Early warning method and system for fires in energy storage power stations based on multiple sensors

Abstract

The invention proposes a multi-sensor-based early warning method and system for fires in energy storage power stations. By using multiple sensors and a normal model obtained through the analysis of a large amount of data, it identifies situations that are different from the norm and adopts a hierarchical cumulative decision-making approach to avoid false alarms. At the same time, it adopts different handling methods for different alarm levels, which can also help regulatory personnel quickly understand the stage of the fire and improve the efficiency of early warning.

Description

Technical Field

[0001] This invention relates to the field of energy storage power station supervision, and in particular to a method and system for early warning of fires in energy storage power stations based on multiple sensors. Background Technology

[0002] With the continuous updating and development of science and technology, the emergence of technologies such as computers, the Internet, the Internet of Things, and artificial intelligence has enabled the power industry to move towards automation and intelligence. Among these advancements, energy storage power stations play a crucial role in the development of the power industry. However, due to deficiencies in safety protection measures during the actual application of energy storage power stations, problems may arise in the operation of equipment and systems, seriously affecting the security and stability of power supply.

[0003] Currently, fire prevention methods for energy storage power stations typically include the following two approaches:

[0004] (1) The battery thermal runaway state is detected by monitoring the battery surface temperature and fire smoke concentration. This method is often delayed and may result in false alarms or missed alarms.

[0005] (2) Characteristic gas early warning technology, represented by the identification of thermal runaway gas production such as hydrogen and carbon monoxide, is accurate and reliable. However, it can only identify and issue an early warning after the battery electrolyte has vaporized and dispersed to the vicinity of the gas sensor. The gas diffusion pattern is easily affected by the cabin environment, and the early warning concentration threshold depends on expert experience. Alternatively, AI algorithms can be used to determine the corresponding abnormal threshold through a large number of annotations. The early warning timeliness is prone to fluctuation, and there are problems such as the early warning time not being early enough and the number of false alarms being high.

[0006] In addition, inadequate alarm filtering and non-tiered alarms cause confusion for users. Summary of the Invention

[0007] To address the aforementioned issues, this invention proposes a method and system for early fire warning in energy storage power stations based on multiple sensors.

[0008] The main contents of this invention include:

[0009] Early warning methods for fires in energy storage power stations based on multiple sensors include:

[0010] The system periodically acquires data collected by a sensor group, which includes a pyrolysis particle sensor, an electrical fire sensor, a hydrogen sensor, and a carbon monoxide sensor.

[0011] Based on the data collected by the sensor array, the current environmental state is determined, including both a safe state and a warning state.

[0012] When the current environmental state is a warning state, the growth trend of the data collected by the sensor group at a single time point is predicted by a time window, and the corresponding warning level is calculated based on the predicted value. The warning level includes several sub-warning levels and several sub-alarm levels.

[0013] When the current warning level is a sub-warning level, the current environmental status is determined according to the warning escalation rules and the concentration of pyrolysis particles collected by the pyrolysis particle sensor.

[0014] When the current warning level is a sub-alarm level, the corresponding terminal is driven to issue an alarm according to the alarm escalation rules and the data collected by the sensor group.

[0015] Preferably, determining the current environmental state based on the data collected by the sensor array includes the following steps:

[0016] The concentration of pyrolysis particles collected by the pyrolysis particle sensor is obtained, and it is determined whether the current concentration of pyrolysis particles is lower than the safety threshold. If it is, the current environmental state is determined to be a safe state. If not, the predicted values ​​of the data collected by the electrical fire sensor, hydrogen sensor and carbon monoxide sensor are obtained, and after calculation, it is determined whether they are lower than the alarm threshold. If they are, the current environmental state is determined to be a sub-warning level. If not, the current environmental state is determined to be a sub-alarm level.

[0017] Wherein, the safety setting threshold is the upper limit value of the pyrolysis particle concentration under the safe operation environment of the energy storage power station to be monitored, obtained through the normal environment model; the alarm setting threshold is the concentration value of pyrolysis particles obtained through the abnormal environment model.

[0018] The conventional environment model is an AI model that uses normal environmental data of the safe operation environment of the energy storage power station as the training set; the abnormal environment model is a fire simulation model of the energy storage power station to be tested.

[0019] Preferably, when the current warning level is a sub-warning level, the current environmental state is further determined according to the warning escalation rules and the concentration of pyrolysis particles collected by the pyrolysis particle sensor, including:

[0020] Data collected by the pyrolysis particle sensor is acquired periodically;

[0021] The current sub-warning level is determined based on the predicted value of the concentration of pyrolysis particles collected at a single time point and the number of accumulated points of the corresponding sub-warning level.

[0022] Preferably, the current sub-warning level is determined based on the predicted value of the pyrolysis particle concentration collected at a single time point and the number of accumulated points of the corresponding sub-warning level, including:

[0023] When the predicted value of the concentration of pyrolysis particles collected at a single time point meets the set conditions of the corresponding sub-warning level, it is recorded as an accumulation point of the corresponding sub-warning level. When n consecutive time points are all accumulation points, a corresponding sub-warning level is recorded.

[0024] The number of corresponding sub-warning levels is obtained within a certain period of time. When the number of corresponding sub-warning levels reaches a set value, the corresponding sub-warning level is upgraded to the next level of sub-warning level and recorded as an accumulation point of the next level of sub-warning level.

[0025] Preferably, when the current corresponding warning level is a sub-alarm level, according to the alarm escalation rules, the corresponding terminal is driven to issue an alarm based on the data collected by the sensor group; including:

[0026] Acquire data collected by the sensor group at individual time points at regular intervals;

[0027] The current sub-alarm level is determined based on the predicted value of the data collected by the sensor group at a single time point and the number of accumulated points of the corresponding sub-alarm level.

[0028] Preferably, the current sub-alarm level is determined based on the predicted value of the data collected by the electrical fire sensor at a single time point and the number of accumulated points of the corresponding sub-alarm level, including:

[0029] The predicted value of the data collected by the sensor group at a single time point is compared with the corresponding set threshold. If the predicted value is greater than the alarm set threshold, it is recorded as an accumulation point of the corresponding sub-alarm level.

[0030] The number of accumulated points of the corresponding sub-alarm level within a certain period of time is obtained. When the number of accumulated points of the corresponding sub-alarm level reaches a set value, the corresponding sub-alarm level is upgraded to the next level of sub-alarm level and recorded as an accumulated point of the next level.

[0031] Preferably, the sub-alarm levels include a primary alarm level and a final alarm level;

[0032] When the concentration of pyrolysis particles collected by the sensor group exceeds the preset primary alarm threshold, it is recorded as an accumulation point of a primary alarm level.

[0033] When the concentration of pyrolysis particles collected by the sensor exceeds the preset final alarm threshold, it is recorded as an accumulation point of a final alarm.

[0034] Wherein, the primary alarm threshold refers to the alarm setting threshold; the final alarm threshold refers to the average of the alarm setting threshold and the safety setting threshold; and the alarm standard warning threshold is obtained through the abnormal environment model.

[0035] Preferably, the feature is that when the number of accumulated points of the highest-level sub-warning level reaches a set value, it is recorded as an accumulated point of a primary alarm level.

[0036] This invention also proposes a multi-sensor-based early warning system for fires in energy storage power stations, comprising:

[0037] The data acquisition module includes a pyrolysis particle sensor, an electrical fire sensor, a hydrogen sensor, and a carbon monoxide sensor, which are used to collect environmental data of the energy storage power station to be monitored.

[0038] The processing module, connected to the acquisition module, is used to execute the aforementioned early warning methods;

[0039] A storage module, connected to the processing module, is used to store the accumulated points of the corresponding sub-early warning level and sub-alarm level;

[0040] The alarm module, together with the processing module, issues corresponding alarms based on the corresponding sub-alarm levels.

[0041] The beneficial effects of this invention are as follows: The multi-sensor-based early warning method and system for fires in energy storage power stations proposed in this invention identify abnormal situations through multiple sensors and a normal model obtained through the analysis of a large amount of data. It adopts a hierarchical cumulative decision-making approach to avoid false alarms. At the same time, it adopts different handling methods for different alarm levels, which can also help regulatory personnel quickly know the stage of the fire and improve the efficiency of early warning. Attached Figure Description

[0042] Figure 1 This is a flowchart of the early warning process of the present invention. Detailed Implementation

[0043] The technical solution protected by this invention will be described in detail below with reference to the accompanying drawings.

[0044] Please refer to Figure 1 This invention proposes a multi-sensor-based early warning system for fires in energy storage power stations, comprising a data acquisition module, a processing module, a storage module, and an alarm module. The data acquisition module includes a pyrolysis particle sensor, an electrical fire sensor, a hydrogen sensor, and a carbon monoxide sensor, used to collect environmental data from the energy storage power station under monitoring. The storage module stores pre-set thresholds and the accumulated points of each sub-early warning level and sub-alarm level during monitoring; that is, the storage module can be a fixed storage area connected to an AI model. The storage module can perform calculations and analysis on the environmental data of the monitored energy storage power station according to set escalation accumulation rules to drive the corresponding alarm module and issue appropriate alarms.

[0045] Furthermore, some thresholds are outputs of conventional environmental models. Conventional environmental models refer to AI models that use a large amount of data under normal conditions as their training set. For example, the safety threshold is the upper limit of the pyrolysis particle concentration under safe operating conditions obtained by inputting environmental data of the energy storage power station to be monitored within a certain period of time into the AI ​​model.

[0046] Furthermore, some thresholds, such as alarm setting thresholds, final alarm thresholds, and time window predictions, are designed using an abnormal environment model. This abnormal environment model is created by simulating a fire in the energy storage power station to be monitored. For example, if the energy storage power station to be monitored uses a large number of certain types of wires, a destructive test can be performed on this type of wire in an experimental chamber, and the concentration of pyrolysis particles as the temperature rises can be recorded. The abrupt change in the concentration of pyrolysis particles during the simulation can be used as a parameter to determine the alarm setting threshold. In other application scenarios, the alarm setting threshold will be different depending on the simulated object. In this embodiment, the alarm setting threshold is obtained by simulating the battery fire process. The final alarm threshold is the average of the concentration of pyrolysis particles obtained from the abnormal environment model and the upper limit of the concentration of pyrolysis particles under safe operating conditions output by the conventional environment model.

[0047] Specifically, please refer to Figure 1 The present invention proposes a multi-sensor-based early warning method for fires in energy storage power stations, comprising:

[0048] The system periodically acquires data from a sensor array, which includes a pyrolysis particle sensor, an electrical fire sensor, a hydrogen sensor, and a carbon monoxide sensor. Specifically, it first determines whether the concentration of pyrolysis particles collected by the pyrolysis particle sensor is lower than a preset safety threshold, i.e., whether the concentration is within the normal operating range. The safety threshold is the upper limit of the pyrolysis particle concentration under the safe operating environment of the energy storage power station being monitored. If the concentration of pyrolysis particles exceeds the safety threshold, it indicates that a fire may have occurred in the environment, or it may be due to interference from non-monitoring entities (such as personnel). To achieve early warning and avoid false alarms, when the detected concentration of pyrolysis particles exceeds the safety threshold, the invention first enters an early warning state and then continuously monitors the concentration of pyrolysis particles. Different warning levels are assigned based on the proportion of the exceedance, and different handling methods are adopted according to different warning levels.

[0049] Specifically, this invention categorizes early warning states into two types: monitoring without notification and issuing alarm notifications. In this embodiment, the early warning level includes several sub-early warning levels and several sub-alarm levels. Specifically, in this embodiment, the several sub-early warning levels can include two levels: blue alarm and yellow alarm. For example, when the predicted concentration of pyrolysis particles is 150%-200% of the safety threshold, it is recorded as a blue alarm accumulation point. If three consecutive time points of blue alarm accumulation points are met, it is recorded as a blue alarm meeting the set conditions. It is set that if the number of blue alarms reaches 8 within 24 hours, it is upgraded to the next sub-early warning level, i.e., upgraded to a yellow alarm. Similarly, when the concentration of pyrolysis particles... When the predicted value of the pyrolysis particle concentration is 200-300% of the safety threshold, it is recorded as a yellow alert accumulation point. A yellow alert is also recorded when the predicted value of the pyrolysis particle concentration meets the yellow alert accumulation point condition for three consecutive time points. If the number of yellow alerts reaches 8 within 24 hours, it is upgraded to the next level. In this embodiment, the highest level of the sub-alarm level is the yellow alert, so the next level after the yellow alert is the sub-alarm level that needs to be notified. Specifically, in this embodiment, the sub-alarm levels include the primary alarm level and the final alarm level, defined by color as orange and red. That is, 8 yellow alerts that meet the set conditions can be upgraded to one orange alert, at which point a corresponding notification will be sent to the user terminal.

[0050] Furthermore, when the predicted concentration of pyrolysis particles exceeds the 300% safety threshold, it is determined whether the concentration of pyrolysis particles exceeds the alarm threshold. In this embodiment, the concentration of pyrolysis particles corresponding to the starting point of the surge in the destructive test wire, PE, and PCB is selected as the alarm threshold, i.e., as the orange alarm accumulation point. If the threshold is reached for two consecutive reporting time points, an orange alarm is recorded. At the same time, if every four orange alarms accumulate within 24 hours to upgrade to a red alarm, the orange alarm needs to notify the user. Combining the upgrade rules of the sub-alarm level, the number of orange alarms also includes those upgraded from yellow alarms.

[0051] In this embodiment, the average of the concentration value of pyrolysis particles corresponding to the destructive starting point and the upper limit value of the concentration of pyrolysis particles under safe operating conditions is selected as the final alarm threshold, that is, as the red alarm accumulation point. If the data at two consecutive time points exceeds the final alarm threshold, it is recorded as a red alarm, and an alarm is immediately sent to the user to ensure that the message can be delivered.

[0052] The notification methods for orange and red alerts can be set by the user, prioritizing the reporting of higher-level warnings. By performing time window predictions on the collected data and comparing the predicted values ​​with the set thresholds, potential hazards can be detected early. At the same time, the strategy of accumulating warning points avoids false alarms. Selecting different notification methods according to different alarm levels can promptly report warnings of higher danger levels, which is also more conducive to staff quickly assessing the fire situation and responding more effectively.

[0053] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for early fire warning in energy storage power stations based on multiple sensors, characterized in that, include: The system periodically acquires data collected by a sensor group, which includes a pyrolysis particle sensor, an electrical fire sensor, a hydrogen sensor, and a carbon monoxide sensor. Based on the data collected by the sensor array, the current environmental state is determined, including both a safe state and a warning state. When the current environmental state is a warning state, the growth trend of the data collected by the sensor group at a single time point is predicted by a time window, and the corresponding warning level is calculated based on the predicted value. The warning level includes several sub-warning levels and several sub-alarm levels. When the current warning level is a sub-warning level, the current environmental status is determined according to the warning escalation rules and the concentration of pyrolysis particles collected by the pyrolysis particle sensor. When the current corresponding warning level is a sub-alarm level, according to the alarm escalation rules, the corresponding terminal is driven to issue an alarm based on the data collected by the sensor group; Based on the data collected by the sensor array, the current environmental state is determined, including the following steps: The concentration of pyrolysis particles collected by the pyrolysis particle sensor is obtained, and it is determined whether the current concentration of pyrolysis particles is lower than the safety threshold. If it is, the current environmental state is determined to be a safe state. If not, the predicted values ​​of the data collected by the electrical fire sensor, hydrogen sensor and carbon monoxide sensor are obtained, and after calculation, it is determined whether they are lower than the alarm threshold. If they are, the current environmental state is determined to be a sub-warning level. If not, the current environmental state is determined to be a sub-alarm level. Wherein, the safety setting threshold is the upper limit value of the pyrolysis particle concentration under the safe operation environment of the energy storage power station to be monitored, obtained through the normal environment model; the alarm setting threshold is the concentration value of pyrolysis particles obtained through the abnormal environment model. The normal environment model is an AI model that uses normal environment data of the safe operation environment of the energy storage power station as the training set; the abnormal environment model is a fire simulation model of the energy storage power station to be tested.

2. The method for early fire warning of energy storage power stations based on multiple sensors according to claim 1, characterized in that, When the current warning level is a sub-warning level, the current environmental state is further determined according to the warning escalation rules and the concentration of pyrolysis particles collected by the pyrolysis particle sensor, including: Data collected by the pyrolysis particle sensor is acquired periodically; The current sub-warning level is determined based on the predicted value of the concentration of pyrolysis particles collected at a single time point and the number of accumulated points of the corresponding sub-warning level.

3. The method for early fire warning of energy storage power stations based on multiple sensors according to claim 2, characterized in that, Based on the predicted concentration of pyrolysis particles collected at a single time point and the number of accumulated points for the corresponding sub-warning level, the current sub-warning level is determined, including: When the predicted value of the concentration of pyrolysis particles collected at a single time point meets the set conditions of the corresponding sub-warning level, it is recorded as an accumulation point of the corresponding sub-warning level. When n consecutive time points are all accumulation points, a corresponding sub-warning level is recorded. The number of corresponding sub-warning levels is obtained within a certain period of time. When the number of corresponding sub-warning levels reaches a set value, the corresponding sub-warning level is upgraded to the next level of sub-warning level and recorded as an accumulation point of the next level of sub-warning level.

4. The method for early fire warning of energy storage power stations based on multiple sensors according to claim 2, characterized in that, When the current warning level is a sub-alarm level, according to the alarm escalation rules, based on the data collected by the sensor group, the corresponding terminal is driven to issue an alarm; including: Acquire data collected by the sensor group at individual time points at regular intervals; The current sub-alarm level is determined based on the predicted value of the data collected by the sensor group at a single time point and the number of accumulated points of the corresponding sub-alarm level.

5. The method for early fire warning of energy storage power stations based on multiple sensors according to claim 4, characterized in that, Based on the predicted values ​​of the data collected by the electrical fire sensor at a single time point and the number of accumulated points for the corresponding sub-alarm level, the current sub-alarm level is determined, including: The predicted value of the data collected by the sensor group at a single time point is compared with the corresponding set threshold. If the predicted value is greater than the alarm set threshold, it is recorded as an accumulation point of the corresponding sub-alarm level. The number of accumulated points of the corresponding sub-alarm level within a certain period of time is obtained. When the number of accumulated points of the corresponding sub-alarm level reaches a set value, the corresponding sub-alarm level is upgraded to the next level of sub-alarm level and recorded as an accumulated point of the next level.

6. The method for early fire warning of energy storage power stations based on multiple sensors according to claim 5, characterized in that, The sub-alarm levels include primary alarm levels and final alarm levels; When the concentration of pyrolysis particles collected by the sensor group exceeds the preset primary alarm threshold, it is recorded as an accumulation point of a primary alarm level. When the concentration of pyrolysis particles collected by the sensor exceeds the preset final alarm threshold, it is recorded as an accumulation point of a final alarm. Wherein, the primary alarm threshold refers to the alarm setting threshold; the final alarm threshold refers to the average of the alarm setting threshold and the security setting threshold.

7. The method for early fire warning of energy storage power stations based on multiple sensors according to claim 6, characterized in that, When the number of accumulated points at the highest-level sub-warning level reaches the set value, it is recorded as an accumulated point at the primary-level alarm level.

8. A multi-sensor-based early warning system for fires in energy storage power stations, characterized in that: include: The data acquisition module includes a pyrolysis particle sensor, an electrical fire sensor, a hydrogen sensor, and a carbon monoxide sensor, which are used to collect environmental data of the energy storage power station to be monitored. A processing module, connected to the acquisition module, is used to execute the early warning method as described in any one of claims 1 to 7; A storage module, connected to the processing module, is used to store the accumulated points of the corresponding sub-early warning level and sub-alarm level; The alarm module, together with the processing module, issues corresponding alarms based on the corresponding sub-alarm levels.

Images