CN-121984188-A - Fault early warning method for multi-source data fusion of energy storage power station

CN121984188ACN 121984188 ACN121984188 ACN 121984188ACN-121984188-A

Abstract

The invention discloses a fault early warning method for multi-source data fusion of an energy storage power station, which belongs to the technical field of fault early warning of the energy storage power station and comprises the steps of collecting multi-source data of each battery cluster in the energy storage power station, respectively inputting the multi-source data of each battery cluster into a pre-trained risk assessment model to obtain corresponding risk assessment values, respectively determining individual correction degrees of each battery cluster based on cluster health degrees, environment temperatures and temperature and pressure range cooperativity of each battery cluster, respectively correcting preset risk threshold values based on the individual correction degrees to obtain target risk threshold values corresponding to each battery cluster, respectively comparing the risk assessment values of each battery cluster with the corresponding target risk threshold values, determining fault states of each battery cluster, and carrying out fault early warning based on the fault states of each battery cluster. The method can improve the fault early warning accuracy of the energy storage power station.

Inventors

XU RUOCHEN
CAO CHUANZHAO
LEI HAODONG
LIU WEI
LIU MINGYI
SUN ZHOUTING
BAI PANXING
Liu Facan
HAN XU
WEI CHEN
LI QIANG
CAO XI

Assignees

中国华能集团清洁能源技术研究院有限公司
华能广西清洁能源有限公司

Dates

Publication Date: 20260505
Application Date: 20251212

Claims (10)

1. The fault early warning method for multi-source data fusion of the energy storage power station is characterized by comprising the following steps of: The method comprises the steps of collecting multi-source data of each battery cluster in an energy storage power station, wherein the multi-source data comprise a first temperature range, a first temperature range change rate, a first voltage range change rate, a first gas concentration and a first gas concentration change rate, and the first gas comprises one or more of VOC, hydrogen and carbon monoxide; Respectively inputting multi-source data of each battery cluster into a pre-trained risk assessment model to obtain corresponding risk assessment values; Based on the cluster health degree, the ambient temperature and the temperature-pressure range synergy degree of each battery cluster, determining the individual correction degree of each battery cluster respectively, and correcting a preset risk threshold based on each individual correction degree to obtain a target risk threshold corresponding to each battery cluster; And comparing the risk evaluation value of each battery cluster with a corresponding target risk threshold value, determining the fault state of each battery cluster, and performing fault early warning based on the fault state of each battery cluster.
2. The method of claim 1, wherein collecting multi-source data for each battery cluster within an energy storage power station comprises: Collecting a first temperature range, a first voltage range and a first gas concentration of each battery cluster in an energy storage power station; Taking the acquisition time of the first temperature range as the first time, extracting all second temperature ranges in a preset period corresponding to the first time, constructing a first change curve, and taking the slope of the first change curve as the change rate of the first temperature range; Taking the acquisition time of the first voltage range as a second time, extracting all second voltage ranges in a preset period corresponding to the second time, constructing a second change curve, and taking the slope of the second change curve as the change rate of the first voltage range; And taking the collection time of the first gas concentration as a third time, extracting all the second gas concentrations in a preset period corresponding to the third time, constructing a third change curve, and taking the slope of the third change curve as the change rate of the first gas concentration.
3. The method of claim 1, wherein the determining the individual correction of each battery cluster based on the cluster health, ambient temperature, and temperature and pressure differential synergy of each battery cluster, respectively, comprises: For each battery cluster, the following steps are performed: Acquiring cluster health degree, calculating first correction degree corresponding to the cluster health degree, and Acquiring the ambient temperature, calculating a second correction degree corresponding to the ambient temperature, and Obtaining the temperature and pressure range degree of synergy, calculating a third correction degree corresponding to the temperature and pressure range degree of synergy, and And carrying out weighted summation on the first correction degree, the second correction degree and the third correction degree to determine the individual correction degree of the battery cluster.
4. The method of claim 1, wherein the correcting the preset risk threshold based on each individual correction degree to obtain the target risk threshold corresponding to each battery cluster comprises: Correcting a preset risk threshold based on each individual correction degree to obtain an individual risk threshold corresponding to each battery cluster; For each battery cluster, the following steps are performed: comparing the risk evaluation value of each associated battery cluster with a corresponding individual risk threshold value, and taking the associated battery cluster with the risk evaluation value larger than the corresponding individual risk threshold value as a target battery cluster, wherein the associated battery cluster represents the battery cluster with the degree of association between the associated battery cluster and the current battery cluster larger than a preset association threshold value, and And determining group correction degree based on the association degree between each target battery cluster and the current battery cluster, and correcting the individual risk threshold based on the group correction degree to obtain the target risk threshold.
5. The method of claim 4, wherein the determining the population correction based on the inter-cluster association of each target battery cluster with the current battery cluster comprises: normalizing the association degree between each target battery cluster and the current battery cluster, and taking the normalized association degree as a corresponding association weight between clusters; and determining the group correction degree based on the risk evaluation value of each target battery cluster and the corresponding association weight.
6. The method of claim 4, wherein before correcting the preset risk threshold based on each individual correction degree to obtain the target risk threshold corresponding to each battery cluster, further comprises: calculating the association degree between every two battery clusters; Based on the degree of association between the clusters, the associated battery clusters corresponding to the battery clusters are determined.
7. The method of claim 6, wherein calculating the inter-cluster association between the battery clusters comprises: Determining a thermal correlation coefficient based on a spatial position and a thermal conduction relation between a first battery cluster and a second battery cluster, wherein the first battery cluster represents one battery cluster in every two battery clusters, and the second battery cluster represents the other battery cluster in every two battery clusters; determining an electrical association coefficient based on a current relationship and a voltage relationship between the first battery cluster and the second battery cluster; Determining an energy consumption association coefficient based on energy consumption cooperativity between the first battery cluster and the second battery cluster; based on the thermal, electrical, and energy consumption correlation coefficients, an inter-cluster correlation between the first and second battery clusters is determined.
8. The method of claim 7, wherein determining the thermal association coefficient based on the spatial location and the thermal conductivity relationship between the first battery cluster and the second battery cluster comprises: acquiring the linear distance and the absolute value of the temperature difference of the first battery cluster and the second battery cluster; Based on the linear distance and the absolute value of the temperature difference, the thermal correlation coefficient is determined by the following formula: Wherein, the The thermal-related coefficient is characterized by, The straight-line distance is characterized by, The preset reference straight line distance is characterized, The absolute value of the temperature difference is characterized, The preset reference temperature difference is characterized in that, And (3) characterizing an exponential function, and enabling the thermal correlation coefficient to exponentially decay with the linear distance or the temperature difference.
9. The method according to any one of claims 6 to 8, further comprising: Constructing a basic topological framework based on each battery cluster, wherein the basic topological framework comprises a plurality of nodes, and one node corresponds to one battery cluster; establishing a connection edge for each node in a basic topological frame based on the inter-cluster association degree of each battery cluster to obtain an initial topological graph; Configuring a preset color for each node in the initial topological graph based on the fault state of each battery cluster, and configuring a preset color for a connecting edge in the initial topological graph based on the inter-cluster association degree of each battery cluster to obtain a target topological graph; And displaying the target topological graph on a preset display interface.
10. The method as recited in claim 9, further comprising: the method comprises the steps of obtaining fault detail information of each node, wherein the fault detail information comprises fault states, the number of associated battery clusters, an associated battery cluster list and the degree of association among clusters corresponding to connecting edges; Associating each node of the target topology graph with corresponding fault information; Responding to a node clicking instruction of a user, and displaying fault detail information of the node pointed by the node clicking instruction on a preset display interface.

Description

Fault early warning method for multi-source data fusion of energy storage power station Technical Field The invention belongs to the technical field of fault early warning of energy storage power stations, and particularly relates to a fault early warning method for multi-source data fusion of an energy storage power station. Background With the rapid development of new energy industry, the energy storage power station is used as a core infrastructure for stabilizing power grid fluctuation and absorbing renewable energy, and the installed capacity and the application scale of the energy storage power station are continuously enlarged. However, the safe operation of the energy storage power station faces a serious challenge, namely, the battery cluster is used as a core constituent unit of the energy storage power station, and under the influence of factors such as long-term charge-discharge cycle and the like, faults such as thermal runaway, electrical faults, performance decay and the like are easy to occur, so that fire or explosion accidents of the whole station are caused, and great economic loss and safety risks are caused. In order to prevent and control the risks, the existing energy storage power stations mostly adopt a fault early warning method based on the combination of a single parameter and a fixed threshold value, and the fault early warning method is used for carrying out fault early warning by monitoring the temperature, voltage and other electrical parameters of a battery cluster, comparing the electrical parameters with the preset fixed threshold value. However, by adopting the fault early warning method based on the combination of the single parameter and the fixed threshold, the monitoring dimension is single and the threshold is set and solidified, so that the fault early warning accuracy of the energy storage power station is lower. Therefore, a fault early warning method for multi-source data fusion of an energy storage power station is needed, so that the fault early warning accuracy of the energy storage power station can be improved. It should be noted that the information disclosed in the above background section is only for enhancing understanding of the background of the application and thus may include information that does not form the prior art that is already known to those of ordinary skill in the art. Disclosure of Invention The following presents a simplified summary in order to provide a basic understanding of some aspects of the disclosed embodiments. This summary is not an extensive overview, and is intended to neither identify key/critical elements nor delineate the scope of such embodiments, but is intended as a prelude to the more detailed description that follows. The embodiment of the disclosure provides a fault early warning method for multi-source data fusion of an energy storage power station, so that the fault early warning accuracy of the energy storage power station can be improved. In some embodiments, a fault early warning method for multi-source data fusion of an energy storage power station includes: The method comprises the steps of collecting multi-source data of each battery cluster in an energy storage power station, wherein the multi-source data comprise a first temperature range, a first temperature range change rate, a first voltage range change rate, a first gas concentration and a first gas concentration change rate, and the first gas comprises one or more of VOC, hydrogen and carbon monoxide; Respectively inputting multi-source data of each battery cluster into a pre-trained risk assessment model to obtain corresponding risk assessment values; Based on the cluster health degree, the ambient temperature and the temperature-pressure range synergy degree of each battery cluster, determining the individual correction degree of each battery cluster respectively, and correcting a preset risk threshold based on each individual correction degree to obtain a target risk threshold corresponding to each battery cluster; And comparing the risk evaluation value of each battery cluster with a corresponding target risk threshold value, determining the fault state of each battery cluster, and performing fault early warning based on the fault state of each battery cluster. The risk assessment method and the risk assessment device have the advantages that the risk assessment value corresponding to each battery cluster is obtained through collecting the first temperature range, the first temperature range change rate, the first voltage range change rate, the first gas concentration and the first gas concentration change rate of each battery cluster in the energy storage power station and inputting the first temperature range, the first temperature range change rate, the first voltage range change rate, the first gas concentration and the first gas concentration change rate into the pre-trained risk assessment model, so that the risk condition of each battery clu