Electric power early warning blocking method for identifying fire occurrence

Abstract

The invention relates to the technical field of fire early warning, in particular to an electric power early warning blocking method for identifying fire occurrence. On the basis of the obtained electric power early warning signal, continuously receiving electrical parameter data uploaded by sensors deployed at all power grid nodes, carrying out abnormal fluctuation detection on the electrical parameter data, judging the abnormal type of the electric power parameter data, and setting an early warning parameter set of the electrical parameter data; judging an abnormal rule conforming to the early warning parameter set in a preset time period, and setting a blocking threshold value of each power grid node by using the power grid node corresponding to the abnormal rule; performing adjacent region analysis on the power grid nodes with the blocking threshold values, and setting a hierarchical blocking mode; connecting the power grid nodes corresponding to the hierarchical blocking mode to form a blocking path, and completing blocking processing after early warning of the power grid nodes; the timeliness of power early warning blocking response and the fire early warning efficiency are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of fire early warning, and in particular to a method for identifying and blocking power early warning of fire. Background Art

[0002] Traditional power fire early warning systems typically rely on fixed thresholds or single-parameter monitoring, resulting in delayed response, high false alarm rates, and an inability to accurately locate the fire source. Due to the complex topology of power grids and the strong interconnectedness of equipment, a single node anomaly can trigger a chain reaction. Traditional methods lack the ability to dynamically analyze the propagation paths of anomalies, resulting in crude blocking strategies that can easily cause widespread power outages or inadequate protection of critical areas.

[0003] For example, Chinese Patent Publication No. CN112330910A discloses an information processing method, device, storage medium, and processor for fire early warning. The method includes: a fire monitoring platform obtaining a first early warning message sent by a fire early warning platform, wherein the first early warning message is uploaded to the fire early warning platform by a fire early warning device; the fire monitoring platform obtaining a review message corresponding to the first early warning message, wherein the review message is obtained by the fire early warning platform reviewing the first early warning message; and the fire monitoring platform determining whether to retain the first early warning message based on the review message.

[0004] For example, Chinese Patent Publication No. CN112991658A discloses a fire warning method and uninterruptible power supply. The method includes the following steps, executed according to a preset cycle: adjusting the operating state of the uninterruptible power supply according to a preset method under the current operating state of the uninterruptible power supply; generating a first cumulative factor based on the actual temperature parameters within the cabinet before and after the adjustment of the operating state of the uninterruptible power supply and pre-stored reference temperature parameters under the same operating conditions; generating a fire warning index based on the first cumulative factor, and outputting an alarm signal when the fire warning index reaches a preset value. The present invention provides a fire warning by actively adjusting the operating state of the uninterruptible power supply, comparing the actual temperature change parameters before and after the adjustment with the pre-stored reference temperature change parameters, and generating a fire warning index based on the first cumulative factor.

[0005] In the existing technology, fire information processing is completed by reviewing the fire warning information, and the power supply status is used with temperature as a reference factor to identify the temperature change of electronic components in the current fire scenario, so as to realize the temperature change rate of electronic components under fire warning; however, the existing technology tends to process single data anomalies, and it is necessary to comprehensively identify the current and other parameters of related equipment in the historical data, and it is necessary to use the abnormal composite method to perform topological hierarchical perception of the nodes to realize comprehensive control of multiple groups of nodes, complete the abnormal propagation and blocking settings of multiple power grid nodes in the fire warning scenario, and ultimately reduce the continuous damage caused by the fire due to equipment operation. Summary of the Invention

[0006] To solve the above technical problems, the present invention adopts a technical solution: a method for identifying fire-related power warning and blocking, comprising: S1, based on the acquired power warning signal, continuously receiving electrical parameter data uploaded by sensors deployed at each power grid node. The electrical parameter data includes voltage, current, and temperature.

[0007] S2, detect abnormal fluctuations in the electrical parameter data, determine the abnormal type of the power parameter data, measure the source location of the power warning signal based on the abnormal type, and set a warning parameter set for the electrical parameter data.

[0008] S3, determining the abnormal rules that the warning parameter set complies with within the preset time period, and using the power grid nodes corresponding to the abnormal rules to set the blocking threshold of each power grid node.

[0009] S4, conducts adjacent area analysis on the grid node with the blocking threshold set, verifies whether the status identifier of the nearest grid node corresponding to the current grid node is consistent, and if so, sets a hierarchical blocking method according to the consistent number; if not, traverses the blocking thresholds of each grid node in multiple time intervals, and analyzes the abnormal propagation path under the hierarchical blocking method; based on the analysis results of the abnormal propagation path, sets a hierarchical blocking method.

[0010] S5, connecting the grid nodes corresponding to the hierarchical blocking mode to form a blocking path, and completing the blocking process of the grid nodes after the early warning.

[0011] The beneficial effects of the present invention are: 1. The present invention continuously collects voltage, current, and temperature parameters, calculates the parameter change rate in combination with the working status of the equipment, identifies abnormal fluctuations by comparing the sliding average with historical data, and uses context mark weighting to determine the abnormality type under different trigger conditions. The warning parameters of the abnormal data are matched with the target value through the instantaneous change rate in the form of current-voltage-temperature multi-parameter association based on data such as the frequency of the abnormal type, and combined with the time classification analysis of the abnormal type to improve the matching accuracy of power warnings under fire.

[0012] 2. The present invention generates a list of adjacent nodes based on the power grid topology, adopts a subtype inheritance mechanism to realize dynamic comparison of blocking thresholds, optimizes the blocking range through blocking coverage calculation and secondary blocking methods, and quantifies the relative position distribution and path identification of each power grid node in the power blocking scenario, so as to complete the power warning for the common abnormal scenario of multiple nodes in the abnormal propagation path of parameter abnormalities under the hierarchical blocking method, so as to achieve the accuracy of power warning.

[0013] 3. The present invention improves the accuracy of risk trend prediction for multiple regions by setting weights for blocking path nodes, and obtains the frequency in multiple regions by regional embedding, and detects the number of weight changes through a sliding window, thereby completing the processing of power warning trend changes. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The present invention will be further described below with reference to the accompanying drawings and examples.

[0015] Figure 1 The present invention is a flowchart of a method for identifying a fire-related power early warning and blocking method.

[0016] Figure 2 The present invention is a flow chart of step S2 of a method for identifying a fire and preventing power from being pre-warned.

[0017] Figure 3 The present invention is a flow chart of step S3 of a method for identifying a fire and preventing power from being pre-warned.

[0018] Figure 4 The present invention is a flow chart of step S4 of a method for identifying a fire and preventing power from being pre-warned. DETAILED DESCRIPTION

[0019] The following embodiments of the present invention are described in detail. The embodiments described below are exemplary and are only used to explain the present invention, and are not to be construed as limiting the present invention. Where specific techniques or conditions are not specified in the embodiments, the techniques or conditions described in the literature in the art or in the product specifications shall be followed.

[0020] See Figure 1 A method for identifying and blocking power supply warnings for fire occurrences includes: S1, based on an acquired power warning signal, continuously receiving electrical parameter data uploaded by sensors deployed at each power grid node. The electrical parameter data includes voltage, current, and temperature.

[0021] S2, detect abnormal fluctuations in the electrical parameter data, determine the abnormal type of the power parameter data, measure the source location of the power warning signal based on the abnormal type, and set a warning parameter set for the electrical parameter data.

[0022] S3, determining the abnormal rules that the warning parameter set complies with within the preset time period, and using the power grid nodes corresponding to the abnormal rules to set the blocking threshold of each power grid node.

[0023] S4, conducts adjacent area analysis on the grid node with the blocking threshold set, verifies whether the status identifier of the nearest grid node corresponding to the current grid node is consistent, and if so, sets a hierarchical blocking method according to the consistent number; if not, traverses the blocking thresholds of each grid node in multiple time intervals, and analyzes the abnormal propagation path under the hierarchical blocking method; based on the analysis results of the abnormal propagation path, sets a hierarchical blocking method.

[0024] S5, connecting the grid nodes corresponding to the hierarchical blocking mode to form a blocking path, and completing the blocking process of the grid nodes after the early warning.

[0025] Electrical parameter data such as current, voltage, power, leakage current, temperature and other parameters are collected through temperature sensors, voltmeters, ammeters and other equipment to collect relevant data of each power grid node when a fire occurs. The power parameter data received by the power grid node can be used to quickly determine the area that the fire can affect. After confirmation, a power warning is issued as soon as possible and the relevant power supply is automatically switched to prevent the fire from expanding and reduce secondary hazards.

[0026] Preferably, the above-mentioned power warning signal can be a signal issued when an abnormality is identified in any part of the electrical parameter data, or it can be a warning signal issued when a fire is identified. Based on this warning signal, the possible warning form and location can be identified, and then the relevant data is continuously checked as the electrical parameter data for subsequent main analysis.

[0027] Preferably, the equipment represented by the above-mentioned power grid nodes include but are not limited to power generation equipment, transformers, circuit breakers, disconnectors, transmission lines, distribution cabinets and other equipment. These devices will obtain the relative status of each node in the current power grid by configuring corresponding current, voltage and temperature sensors to explain how to block them in the event of a fire.

[0028] When receiving electrical parameter data, step S1 is implemented in a manner that further includes: extracting the rate of change of the electrical parameter data of each grid node according to the uninterrupted working status of each grid node, and comparing them with the preset warning threshold in turn; when the preset warning threshold is met, the electrical parameter data corresponding to each grid node is output according to the triggering condition of the current power warning signal.

[0029] Preferably, the trigger conditions are divided into electrical parameter anomalies and fire confirmation signals. When the electrical parameter data is identified as belonging to any of the above conditions, the grid node at the corresponding location will be monitored. As for the fire confirmation signal, it is the presence of a fire or smoke that is recognized by the sensors configured on the grid node. In addition to temperature, current, and voltage sensors, the sensors configured at this time also include smoke alarms and cameras to collect the working conditions of the equipment at the locations configured at each grid node. The smoke alarms and cameras then transmit the data to the fire confirmation signal section, with the electrical parameter anomalies transmitted by the temperature, current, and voltage sensors as the basis for the current monitoring of the data of each grid node. As long as there is data that meets the current trigger conditions, the electrical parameter data of each grid node will be processed to identify the relevant characteristics of the fire, so as to determine how to cut off the power supply according to the corresponding characteristics when a fire occurs.

[0030] Preferably, the above-mentioned preset warning threshold will be based on the average of the upper limit values ​​of temperature, voltage and current of each grid node under normal operation in historical data as its preset warning threshold to indicate that any one of the current, voltage or temperature of the current grid node is abnormal.

[0031] In one embodiment of the present invention, in step S2, it is mainly used to detect abnormal fluctuations in electrical parameter data to obtain the abnormal type of the current electrical parameter, such as short circuit, overload and other abnormal types, and analyze the time sequence of the received electrical parameter data. Combined with the correlation of multiple abnormal types that appear under fire, the abnormal type is extracted by correlation analysis and data measurement to extract the currently identified abnormal data.

[0032] like Figure 2 As shown, the implementation of step S2 includes: using the time series of electrical parameter data, taking the sliding average of the dimension in which the electrical parameter data is located, comparing the electrical parameter data with historical data, using the current dimension as an influencing factor, and determining the parameter thresholds of different influencing factors before and after the fire occurs; the above-mentioned dimensions represent the currently collected temperature, voltage, and current data divided into multiple dimensions, and anomaly identification is performed for each dimension based on the parameter threshold. The above-mentioned parameter threshold is represented by the average value of the historical data, which is used to determine the sliding average value of the current electrical parameter data in the same dimension.

[0033] Obtain the context flags of the current influencing factors and determine the weights of the influencing factors based on the context flags. The context flags identify the status of the corresponding equipment when the power warning signal is currently issued, such as peak power consumption periods, equipment start and stop events, etc., which indicate the status of the current equipment when the warning occurs. These states are used as the basis for determining whether the current electrical parameters are abnormal.

[0034] Based on the weights and parameter thresholds of the influencing factors, the abnormality type of the current influencing factor is identified. The identified abnormality types may include current overload, harmonics, flicker, and other data abnormalities that indicate equipment at the grid node.

[0035] Preferably, the context mark will set relevant values ​​based on historical data. For example, when judging the peak power consumption period, if the currently collected current exceeds 80% of the average peak power consumption period in the historical data, the current corresponding influencing factor will be marked as peak period or non-peak period, and 1 or 0 will be used to represent its relative situation. The start and stop of the equipment will indicate the length of time the equipment starts and stops. These data will be normalized or otherwise used to mark the currently collected electrical parameters.

[0036] Therefore, the implementation method of judging the weight of the influencing factor based on the context flag also includes: obtaining the number of influencing factors of the current power grid node, summarizing the context flags according to the number of influencing factors, and setting the weight of the current influencing factor according to the average characteristic value of the context flags after summarization; when setting the weight of the current influencing factor, the context flags that appear on a single dimension are summarized. For example, after the current collection is currently used to identify multiple flags for the current power grid node, such as the peak period, equipment startup time, large current fluctuations, etc., the preset weight of each context flag in the database is averaged to obtain the weight of the current influencing factor. As for the number of influencing factors, it represents the gradual processing of the dimensions currently identified to have abnormalities, and the abnormal type of the influencing factor is used to identify whether the current and voltage in these dimensions are abnormal.

[0037] Preferably, the implementation method for identifying the abnormal type of the current influencing factor includes: clustering the electrical parameter data based on the parameter threshold of the influencing factor to obtain multiple clustered abnormal types. Sorting the abnormal types according to the weight of the influencing factor to obtain the order of abnormal types under different influencing factors. For example, the short-circuit cluster characteristics: a sudden increase in current (>300% of the rated value), a rapid rise in temperature (>80℃ / s); the overload cluster characteristics: a continuous high current (120%-150% of the rated value), a slow rise in temperature (<10℃ / s); clustering the currently identified data to obtain different abnormal types. Further, if in high-weight scenarios such as peak hours, the overload-related abnormal types will be determined first, and then the abnormal types identified as instantaneous fluctuations or harmonics will be marked.

[0038] After clustering data with similar patterns, the clustered values ​​are directly compared with the contents represented by short-circuit cluster characteristics, overload cluster characteristics, etc. to determine the abnormality type. The abnormality type determined in this step is used to illustrate obvious problems in the nature of current, voltage, and temperature. It can also be further compared with power reference data and historical data to obtain the current obvious abnormality type from data sources such as databases.

[0039] When setting the warning parameter set for the power warning signal, the corresponding power reference data is further compared with the historical data based on the source location of the power warning signal, such as a specific sensor, device node or regional power grid, to preliminarily determine the current abnormal range; and output is based on the data within the abnormal range.

[0040] Therefore, the implementation method of setting the warning parameter set for the power warning signal also includes: measuring the electrical parameter data based on the co-occurrence frequency of the abnormal type in the power grid node, and analyzing the current, voltage and temperature of the electrical parameter data in turn based on the time when the electrical parameter data after measurement appears; respectively obtaining the current abnormality type, voltage abnormality type and temperature abnormality type in the area where the power grid node is located, and storing the corresponding data in the warning parameter set according to the processing method of the current abnormality type, voltage abnormality type and temperature abnormality type.

[0041] When handling current anomalies, if it's determined to be an overload, a flexible power outage is triggered, such as gradually reducing the load over 30 seconds. If it's a short circuit, the power is immediately cut off. When handling voltage anomalies, if the voltage is overvoltage (>110% of the rated voltage), the voltage regulator is activated. If it's undervoltage (<90% of the average voltage), the backup power source is switched. For temperature anomalies, if the temperature exceeds a threshold, such as 60°C, the cooling system or current limiting protection is activated.

[0042] For current anomalies, the warning thresholds are set to 1.5 times the rated current for overload and 3 times the rated current for short-circuit. These thresholds are used to determine the abnormal current condition and the thresholds used to identify the abnormal condition. For voltage anomalies, the voltage fluctuation range is set to ±10%, and a response time of 50ms triggers voltage anomaly processing. For temperature anomalies, a temperature threshold can be set, using the maximum allowable operating temperature of the equipment at the grid node as the threshold. For example, a conductor temperature >70°C triggers an alarm. This sets a warning threshold for the identified anomaly, associates the warning threshold with the electrical parameter data, and stores it in the warning parameter set.

[0043] The multiple numerical ranges described above are schematic representations of current grid node processing. The specific settings of the numerical ranges need to be adjusted according to the device type of the grid node to achieve monitoring and processing of multiple devices on the grid node.

[0044] In one embodiment of the present invention, when describing the abnormal rules of the current warning parameter set, the compliance degree of each grid node is calculated based on historical data and real-time load, and then the blocking threshold set for each grid node is defined using the compliance degree of each grid node.

[0045] When checking the abnormal rules in step S3, the instantaneous change rate in the warning parameter set is used to identify whether there are sudden increases and decreases in voltage and current, and when the instantaneous change rate changes in a single time period, the abnormal rule matching is performed to obtain historical data similar to the current power grid node, and the current data situation under the abnormal rules is output. If the abnormal rules are overload, short circuit, etc., the source of the power grid node is marked according to the location, indicating the maximum allowable power-off time and minimum safe current of the current power grid node as the blocking threshold of the current power grid node.

[0046] like Figure 3 As shown, the implementation method of step S3 also includes: based on the warning parameter set within the preset time period, extracting the warning threshold of the warning parameter set, and using the warning threshold to compare with the real-time electrical parameter data to obtain the instantaneous change rate of the electrical parameter data under different warning thresholds; at this time, by using the warning threshold set under different abnormality types, the electrical parameter data is divided into multiple data sets, and the data change rate of each data set within the preset time period is checked, that is, the instantaneous change ratio value of temperature, current and voltage within the preset time period. The length of the preset time period can be set to 5 minutes, which is used to continuously monitor the working conditions of each node in the power grid. If some data continues to process short circuit, overload, etc., the relevant nodes will be blocked, and the maximum allowable power-off time and minimum safe current of each power grid node will be extracted to block and control the power grid nodes.

[0047] Events are classified based on the time point corresponding to the instantaneous change rate of the electrical parameter data, and anomaly rule matching is performed based on the anomaly type after event classification.

[0048] Based on the results of abnormal rule matching, the maximum allowable power outage duration and minimum safe current of the current grid node are extracted as the blocking threshold of the current grid node.

[0049] Preferably, when classifying events, the abnormality types in the electrical parameter data are arranged in the order of time points to illustrate the abnormality types that can appear together in a short period of time in the current scenario of fire or fire warning, and the abnormality rules are matched according to the order of appearance of these abnormality types and the time period to illustrate whether the current current, voltage and temperature related abnormalities belong to the scenario of fire.

[0050] That is, when matching abnormal rules, its implementation method includes: converting the instantaneous change rate and time point of the electrical parameter data under the abnormal type into a vector representation, and matching the vector with the preset rule event, using vector cross multiplication to calculate the matching degree. If the matching degree is 0, it is judged as a complete match, and matching is performed one by one at the time point after the event is classified to complete the abnormal rule matching results of the electrical parameter data under different warning thresholds.

[0051] The preset rule event represents a set of data sets of multiple abnormal types of current, voltage and temperature in the scenario of fire occurrence or impending fire. This data set represents a composite data set describing the occurrence of a fire scenario. The preset rule event will set multiple sets of data sets according to the type of power grid node to complete whether the current power grid node has corresponding abnormal conditions under multiple abnormal type combinations.

[0052] After completing the exception rule matching and extracting the maximum allowable power-off duration and minimum safe current of the current grid node, it is necessary to call the corresponding exception types for the current, voltage and temperature of the current electrical parameter data respectively, and combine the called exception types to complete the setting of the blocking threshold and other parts.

[0053] That is, the implementation method of abnormal rule matching also includes: based on pre-set preset rule events, performing abnormal rule matching on electrical parameter data, calling the abnormal type of current, obtaining the target current value, and matching the instantaneous change rate of the target current value in the current abnormal type and the warning threshold. When the current current value is not at the target current value and the matching degree is less than the preset threshold, the abnormal type of the current current is regarded as the result of the current abnormal rule matching.

[0054] After calling the abnormal type of current, the current abnormal rule matching is directed to the abnormal rule matching of voltage. Based on the result of the current abnormal rule matching, the lower limit of the current threshold and the upper limit of the current threshold are obtained. In the scenario where the abnormal type of voltage exists in the current power grid node, the time difference between the instantaneous change rate of voltage being greater than the voltage drop threshold and the current reaching the upper limit of the current threshold is judged. If it is met, a composite abnormality is triggered, and the abnormal type of current and the abnormal type of voltage are combined and regarded as the result of the abnormal rule matching of current and voltage; otherwise, it is regarded as the result of only the abnormal rule matching of current.

[0055] The result of the current anomaly rule matching is directed to the temperature anomaly rule matching. In the scenario where there is a temperature anomaly type at the current power grid node, the temperature anomaly type in which the current reaches the lower limit of the current threshold is determined, and the corresponding temperature anomaly type is combined with the result of the current anomaly rule matching. The multiple combined data pointed to by the result of the current anomaly rule matching are regarded as the output anomaly rule matching result.

[0056] Based on the obtained composite abnormal rule matching, the maximum allowable power outage duration and minimum safe current corresponding to the current scenario are extracted from the database.

[0057] Preferably, the above-mentioned lower limit of the current threshold and upper limit of the current threshold indicate that the current is in the abnormal range value under the current abnormal type scenario, and this value indicates the threshold of the abnormal setting of the current current; the lower limit of the current threshold is used to prevent false alarms, avoid false compound abnormalities caused by short-term current fluctuations, and reduce false blocking; the upper limit of the current threshold is to define serious abnormalities to ensure that high-risk events are handled immediately to avoid the spread of fire.

[0058] Preferably, the target current value is obtained when the abnormal type of current is called, and the target current value is used to represent the target value under a single abnormal type, such as overload is usually set to 1.2-1.5 times the rated current, short circuit is usually set to more than 3 times the rated current, and the total harmonic distortion of the harmonic passing current is greater than 5%, etc. The target current value is used to illustrate that the current current value is in the main judgment value of the corresponding abnormal type, and then the subsequent description of the current current value is not in the target current value, which is used to indicate that the current current value is already in the abnormal range where the current target current value is located. For example, when judging overload, the current current value is not in the range of less than 1.2 times the rated current; as for the calculation method of the matching degree, the form of vector cross product is adopted. In the scenario where the matching degree is less than the preset threshold, it is explained that the current is abnormal and meets the current abnormal type. As for the preset threshold, the average value of the matching of the current abnormality rules in the historical data can be used as the preset threshold set at this time.

[0059] Preferably, when the abnormal rule matching result of current is directed to the abnormal type of voltage, it is mainly judged whether the instantaneous change ratio of voltage is too large or too small. When the voltage is greater than the voltage drop threshold, it means that in addition to the current abnormality, the current voltage is also synchronously abnormal. For example, the voltage drop threshold can be set to 20% to illustrate the change of voltage in unit time when the current changes. If the time difference between the abnormal voltage and current is less than a certain value, it means that the two are close to the concurrent time of the same abnormality, and meet the composite abnormality. For example, if the interval between the abnormalities of the two is less than 10s, it is considered that the currently identified voltage abnormality type and current abnormality type are concurrent events under the fire scenario, and a combined analysis is required to query the blocking method required for the current scenario.

[0060] Preferably, the voltage drop threshold is set based on the average value of voltage drops occurring at the current grid node in historical data to determine the voltage drop threshold at this time.

[0061] Preferably, the average value of the time interval in which the voltage and current in the historical data simultaneously meet the requirements that the instantaneous rate of change of the voltage is greater than the voltage drop threshold and the current reaches the upper limit of the current threshold is used as the current judgment standard. When it is less than the average value of the time interval, it is considered to meet the composite abnormality situation.

[0062] Preferably, when the abnormal rule matching result of the current points to the abnormal type of temperature, it is mainly judged that when the current current begins to be abnormal, the temperature has already met the corresponding abnormal type scenario, and the abnormal type of current and the abnormal type of temperature at this time are combined as the basis for subsequent search for blocking methods. After that, the data of the combination of the three abnormal types of current, voltage and temperature after the abnormal rule matching is used as the output abnormal rule matching result, and according to the result, the maximum allowable power-off duration and minimum safe current that can be selected in the current multi-abnormal type scenario are queried from the database to complete the control and processing of each power grid node.

[0063] The exception rules obtained in step S3 focus on the contextual flags and thresholds of multiple exception type combinations compared to the exception types, and determine whether the risks in the complex scenario are fire risks. The rules also describe the current, voltage, and temperature in the current electrical parameter data in the form of structured thresholds, timing patterns, and vector combinations, and explain the data form of fires that are about to or are occurring, so that subsequent power grid nodes can play a role in fire warning and blocking based on the currently acquired data.

[0064] In one embodiment of the present invention, when setting a hierarchical blocking method, the status and relative situation of the grid nodes are obtained by checking the identification of each grid node. The status identification represents the abnormality type and blocking threshold of the corresponding node, as well as the equipment and location represented by each grid node. These data are used as the status identification of each grid node.

[0065] In the case of inconsistency, it is necessary to determine the propagation path related to the exception type, such as Figure 4 As shown, the implementation of step S4 includes: analyzing multiple groups of adjacent grid nodes, and organizing each grid node into a neighboring node list based on the grid topology; each row of data in the neighboring node list represents a group of adjacent grid nodes.

[0066] The abnormal type of each group of power grid nodes in the adjacent node list is identified. If any abnormal type of the current power grid node belongs to a subtype of the adjacent power grid node, starting from the abnormal type of the adjacent power grid node, the number of subtypes of the abnormal type of the current power grid node is identified, the abnormal type of the current power grid node is connected with the abnormal type of the adjacent power grid node, and the blocking thresholds of the adjacent power grid node and the current power grid node are compared. If the blocking thresholds are consistent, the blocking methods of the current power grid node and the adjacent power grid node are regarded as the same level of hierarchical blocking methods.

[0067] If any of the anomaly types at the current grid node doesn't fall under the subtype of a neighboring grid node, the current grid node's anomaly type is compared with that of the neighboring nodes, with the current grid node designated as the termination node. A hierarchical blocking method is then selected based on the current grid node's blocking threshold. When setting a hierarchical blocking method, current, voltage, and other data are blocked based on the hierarchical blocking methods of multiple current grid nodes to prevent fires and further damage.

[0068] When setting the hierarchical blocking mode of the same level, if the blocking thresholds are inconsistent, the hierarchical blocking mode is set according to the blocking threshold of the adjacent grid node, and the current node is set to the secondary blocking mode of the hierarchical blocking mode of the adjacent grid node.

[0069] Preferably, when judging whether the status identification is consistent, each group of adjacent grid nodes is compared one by one. If there are more than 50% abnormal labels in the grid nodes with adjacent relationships and the blocking threshold is the same in multiple time periods, it is judged to be consistent, and a blocking mode is set for the current grid node and the adjacent grid nodes. For example, the blocking mode can be divided into three levels. The first-level blocking extends the blocked grid node to multiple adjacent grid nodes of the adjacent grid node. The second-level blocking uses the current grid node and the adjacent grid node as the blocking content. The third-level blocking is to block only the current grid node, and then block in real time with the blocking threshold queried by the database.

[0070] Preferably, if there is an inconsistent scenario, it is determined whether the adjacent power grid node and the current power grid node are in a parent-child relationship. Then, when the blocking threshold is consistent with the adjacent node, its blocking area is merged. If it is inconsistent, the current power grid node is regarded as a secondary blocking method to merge the propagation methods of multiple abnormal types when blocking.

[0071] Only for nodes that do not belong to subtypes, time interval traversal is used, triggered by the abnormal type of the current power grid node, and reverse tracing is performed to identify the power grid nodes that have a topological connection relationship with the current power grid node. After checking the blocking threshold of each power grid node in a reverse tracing manner, the hierarchical blocking method is set in sequence according to the value range of the blocking threshold, and the blocking level of each power grid node is set.

[0072] As for the setting method, by using the status identifier of the current power grid node as the search condition, a composite search is performed to check the database for blocking methods that meet multiple conditions such as the current power grid node abnormality type, blocking threshold, etc., and then the hierarchical blocking methods are set one by one.

[0073] When outputting the hierarchical blocking mode, the output will be made one by one in the order of each level according to the level set when the hierarchical blocking mode is set.

[0074] After completing the setting of the hierarchical blocking method, it is necessary to gradually check the hierarchical blocking methods set for multiple power grid nodes in the adjacent node list, that is, the implementation method of step S4 also includes: checking the hierarchical blocking method of each power grid node, identifying the blocking coverage rate under each level for the hierarchical blocking method, starting from any power grid node where the blocking starts, identifying multiple other power grid nodes connected to the currently blocked power grid node, and using the blocking coverage rate under the hierarchical blocking method to output the control commands corresponding to the hierarchical blocking method in sequence.

[0075] At this time, the levels set for the hierarchical blocking method are used. The ratio of the number of abnormal grid nodes covered under each level to the total number of abnormal nodes is used as the basis for sorting. The blocking method under each level is output from large to small to complete the current blocking processing for abnormal situations.

[0076] In one embodiment of the present invention, when forming a blocking path, multiple grid nodes need to be connected, and the number of times each grid node in the blocking path is marked is used to capture each blocked area after the power warning is blocked to complete multi-area identification of power blocking.

[0077] Therefore, the implementation method of step S5 includes: performing data judgment on the power grid nodes on the blocking path, setting the counting weight of the power grid nodes based on the time when each power grid node appears in the blocking path and the graded blocking method, and regionally embedding and associating the power grid nodes. Starting from any power grid node on the current blocking path, the number of changes in the counting weight of each power grid node in multiple time periods is judged. If the number of changes exceeds the preset threshold, the current blocking path is marked and output as power grid warning information.

[0078] Preferably, when implementing step S5, the first blocking time and the most recent blocking time of the grid node will be recorded, and the difference in the time dimension of the corresponding grid node under two or more processings will be checked. Then, the identification of the current grid node as a first-level blocking, a second-level blocking or a third-level blocking will be checked, and the weight of each grid node will be calculated at this time.

[0079] For example, the weights of the first, second, and third level blocking are set to 2.0, 1.0, and 0.5, respectively, to illustrate the degree of verification under the current blocking process. The weights are then set using the blocking time difference of the grid node under multiple blocking conditions. The weights can be expressed as ;in, represents the counting weight, Indicates the initial weight, which is set based on whether the current grid node is in the first-level blocking, second-level blocking or third-level blocking. represents the attenuation coefficient, where the attenuation coefficient is set to 0.05, which is used to weaken the difference between the first blocking time and the latest blocking time in a short period of time, to emphasize the smoothness of the overall blocking method and retain the relative weight of each grid node under long-term risks; Indicates the time difference between the first blocking time and the most recent blocking time. When calculating this value, the current time difference is processed in a normalized form to eliminate the dimension when calculating the weight; Represents an exponential constant. This counting weight is used to identify whether the blocking mode of the current grid node exhibits time decay over different time periods. This facilitates subsequent traversal of the abnormalities experienced by multiple grid nodes on the blocking path over multiple time periods, facilitating subsequent prediction of the current grid device status in the event of a fire risk.

[0080] Preferably, when performing regional embedding association, multiple power grid nodes are clustered according to the coordinates of the regions in which they are located to identify the location of the region where the current blocking path is located. Then, by comparing whether the counting weights have numerical changes in multiple time periods, it is described whether the region where the current power grid node is located is a high-frequency occurrence region. If the counting weight value changes, it means that the corresponding power blocking event still occurs in the corresponding region, and a mark needs to be set on the relevant blocking path. For example, if a node changes more than 2 times in 3 consecutive windows, a regional-level warning is triggered; in this way, sudden blocking fluctuations caused by short circuits, lightning strikes, etc. are quickly identified, and the changing trend in the corresponding time period is described in the form of a time series to achieve preventive maintenance and complete the adaptive optimization of power grid node blocking.

[0081] Although the embodiments of the present invention have been shown and described above, it is understood that the above embodiments are illustrative and are not to be construed as limitations on the present invention. A person skilled in the art may change, modify, replace and modify the above embodiments within the scope of the present invention, and these changes shall still be within the scope of protection of the present invention.

Claims

1. A method for identifying fire-related power early warning and blocking, characterized in that: include: S1, based on the acquired power warning signal, continuously receives electrical parameter data uploaded by sensors deployed at each grid node. The electrical parameter data includes voltage, current, and temperature. S2, detecting abnormal fluctuations in the electrical parameter data, determining the abnormal type of the power parameter data, and measuring the source location of the power warning signal based on the abnormal type, and setting a warning parameter set for the electrical parameter data; S3, determining the abnormal rules that the warning parameter set meets within the preset time period, and using the grid nodes corresponding to the abnormal rules to set the blocking threshold of each grid node; S4: Analyze the adjacent areas of the grid node for which the blocking threshold is set to verify whether the status identifiers of the nearest grid nodes corresponding to the current grid node are consistent. If they are consistent, set a hierarchical blocking method according to the consistent number; if they are inconsistent, traverse the blocking thresholds of each grid node in multiple time intervals to analyze the abnormal propagation path under the hierarchical blocking method; set a hierarchical blocking method based on the analysis results of the abnormal propagation path; S5, connecting the grid nodes corresponding to the hierarchical blocking mode to form a blocking path, and completing the blocking process of the grid nodes after the early warning.

2. A method for identifying fire occurrence and power early warning and blocking according to claim 1, characterized in that: When receiving electrical parameter data, step S1 may also include: According to the uninterrupted working status of each grid node, the change rate of the electrical parameter data of each grid node is extracted and compared with the preset warning threshold in turn. When the preset warning threshold is met, the electrical parameter data corresponding to each grid node is output according to the trigger condition of the current power warning signal.

3. A method for identifying fire occurrence and power early warning and blocking according to claim 1, characterized in that: The implementation of step S2 includes: Using the time series of electrical parameter data, the sliding average of the dimension where the electrical parameter data is located is used to compare the electrical parameter data with historical data. The current dimension is used as the influencing factor to determine the parameter thresholds of different influencing factors before and after the fire occurs. Obtain the context flag of the current influencing factor and determine the weight of the influencing factor based on the context flag; Based on the weights and parameter thresholds of the influencing factors, identify the abnormal types of the current influencing factors.

4. A method for identifying fire occurrence and power early warning and blocking according to claim 3, characterized in that: Methods for determining the weight of influencing factors based on contextual markers also include: The number of influencing factors of the current power grid node is obtained, the context flags are summarized according to the number of influencing factors, and the weight of the current influencing factor is set according to the average characteristic value of the context flags after the aggregation.

5. The method for identifying fire occurrence and power early warning and blocking according to claim 1, characterized in that: The implementation method of setting the warning parameter set of the power warning signal also includes: The electrical parameter data is metered based on the co-occurrence frequency of the abnormality type at the power grid node, and the current, voltage and temperature of the electrical parameter data are analyzed in sequence based on the appearance time of the measured electrical parameter data. The current abnormality type, voltage abnormality type and temperature abnormality type in the area where the power grid node is located are obtained respectively, and the corresponding data are stored in the early warning parameter set according to the processing method of the current abnormality type, voltage abnormality type and temperature abnormality type.

6. A method for identifying fire occurrence and power early warning and blocking according to claim 1, characterized in that: The implementation of step S3 further includes: Based on the warning parameter set within a preset time period, the warning threshold of the warning parameter set is extracted, and the warning threshold is compared with the real-time electrical parameter data to obtain the instantaneous change rate of the electrical parameter data under different warning thresholds; Events are classified based on the time point corresponding to the instantaneous change rate of electrical parameter data, and anomaly rule matching is performed based on the anomaly type after event classification; Based on the results of abnormal rule matching, the maximum allowable power outage duration and minimum safe current of the current grid node are extracted as the blocking threshold of the current grid node.

7. A method for identifying fire occurrence and power early warning and blocking according to claim 6, characterized in that: Exception rule matching can also be implemented by: Based on pre-set rule events, the electrical parameter data is matched with abnormal rules, the abnormal type of the current is called, the target current value is obtained, and the instantaneous change rate of the target current value in the current abnormal type is matched with the warning threshold. If the current current value is not within the target current value and the matching degree is less than the preset threshold, the abnormal type of the current current is regarded as the result of the current abnormal rule matching; Direct the current anomaly rule matching to the voltage anomaly rule matching. Based on the result of the current anomaly rule matching, obtain the current threshold lower limit and the current threshold upper limit. In the scenario where the current grid node has a voltage anomaly type, determine the time difference between the instantaneous rate of change of the voltage being greater than the voltage drop threshold and the current reaching the current threshold upper limit. If the difference is met, a composite anomaly is triggered. The combination of the current anomaly type and the voltage anomaly type is considered the result of the current and voltage anomaly rule matching; otherwise, it is considered the result of only the current anomaly rule matching. The result of the current anomaly rule matching is directed to the temperature anomaly rule matching. In the scenario where there is a temperature anomaly type at the current power grid node, the temperature anomaly type in which the current reaches the lower limit of the current threshold is determined, and the corresponding temperature anomaly type is combined with the result of the current anomaly rule matching. The multiple combined data pointed to by the result of the current anomaly rule matching are regarded as the output anomaly rule matching result.

8. The method for identifying fire occurrence and power early warning and blocking according to claim 1, characterized in that: The implementation of step S4 includes: Analyze multiple groups of adjacent grid nodes and organize each grid node into a list of adjacent nodes based on the grid topology. Each row of data in the adjacent node list represents a group of adjacent grid nodes. Identify the abnormality type of each group of grid nodes in the adjacent node list. If any abnormality type of the current grid node belongs to a subtype of the adjacent grid node, then starting from the abnormality type of the adjacent grid node, identify the number of subtypes of the abnormality type of the current grid node, connect the abnormality types of the current grid node with those of the adjacent grid nodes, and compare the blocking thresholds of the adjacent grid nodes with the current grid node. If the blocking thresholds are consistent, the blocking methods of the current grid node and the adjacent grid nodes are regarded as the same level of hierarchical blocking methods; When setting the hierarchical blocking mode of the same level, if the blocking thresholds are inconsistent, the hierarchical blocking mode is set according to the blocking threshold of the adjacent grid node, and the current node is set to the secondary blocking mode of the hierarchical blocking mode of the adjacent grid node; If any abnormal type of the current grid node does not belong to the subtype of the adjacent grid node, the abnormal type of the current grid node is compared with the adjacent grid nodes, the current grid node is made the termination node, and the hierarchical blocking method is selected based on the blocking threshold of the current grid node.

9. The method for identifying fire occurrence and power early warning and blocking according to claim 8, characterized in that: The implementation of step S4 further includes: Check the hierarchical blocking mode of each power grid node, identify the blocking coverage rate under each level for the hierarchical blocking mode, start from any grid node where the blocking starts, identify multiple other grid nodes connected to the currently blocked grid node, and use the blocking coverage rate under the hierarchical blocking mode to output the control commands corresponding to the hierarchical blocking mode in sequence.

10. The method for identifying fire occurrence and power early warning and blocking according to claim 1, characterized in that: The implementation of step S5 includes: Data judgment is performed on the power grid nodes on the blocking path. The counting weight of the power grid nodes is set according to the time when each power grid node appears in the blocking path and the hierarchical blocking method. The power grid nodes are regionally embedded and associated. Starting from any power grid node on the current blocking path, the number of changes in the counting weight of each power grid node in multiple time periods is judged. If the number of changes exceeds the preset threshold, the current blocking path is marked and output as power grid early warning information.

Images