurate diagnosis of the fire blast risk of the switch cabinet are realized, the contribution weight of each inducement parameter to a fault result is quantified through a dynamic correlation analysis algorithm, and a time sequence prediction model is introduced to establish an advanced early warning mechanism of parameter change, so that the early warning of the fire blast risk of the switch cabinet is realized. According to the method, when direct fault indexes such as temperature and gas concentration do not exceed the standard, a potential fault development path can be recognized in advance by analyzing the change trend of incentive parameters such as current and voltage, so that an accident handling mode is converted from post-remedy to pre-prevention, and the active safety guarantee capability of operation of the switch cabinet is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of switch cabinet state monitoring, in particular to an intelligent switch cabinet anti-explosion abnormality monitoring method and system. BACKGROUND

[0002] As a key device for realizing power distribution, control and protection in the power system, the operation reliability of the switch cabinet is directly related to the safety and stability of the power supply network. Due to long-term operation in a harsh environment of strong current and high voltage, hidden dangers such as loose connection, insulation aging and increased contact resistance are prone to occur inside the switch cabinet. These hidden dangers may develop into local overheating, insulation deterioration and even internal arc fault, eventually causing serious explosion accidents. The traditional switch cabinet monitoring method mainly relies on periodic manual inspection and simple temperature measurement. This method is not only inefficient, but also difficult to find latent faults in time, and has been unable to meet the needs of real-time sensing and early warning of device state in modern smart grids.

[0003] With the development of sensing technology, existing technologies have begun to use online monitoring devices to monitor single or a few parameters such as local temperature and partial discharge of the switch cabinet. Although this method has made progress from manual inspection to automatic monitoring, it still lacks in-depth mining of the correlation between multiple parameters. The main defects are as follows: on the one hand, the monitoring system usually sets a fixed alarm threshold, and only triggers an alarm when the measured value exceeds the threshold. This kind of lagging alarm cannot provide enough buffer time for fault handling. On the other hand, different types of faults often show different characteristic parameter changes in the early stage, and isolated parameter monitoring cannot accurately identify complex fault patterns, which may lead to false alarms or missed alarms, affecting the accuracy of early warning.

[0004] Therefore, we propose an intelligent switch cabinet anti-explosion abnormality monitoring method and system. By building a monitoring system that integrates multi-source sensing and intelligent analysis, establishing a correlation analysis model centered on result data and cause data, and introducing a dynamic optimization mechanism with self-learning ability, we can provide an intelligent solution for switch cabinet anti-explosion by providing early warning in the early stage of fault and continuously optimizing the diagnosis accuracy through continuous learning. SUMMARY

[0005] To achieve the above purpose, the present application proposes an intelligent switch cabinet anti-explosion abnormality monitoring method, which includes the following steps: Step 1: Multi-source sensing and data collection, comprehensive state detection is performed by deploying multi-source sensors at key nodes inside the switch cabinet; Second step: data correlation modeling and archiving step, the correlation between the result data and the cause data is established, the parameters directly representing the fault results are defined as the result data, the potential cause parameters are defined as the cause data, and a cause data mapping table with the result data as the core is constructed; Third step: dynamic correlation analysis and weight distribution, the curve similarity of each cause data sequence and the result data sequence is calculated to quantify the contribution weight of each cause parameter to the fault result; Fourth step: time series modeling and prediction weight calculation step, the time series analysis method is used to determine the best time lag interval of each cause data relative to the result data, and a prediction model is established; Fifth step: real-time risk prediction and early warning step, based on the weight model and the time effectiveness model, the comprehensive risk index is calculated to realize graded early warning; Sixth step: misjudgment feedback and model self-correction, when a judgment error sample occurs, collect negative sample cases, dynamically adjust the weight distribution and early warning threshold, and realize the self-learning optimization of the model.

[0006] In one example, the third step calculates the importance of different cause data under the same result data. In the same result data, different cause data corresponding to different state result data are selected, and a circuit diagram is generated for curvature calculation. The dynamic correlation strength between the cause and the result is quantified by analyzing the morphological similarity of the data curve.

[0007] In one example, the fourth step uses a time series analysis algorithm to automatically find the best time lag interval between different cause data and result data, and assigns prediction weights to these lagged cause data.

[0008] In one example, the sixth step monitors the occurrence of significant fluctuations in single or multiple cause data and issues a warning accordingly. If no expected change in the corresponding result data is observed within the set time extension interval, the system will determine that this warning is a misjudgment.

[0009] In one example, the sixth step, if the number of misjudgments exceeds the set number threshold, it is determined to be a negative sample case. In the model self-correction, the weight of the specific cause data that leads to this misjudgment is adjusted, the weight value of the second weight is increased, and the determination is made again. If the second weight still misjudges, compare the data in the negative sample case with the data when the fault occurs in the database, find the same misjudgment data, and find the cause data with different values under the same misjudgment data. The cause data and the previous misjudgment data are combined for monitoring, and the update of the determination is completed.

[0010] In one example, an intelligent switch cabinet explosion-proof abnormality monitoring system comprises a data acquisition module, a data storage and management module, an association analysis module, a time series prediction module, a risk assessment and early warning module, a communication interface module and a human-computer interaction module.

[0011] In one example, the data acquisition module is composed of multiple types of sensor units, a signal conditioning unit and an edge computing node, and the data storage and management module constructs a hybrid storage architecture of a time series database and a relational database.

[0012] In one example, the association analysis module realizes dynamic association weight calculation function, adopts dynamic time warping algorithm to calculate the curve similarity of cause data sequence and result data sequence, outputs 0-1 interval association weight value, and is provided with a weight updating unit.

[0013] In one example, the time series prediction module integrates autoregressive integrated moving average model and long short-term memory neural network algorithm, a lag analysis engine automatically performs cross-correlation function calculation to determine the best time lag interval of each cause data relative to the result data, generate a lag parameter table, and a weight time extension unit performs time translation on historical cause data according to the lag interval to construct a prediction feature vector.

[0014] In one example, the risk assessment and early warning module includes a risk index calculator and a hierarchical early warning generator, a self-learning optimization module is composed of a misjudgment identification unit and a negative sample processing unit, a communication interface module realizes connection with a substation integrated automation system, a human-computer interaction module is based on Web architecture, provides real-time monitoring pictures, trend curve display, early warning information list, model parameter configuration and historical data query function, and supports mobile terminal access.

[0015] The intelligent switch cabinet explosion-proof abnormality monitoring method and system can bring the following beneficial effects: 1. The monitoring architecture integrating multi-source perception and intelligent analysis is set in the present application, early prediction and accurate diagnosis of switch cabinet explosion risk are realized, the contribution weight of each cause parameter to the fault result is quantified through a dynamic association analysis algorithm, a time series prediction model is introduced to establish a lead warning mechanism of parameter change, potential fault development paths can be identified in advance by analyzing the change trend of current, voltage and other cause parameters when temperature, gas concentration and other direct fault indicators have not exceeded the standard, thereby changing the accident handling mode from post-repair to pre-prevention, and the active safety protection capability of switch cabinet operation is significantly improved.

[0016] 2.The application automatically collects negative sample cases through the misjudgment feedback mechanism, uses the difference analysis engine to compare the data characteristics under the normal and misjudgment states, dynamically adjusts the weight distribution and early warning threshold of each inducement parameter, and the self-correcting mechanism enables the system to adapt to changes in different operating environments and device states, effectively reduces false positives and false negatives, and optimizes the diagnosis rules through the continuously accumulated fault case library, so that the judgment accuracy and reliability of the system are continuously improved as the running time is prolonged, providing intelligent protection for the continuous evolution of the long-term stable operation of the switch cabinet. BRIEF DESCRIPTION OF DRAWINGS

[0017] The drawings described herein are intended to provide further understanding of the present application, form a part of the present application, and the illustrative embodiments of the present application and the description thereof are used to explain the present application, and do not constitute an improper limitation on the present application. In the drawings: Fig. 1 Method flow diagram of the intelligent switch cabinet anti-explosion abnormality monitoring method and system; Fig. 2 System architecture diagram of the intelligent switch cabinet anti-explosion abnormality monitoring method and system. DETAILED DESCRIPTION

[0018] In order to more clearly explain the overall concept of the present application, the following will be described in detail with reference to the drawings.

[0019] In the description of the present application, it should be understood that the terms "center", "upper", "lower", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "axial", "radial", "circumferential" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present application.

[0020] In addition, the terms "first", "second" are only for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the technical features indicated. Therefore, the features defined with "first", "second" can explicitly or implicitly include one or more of the features. In the description of the present application, the meaning of "a plurality of" is two or more, unless otherwise specifically limited.

[0021] In the present application, unless specifically defined and limited otherwise, the terms "mounting", "connection", "connecting", "fixing" and the like should be interpreted broadly, for example, can be fixed connection, can also be detachable connection, or integral; can be mechanical connection, can also be electrical connection, and can also be communication; can be directly connected, can also be indirectly connected through an intermediate medium, can be internal communication of two elements or interaction relationship between two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0022] In the present application, unless specifically defined and limited otherwise, the first feature is "on" or "under" the second feature. The first and second features can be in direct contact, or the first and second features can be in indirect contact through an intermediate medium. In the description of the present application, the description of the terms "one scheme", "some schemes", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in connection with the scheme or example are included in at least one scheme or example of the present application. In the present application, the illustrative description of the above terms does not necessarily refer to the same scheme or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more schemes or examples in a suitable manner.

[0023] As shown in Figs. 1-2 The present application proposes a kind of intelligent switchgear anti-explosion abnormal monitoring method, comprising the following steps: First step: multi-source perception and data collection, utilize through the multi-source sensor of deployment in switchgear internal key node and carry out comprehensive state detection.Data acquisition core is to cover the main fault type leading to explosion, including through the wireless passive temperature sensor installed in circuit breaker contact and other high-voltage parts real-time monitoring overheating risk, through ultra-high frequency and ultrasonic sensor capture local discharge electromagnetic and acoustic signal to assess insulation deterioration, through electrochemical gas sensor monitoring hydrogen, carbon monoxide and other insulation material decomposition characteristic gas concentration, and through high-speed arc light sensor as the last line of defense of internal arc fault.All these data are collected through the edge computing gateway in cabinet, which is responsible for preliminary data cleaning, time synchronization and caching, and then uploaded to central analysis system through industrial network, to provide uniform, high-quality data basis for subsequent analysis.

[0024] Second step: data correlation modeling and archiving, establish the relationship between results and cause data, define parameters such as temperature, characteristic gas concentration and partial discharge intensity as result data, define potential cause parameters such as current, voltage and humidity as cause data, and group the result data as the main body, collect data in cause data, such as setting up a temperature database, collecting current, voltage and humidity corresponding to different temperatures.

[0025] Third step: dynamic correlation analysis and weight distribution, calculate the importance of different cause data under the same result data, in the same result data, select different cause data corresponding to different state result data, and generate a circuit diagram for curvature calculation, such as generating a temperature change curve with time as the horizontal axis and temperature as the vertical axis, generating a voltage change curve with time as the horizontal axis and voltage as the vertical axis, calculating the curvature of the temperature change curve and the curvature of the voltage change curve, comparing the curvature of the temperature change curve, and the closer it is, the higher the influence on temperature change, so as to determine the weight of each cause data in the same result data. The core of this step is to quantify the dynamic correlation strength between cause and result by analyzing the morphological similarity of data curve (not just the numerical value), which can more sensitively capture the key causes that evolve synchronously with the fault.

[0026] Fourth step: time series modeling and prediction weight calculation, for the cause data corresponding to the same result data, since the result has a delay, such as the increase of voltage in the previous one or two minutes, which leads to the increase of temperature in the subsequent one or two minutes, the weight of the cause data is extended in time, such as the cause data will lead to the increase or decrease of the subsequent result data.

[0027] The system will introduce time series analysis algorithm to automatically find the best time lag interval between different cause data and result data, and give prediction weight to these lagged cause data, so that the system can predict the trend of result data in the future based on the change of current cause.

[0028] Fifth step: real-time risk prediction and early warning The detected result data or cause data is transmitted to the weight model established in the third step and the time effectiveness model established in the fourth step for real-time risk assessment. Instead of passively waiting for the result data to exceed the standard, the system continuously calculates a "risk index" that integrates the changes of all high-weight, time-effective cause data. When the index indicates that the system state is developing towards a known failure mode, even if all result data is still within the normal range, the system will trigger early warning.

[0029] Sixth step: false alarm feedback and model self-correction When the system detects that a single or multiple cause data has a significant fluctuation and issues a warning accordingly, if no expected change in the corresponding result data is observed within the set time extension interval, the system will determine that this warning is a false alarm or an over-sensitive warning.

[0030] When the above situation occurs several times, the specific number of times can be set to a threshold number of times, such as twice. After two false alarms, the system automatically records the complete data sequence of this event to form a negative sample case.

[0031] Based on the negative sample case, the system will start the self-correction mechanism. First, it will lower the weight of the specific inducement data that caused this misjudgment, and the weight of the result data, so that it will not easily trigger a high-level alarm in similar future scenarios. And in turn, it will continue to compare the second weight when the result data appears, increase the weight value of the second weight, and re-determine. If the second weight still misjudges, it needs to consider each inducement data, record each data in the negative sample case, and compare it with the data in the previous database when the fault occurs. Find out the same misjudgment data and the inducement data with different values under the same misjudgment data. Combine the inducement data with the previous misjudgment data for monitoring and complete the update of the determination.

[0032] The system architecture based on the above intelligent switch cabinet anti-explosion abnormal monitoring method includes a data acquisition module, a data storage and management module, an association analysis module, a time series prediction module, a risk assessment and early warning module, a communication interface module, and a human-computer interaction module.

[0033] The data acquisition module is composed of multiple types of sensor units, signal conditioning units, and edge computing nodes. The sensor units include temperature sensor arrays, characteristic gas sensor groups, partial discharge detectors, current sensors, voltage transformers, and humidity sensors. The signal conditioning unit performs analog-to-digital conversion, filtering and denoising, and standardized output. The edge computing node is deployed on-site in the switch cabinet to realize raw data compression and preprocessing, and transmits data to the upper layer through industrial Ethernet or RS485 bus.

[0034] The data storage and management module constructs a hybrid storage architecture of time series database and relational database. The time series database is responsible for storing real-time monitoring data streams, and the tables are built according to result data and inducement data. The result data table is indexed by temperature, gas concentration, and partial discharge intensity, and the inducement data table is indexed by current, voltage, and humidity. The association is established through timestamp. The relational database is used to store weight model parameters, warning records, negative sample case library, and system configuration information. The module has a built-in data archiving engine that performs data grouping and collection operations at a preset period to build an inducement data mapping table centered on result data.

[0035] The association analysis module realizes dynamic association weight calculation function. The dynamic time warping algorithm is used to calculate the curve similarity of inducement data sequence and result data sequence, and the 0-1 interval association weight value is output. There is a weight update unit that automatically triggers weight recalculation when a new fault case is detected. Gradient descent method is used to optimize weight distribution. The threshold manager is configured to filter weakly associated inducements below the sensitivity threshold, and only keep the high-weight inducement set for subsequent prediction.

[0036] The time series prediction module integrates an autoregressive integrated moving average model and a long short-term memory neural network algorithm. The lag analysis engine automatically performs a cross-correlation function calculation to determine the optimal time lag interval of each cause data relative to the result data, generate a lag parameter table, and the weight time extension unit performs time translation on the historical cause data according to the lag interval to construct a prediction feature vector. The prediction model trainer adopts an online learning method, continuously incorporates new collected data into the training set, updates the model parameters once every 24 hours, and outputs the prediction value and confidence interval of the future 60-second result data.

[0037] The risk assessment and early warning module includes a risk index calculator and a graded early warning generator. The risk index calculator reads the current cause data in real time, substitutes it into the weight model and the time model, and calculates the comprehensive risk score by weighting. The graded early warning generator sets three threshold values (attention, abnormal, and emergency). When the threshold value is exceeded, the sound and light alarm and the SMS gateway are linked, and the trip or ventilation instruction is sent to the execution layer.

[0038] The self-learning optimization module consists of a misjudgment identification unit and a negative sample processing unit. The misjudgment identification unit monitors the result data change within a set period after the early warning. If the expected fluctuation amplitude is not reached, it is determined as a misjudgment, and the complete data sequence is automatically recorded and marked as a negative sample. When the number of negative samples accumulates to the threshold value (default 2 times), the difference analysis engine is started, the feature difference of the cause data of the misjudgment case and the positive sample is extracted, the avoidance rule is injected into the correlation analysis module, and the model is self-corrected.

[0039] The communication interface module realizes the connection with the substation integrated automation system, reserves the MQTT interface for cloud data synchronization, and uses the hardware watchdog and double-redundancy communication link to ensure reliability.

[0040] The human-computer interaction module is based on Web architecture and provides real-time monitoring pictures, trend curve display, early warning information list, model parameter configuration, and historical data query functions, and supports mobile access.

[0041] Each embodiment in the specification is described in a progressive manner, and the same and similar parts of each embodiment can be referred to each other. Each embodiment focuses on the difference from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the part of the method embodiment.

[0042] The above only describes the embodiments of the present application and is not used to limit the present application. The present application can have various changes and modifications for those skilled in the art. Any modification, equivalent replacement, improvement, etc. within the spirit and principles of the present application shall be included in the scope of the claims of the present application.

Claims

1. An intelligent switch cabinet anti-explosion abnormality monitoring method, characterized in that: The method comprises the following steps: The first step: multi-source perception and data collection, comprehensive state detection is performed by deploying multi-source sensors at key nodes inside the switch cabinet; The second step: data correlation modeling and archiving, the correlation between result data and cause data is established, the parameters directly representing the fault results are defined as result data, the potential cause parameters are defined as cause data, and a cause data mapping table with result data as the core is constructed; The third step: dynamic correlation analysis and weight distribution, the curve similarity of each cause data sequence and result data sequence is calculated to quantify the contribution weight of each cause parameter to the fault result; The fourth step: time series modeling and prediction weight calculation step, the time series analysis method is used to determine the best time lag interval of each cause data relative to the result data, and a prediction model is established; The fifth step: real-time risk prediction and early warning step, the comprehensive risk index is calculated based on the weight model and the time effectiveness model to realize graded warning; The sixth step: false judgment feedback and model self-correction, when a judgment error sample occurs, collect negative sample cases, dynamically adjust the weight distribution and warning threshold, and realize self-learning optimization of the model. 2.The intelligent switch cabinet anti-explosion abnormality monitoring method according to claim 1, characterized in that: In the third step, the importance of different cause data under the same result data is calculated, different cause data corresponding to different state result data is selected in the same result data, and a line graph is generated for curvature calculation, and the dynamic correlation strength between the cause and the result is quantified by analyzing the shape similarity of the data curve. 3.The intelligent switch cabinet anti-explosion abnormality monitoring method according to claim 1, characterized in that: In the fourth step, the time series analysis algorithm automatically finds the best time lag interval between different cause data and result data, and gives prediction weights to these lagged cause data.

4. The intelligent switch cabinet anti-explosion abnormality monitoring method according to claim 1, characterized in that: In the sixth step, if no expected change in the corresponding result data is observed within the set time extension interval after a single or multiple cause data is monitored to have a significant fluctuation and a warning is issued accordingly, the system will determine that this warning is a false alarm.

5. The intelligent switch cabinet anti-explosion abnormality monitoring method according to claim 1, characterized in that: In the sixth step, if the number of false alarms exceeds the set number threshold, it is determined as a negative sample case, and the model self-correction will adjust the weight of the specific cause data that leads to the false alarm, increase the weight value of the second weight, and determine again. If the second weight still has a false alarm, compare the data in the negative sample case with the data when the fault occurs in the database, find the same false alarm data and different values of the cause data under the same false alarm data, combine the cause data with the previous false alarm data for monitoring, and update the determination.

6. The intelligent switch cabinet anti-explosion abnormality monitoring system of any one of claims 1 to 5, characterized in that: The method comprises a data acquisition module, a data storage and management module, an association analysis module, a time series prediction module, a risk assessment and warning module, a communication interface module, and a human-computer interaction module. 7.The intelligent switch cabinet anti-explosion abnormality monitoring system according to claim 6, characterized in that: The data acquisition module is composed of multiple types of sensor units, signal conditioning units, and edge computing nodes, and the data storage and management module constructs a hybrid storage architecture of time series database and relational database. 8.The intelligent switch cabinet anti-explosion abnormality monitoring system according to claim 6, characterized in that: The association analysis module realizes dynamic correlation weight calculation function, adopts dynamic time warping algorithm to calculate the curve similarity of cause data sequence and result data sequence, outputs 0-1 interval correlation weight value, and is provided with a weight updating unit. 9.The intelligent switch cabinet anti-explosion abnormality monitoring system according to claim 6, characterized in that: The time series prediction module integrates an autoregressive integrated moving average model and a long short-term memory neural network algorithm, a lag analysis engine automatically performs cross-correlation function calculation, determines the best time lag interval of each cause data relative to the result data, generates a lag parameter table, and a weight time extension unit performs time translation on historical cause data according to the lag interval to construct a prediction feature vector. 10.The intelligent switch cabinet anti-explosion abnormality monitoring system according to claim 6, characterized in that: The risk assessment and early warning module includes a risk index calculator and a hierarchical early warning generator, a self-learning optimization module is composed of a misjudgment identification unit and a negative sample processing unit, a communication interface module realizes the docking with the transformer station integrated self system, a man-machine interaction module is based on a Web architecture, provides real-time monitoring pictures, trend curve display, early warning information list, model parameter configuration and historical data query functions, and supports mobile terminal access.

Images