Fire treatment system and method applied to rail transit lithium battery energy storage power station

Abstract

The invention discloses a fire treatment system and method applied to a rail transit lithium battery energy storage power station. The system mainly comprises a sensor module, a Bayesian network reasoning module, a gas fire extinguishing system, a water-based fire extinguishing system, a waste liquid treatment system and a control module. The sensors collect data such as temperature, smoke concentration, combustible gas concentration and humidity in the power station in real time and transmit the data to the control module, and the control module conducts probabilistic reasoning analysis on the real-time data according to a Bayesian network model and predicts the fire risk. At the initial stage of a fire, the system starts the gas fire extinguishing device and sprays perfluorohexanone for local suppression; and if the fire behavior expands, a high-pressure water mist system is switched to, and large-area fire extinguishing is carried out. Meanwhile, the waste liquid treatment system is started, waste liquid generated by fire extinguishing is collected, components of the waste liquid are monitored and treated, and it is ensured that emission meets the environmental protection standard. The system has the advantages of accurate early warning, efficient fire extinguishing, environment-friendly waste liquid treatment and the like, and is particularly suitable for safety protection of industrial energy storage power stations such as rail transit.

Description

Technical Field

[0001] The present invention belongs to the technical field of energy storage power stations, and in particular relates to a fire handling system and method for a lithium battery energy storage power station applied to rail transit. Background Art

[0002] With the continuous advancement of global industrial energy conservation and emission reduction and the application of renewable energy, rail transit, as a large-scale public infrastructure, faces huge energy management challenges and pressure to meet carbon emission reduction targets. The application of lithium battery energy storage systems has gradually become a key means for rail transit to optimize energy management, reduce electricity costs, and improve energy utilization. However, lithium batteries are at risk of thermal runaway under extreme conditions. Once a fire occurs, a large amount of toxic and harmful gases will be produced, and harmful waste liquid will be produced after the fire is extinguished, which seriously threatens environmental safety and the normal operation of rail transit. Therefore, the safety of lithium battery energy storage systems has become a focus of attention. Designing an intelligent fire warning and waste liquid treatment system is crucial to the safe operation of rail transit lithium battery energy storage power stations.

[0003] In the prior art, patent CN117854216 adopts a traditional alarm module, which integrates sound and gas signals to warn of fires in energy storage power stations. The sound measuring instrument installed in the energy storage compartment collects sound signals in real time. By analyzing these sound signals, the system can identify the opening action of the safety valve, and automatically cut off the power to the battery to issue a first-level alarm, and use the hydrogen sensor in the energy storage compartment to monitor the hydrogen concentration in real time. When the monitored hydrogen concentration reaches the preset threshold, the system will immediately cut off the power to the battery to issue a second-level alarm.

[0004] In the prior art, for example, patent CN116459470 uses a combination of gas and liquid fire extinguishing methods to carry out fire rescue inside the energy storage power station, and arranges a fire hydrant fire extinguishing system outside to improve the safety of the energy storage power station. However, it does not consider how to deal with the toxic and harmful liquids generated after the fire is extinguished.

[0005] In summary, the traditional fire warning system of energy storage power stations not only has disadvantages in real-time data processing, intelligence level and warning accuracy, but the existing fire extinguishing system also does not fully consider the treatment of fire extinguishing waste liquid. It is urgent to design an intelligent and integrated fire prevention and waste liquid treatment system to improve the safety, reliability and environmental protection of energy storage power stations. Summary of the invention

[0006] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and provide a fire handling system and method for rail transit lithium battery energy storage power station.

[0007] The purpose of the present invention can be achieved by the following technical solutions:

[0008] On the one hand, the present invention provides a fire handling system for a lithium battery energy storage power station for rail transit, the system comprising: an energy storage battery, a sensor module, a fire extinguishing system, a waste liquid treatment system and a control module, the fire extinguishing system comprising a gas fire extinguishing system and a water-based fire extinguishing system, the sensor module is connected to the control module, the control module is connected to the gas fire extinguishing system, the water-based fire extinguishing system and the waste liquid treatment system, the control module is deployed with a Bayesian network, the Bayesian network is used to predict the probability of fire discrimination, the control module is used to judge the working state of the energy storage battery according to the fire discrimination probability, and control the corresponding fire extinguishing system and waste liquid treatment system to work according to the working state of the energy storage battery.

[0009] Furthermore, the sensor module includes a voltage sensor, a smoke sensor, a temperature sensor, a combustible gas sensor, and a humidity sensor. The sensor module is used to collect fire data and transmit the fire data to the control module. The fire data is collected through various sensors of the sensor module, including the energy storage battery voltage, the indoor smoke concentration of the energy storage power station, the indoor temperature of the energy storage power station, the indoor combustible gas concentration of the energy storage power station, and the indoor humidity of the energy storage power station.

[0010] Furthermore, the working state of the energy storage battery includes a normal working state, an early stage of thermal runaway and a late stage of thermal runaway.

[0011] Furthermore, the control module functions include: determining whether the indoor smoke concentration and indoor temperature of the energy storage power station are greater than a first preset threshold and a second preset threshold, respectively; if one is greater than the preset threshold, controlling the start of a water-based fire extinguishing system and a waste liquid treatment system; if neither is greater than the preset threshold, no treatment is performed.

[0012] Furthermore, the gas fire extinguishing system is a prefabricated perfluorohexanone fire extinguishing system, which is connected and controlled by a control module and is used to start fire extinguishing when the energy storage battery is in the early stage of thermal runaway; the water-based fire extinguishing system is a high-pressure fine water mist fire extinguishing system, which is connected and controlled by a control module and is used for fire extinguishing. The water-based fire extinguishing system will produce fire extinguishing waste liquid when extinguishing a fire.

[0013] Furthermore, the waste liquid treatment system includes a waste liquid collection module and a waste liquid component real-time monitoring module. The waste liquid treatment system and the water-based fire extinguishing system are started simultaneously. The waste liquid collection module includes a waste liquid collection pipe and a waste liquid tank. The waste liquid collection pipe is made of stainless steel pipe. The outer wall of the stainless steel pipe is painted with anti-corrosion paint. The waste liquid collection pipe is installed with an electric valve, and the electric valve is connected and controlled by a control module. The waste liquid collection pipe is used to guide the fire extinguishing waste liquid into the waste liquid tank. The waste liquid tank is made of stainless steel, and the inner wall of the waste liquid tank is anti-seepage treated. The waste liquid component real-time monitoring module includes a PH sensor and a heavy metal ion detection sensor, which are used to monitor the pH value and heavy metal ion concentration of the waste liquid in real time. The PH sensor and heavy metal ion detection sensor are installed on the inner wall of the waste liquid tank.

[0014] Another aspect of the present invention provides a fire treatment method for a rail transit lithium battery energy storage power station, comprising the following steps:

[0015] Step S1: construct a Bayesian network for predicting the probability of fire identification through historical data, and deploy the trained Bayesian network in the control module;

[0016] Step S2: Collect fire data through the sensor module, including: energy storage battery voltage, indoor smoke concentration of the energy storage power station, indoor temperature of the energy storage power station, indoor combustible gas concentration of the energy storage power station and indoor humidity of the energy storage power station, and transmit the fire data to the control module;

[0017] Step S3: the control module inputs the fire data into the Bayesian network, the Bayesian network outputs the fire discrimination probability, and the control module determines the working state of the energy storage battery according to the fire discrimination probability;

[0018] Step S4: the control module selects a corresponding fire extinguishing system according to the working state of the energy storage battery, and sends a control signal to the corresponding fire extinguishing system and waste liquid treatment system; the control module determines whether the indoor smoke concentration and indoor temperature of the energy storage power station are greater than the first preset threshold and the second preset threshold respectively, and if one is greater than the preset threshold, the control module sends a control signal to the water-based fire extinguishing system and the waste liquid treatment system, and if neither is greater than the preset threshold, no processing is performed;

[0019] Step S5: After receiving the control signal, the fire extinguishing system starts to start and extinguish the fire, and after receiving the control signal, the waste liquid treatment system starts to collect waste liquid.

[0020] Furthermore, in step S1, constructing a Bayesian network for predicting the probability of fire identification through historical data includes the following steps:

[0021] Obtain historical voltage, temperature, smoke concentration, combustible gas concentration and indoor humidity data of the energy storage power station, including data under normal operation and abnormal conditions;

[0022] Preprocessing the acquired data, including removing noise and outliers, formatting the data into an input form suitable for a machine learning algorithm, and labeling the data;

[0023] The noise removal comprises the following steps:

[0024] The temperature data and voltage data are subjected to noise reduction by using the moving average method;

[0025]

[0026] Where T(t) is the temperature or voltage value at time t, and n is the window length, which can be adjusted according to the system's requirements for fluctuation sensitivity;

[0027] Exponential smoothing method is used to denoise the humidity, smoke concentration and combustible gas concentration data. The formula is as follows:

[0028]

[0029] Where H(t) is the humidity value or combustible gas concentration value at time t, and α is the smoothing coefficient;

[0030] The input format suitable for machine learning algorithms includes:

[0031] Normalize the temperature and voltage data by Z-score, the formula is:

[0032]

[0033] where x i is the temperature value at time i, μ is the data mean, and σ is the data standard deviation;

[0034] The collected gas concentration data is logarithmically normalized, and the formula is as follows:

[0035] x1=log(x+1)

[0036] Where x is the gas concentration;

[0037] The collected humidity data is processed by Min-Max standardization, and the formula is as follows:

[0038]

[0039] Where x is the current humidity value, x min =0,x max =1;

[0040] The parameters of the Bayesian network are learned by using the maximum likelihood estimation method through the preprocessed data to obtain the trained Bayesian network.

[0041] Furthermore, in step S3, the control module inputs the fire data into the Bayesian network, and the Bayesian network outputs the fire discrimination probability. The fire discrimination probability formula is:

[0042]

[0043] Among them, P(S,T,G,H|F) is the conditional probability of the currently detected smoke, temperature, combustible gas concentration and humidity when a fire occurs, P(F) is the prior probability of fire occurrence, P(S,T,G,H) is the marginal probability of sensor data, and P(F|S,T,G,H) is the fire discrimination probability output by the Bayesian network.

[0044] Furthermore, the control module determines the working state of the energy storage battery according to the fire identification probability, including:

[0045] When the fire judgment probability is ≤0.1, the control module determines that the energy storage battery is in normal working condition and does not trigger the warning;

[0046] When the fire identification probability is between 0.1 and 0.3, the control module determines that the energy storage battery is in the early stage of thermal runaway, and the control module sends a control signal to the gas fire extinguishing system;

[0047] When the fire identification probability is between 0.3 and 1, the control module determines that the energy storage battery is in the late stage of thermal runaway, and the control module sends a control signal to the water-based fire extinguishing system and the waste liquid treatment system.

[0048] Compared with the prior art, the present invention has the following advantages:

[0049] (1) The present invention has significant advantages over the prior art, especially for the deficiencies of lithium battery energy storage power stations for rail transit in fire warning, fire extinguishing response and waste liquid treatment, and has made innovative improvements. First, the present invention realizes real-time fire probability reasoning based on multi-dimensional data (such as voltage, temperature, smoke concentration, combustible gas concentration, humidity, etc.) through the Bayesian network model, which can accurately warn in the early stage of thermal runaway of lithium batteries, and dynamically adjust the warning threshold according to historical data, significantly improving the accuracy of the warning. Secondly, the present invention adopts an intelligent fire extinguishing system, which can switch the fire extinguishing means according to the different stages of the fire: perfluorohexanone gas is used to suppress the fire source in the early stage, and then switched to a high-pressure fine water mist system in the later stage, which effectively reduces the risk of fire spread and avoids the damage of water-based fire extinguishing to the battery. Finally, in view of the problem of imperfect waste liquid treatment in traditional fire extinguishing systems, the present invention designs a waste liquid collection and monitoring system, which can effectively collect, monitor and treat the waste liquid generated during the fire extinguishing process, ensuring that the waste liquid discharge meets environmental protection standards, and is particularly suitable for industrial scenarios such as rail transit that have high requirements for the environment.

[0050] (2) The present invention is applied to rail transit lithium battery energy storage power stations, and has intelligent fire warning, automatic fire extinguishing and waste liquid treatment functions. Through real-time data collection by Bayesian network models and multiple sensors, the system can accurately predict the risk of thermal runaway of lithium batteries, and use edge computing technology to realize local data processing and improve response speed. The fire extinguishing system intelligently selects gas or water-based fire extinguishing methods according to the fire stage to effectively control the fire. The waste liquid treatment system automatically collects the waste liquid generated by fire extinguishing, and monitors the composition of the waste liquid through sensors to ensure that it meets environmental protection requirements. The system integrates intelligent warning, fire extinguishing and environmental protection treatment, improving the safety and reliability of energy storage power stations. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is a schematic diagram of the overall system architecture of the present invention;

[0052] Figure 2 This is a schematic diagram of the overall layout of the energy storage power station sensors, fire extinguishing equipment and waste liquid collection of the present invention;

[0053] Figure 3 The present invention is a flow chart of the steps of a fire warning and waste liquid treatment system applied to a rail transit lithium battery energy storage power station. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0055] Embodiment 1:

[0056] This embodiment proposes the following technical solution: a fire warning and waste liquid treatment system applied to rail transit lithium battery energy storage power stations, including a sensor module, a Bayesian network reasoning module, a gas fire extinguishing system, a water-based fire extinguishing system, a waste liquid treatment system and a control module.

[0057] The fire data corresponding to the fire data sources such as voltage, smoke concentration, temperature, combustible gas concentration and environmental humidity in the energy storage power station are collected in real time through sensors; the sensors include: voltage sensor, smoke sensor, temperature sensor, combustible gas sensor, humidity sensor; the Bayesian network of energy storage power station fire is constructed based on historical data, and the collected fire data is input into the Bayesian network to obtain the fire discrimination probability; when the system determines that the fire discrimination probability reaches the preset threshold, the fire extinguishing device system is automatically started. The fire extinguishing system includes a gas fire extinguishing system and a water-based fire extinguishing system, and the gas fire extinguishing system is activated with a delay of 0-5s according to the extremely early stage of the fire, and the water-based fire extinguishing system is activated in the late stage of the fire with a delay of 0-30s. The nozzle adopts a swirl nozzle structure. After the water-based fire extinguishing system is started, the waste liquid treatment system will be automatically activated to prevent the waste liquid generated during the fire extinguishing process from entering other sensitive areas of the power station or polluting the surrounding environment.

[0058] The fire data sources include energy storage battery voltage, smoke concentration, temperature, combustible gas concentration and ambient humidity; the construction of the Bayesian network is based on experimental data and historical data, and the data includes voltage, combustible gas, humidity and temperature. After calculation by the Bayesian network, the prior probability and conditional probability of the lithium battery in the normal working stage, the early stage of thermal runaway and the late stage of thermal runaway are obtained; the system collects voltage, combustible gas, smoke and temperature data in real time, and calculates the posterior probability of the stage the lithium battery is in based on the prior probability and conditional probability, and judges the stage of the lithium battery based on the calculated posterior probability; when the fire discrimination probability exceeds the threshold, the system triggers a fire warning signal and initiates emergency treatment measures.

[0059] The waste liquid treatment system includes a waste liquid collection module and a waste liquid composition real-time monitoring module; the waste liquid collection module includes a dedicated waste liquid collection pipeline and a waste liquid tank; when the water-based fire extinguishing system is started, the waste liquid treatment system is activated at the same time, and the pipeline automatically guides the fire extinguishing waste liquid into the waste liquid tank. The waste liquid collection pipeline must have corrosion resistance and high temperature resistance. Stainless steel pipes are used, and anti-corrosion paint is painted on the outer wall of the stainless steel pipe. Then bricks and concrete are used to build grooves and wrap the stainless steel pipe; an electric valve is installed on the waste liquid collection pipeline. After receiving the start signal, the pipeline valve automatically opens to transport the waste liquid generated during the fire extinguishing process to the waste liquid tank. The waste liquid tank is made of 316L stainless steel, and the inner wall of the waste liquid tank has been treated for anti-seepage;

[0060] The real-time monitoring module includes a pH sensor and a heavy metal ion detection sensor. During the waste liquid collection process, the real-time monitoring module monitors the waste liquid composition (such as pH value, heavy metal ion concentration, etc.) in real time through the pH sensor and the heavy metal ion detection sensor to ensure that the collected waste liquid meets the environmental emission standards; if it is detected that the collected waste liquid does not meet the standards, the waste liquid in the waste liquid tank will be temporarily stored and then sent to a designated industrial waste liquid treatment station for centralized treatment.

[0061] The collected temperature, voltage, combustible gas concentration and humidity information are preprocessed as follows:

[0062] 1) Obtain historical voltage, temperature and gas data of the energy storage power station. The data should cover normal operation and abnormal conditions (such as battery failure, overheating, gas leakage, etc.);

[0063] 2) Clean the historical data, remove noise and outliers, format it into an input form suitable for machine learning algorithms, and provide annotations for the data;

[0064] 3) The collected data is first subjected to denoising, and the temperature data and voltage data are subjected to denoising using the moving average method;

[0065] 4) The Bayesian network model uses the pgmpy library for structure learning and parameter learning, uses the constrained PC algorithm for structure learning, and uses the maximum likelihood estimation method for parameter learning of the data.

[0066] The present invention adopts a method combining gas and liquid fire extinguishing systems, which can effectively control the fire of energy storage power station, quickly extinguish open flames, stop the fire from spreading, and prevent battery fire and explosion.

[0067] After the Bayesian network is built, the Bayesian network model is serialized and deployed to the control module. The selected control module model can be an embedded device NVIDIA Jetson Nano with sufficient computing power. During deployment, the trained Bayesian network model is serialized into a file and saved to the device through the joblib tool. When the device is running, Jetson Nano connects to the sensor module through the GPIO or I2C interface to collect key fire data in the energy storage power station, such as temperature, humidity, combustible gas concentration and voltage. Through the reasoning results of the Bayesian network, the probability of fire occurrence and the stage of battery thermal runaway are judged. When the probability exceeds the set threshold, the control module will trigger a fire warning signal and automatically start the corresponding fire extinguishing and waste liquid treatment system.

[0068] The gas fire extinguishing system is a prefabricated perfluorohexanone fire extinguishing system. When the fire is in the early stage of battery thermal runaway, the system automatically triggers the gas fire extinguishing to locally suppress the fire source. The perfluorohexanone is in accordance with the group standard T / CECS 10171-2022 "Prefabricated perfluorohexanone fire extinguishing device". The spray volume of perfluorohexanone is 3L, the spray time is 20 seconds, the spray interval is 30 seconds, and a total of 15 sprays; the system is also equipped with emergency start and emergency stop switches;

[0069] The water-based fire extinguishing system is a high-pressure water mist fire extinguishing system. When the early warning device detects that the fire is in the late stage of battery thermal runaway, it automatically switches to the water-based fire extinguishing system.

[0070] The water supply pipeline of the high-pressure fine water mist fire extinguishing system is equipped with an electric ball valve and a rising stem gate valve, and the electric ball valve is automatically opened when the early warning signal is detected to be in the late stage of thermal runaway.

[0071] Furthermore, the waste liquid collection needs to consider the maximum estimated amount of waste liquid M in a fire accident to be at least 30%-50% more, M=N+L;

[0072] Where N is the estimated volume of the converted transfer chemicals, and L is the total water output within 1.0 hour of firefighting time.

[0073] Embodiment 2:

[0074] The parts not mentioned in this embodiment are the same as those in Embodiment 1.

[0075] This embodiment provides a fire warning and waste liquid treatment system for rail transit lithium battery energy storage power stations, including real-time collection of fire data corresponding to fire data sources such as smoke concentration, temperature, combustible gas concentration and ambient humidity in the energy storage power station through sensors; constructing a Bayesian network for energy storage power station fires based on historical data, inputting the collected fire data into the Bayesian network, and obtaining the probability of fire discrimination; when the system determines that the fire risk reaches a preset threshold, the fire extinguishing device is automatically started. The fire extinguishing system includes a gas fire extinguishing system and a water-based fire extinguishing system, and selects a suitable type of fire extinguishing system according to different stages of the fire. The system monitors the fire in real time and adjusts the fire extinguishing means when the fire changes. After the high-pressure fine water mist fire extinguishing system is started, the waste liquid treatment system will be automatically activated to prevent the waste liquid generated during the fire extinguishing process from entering other sensitive areas of the power station or polluting the surrounding environment.

[0076] This embodiment greatly increases the reliability of fire warning of rail transit energy storage power stations by combining early warning, fire extinguishing, and waste liquid collection systems, improves the fire extinguishing capabilities of energy storage power stations, and thus helps to improve the safety of energy storage power stations.

[0077] In this embodiment, the fire data source including smoke concentration, temperature, combustible gas concentration and ambient humidity is collected as experimental data; a Bayesian network based on voltage, combustible gas, humidity and temperature is constructed, and the prior probability and conditional probability of the lithium battery in the normal working stage, the very early stage of thermal runaway, the early and middle stage of thermal runaway and the late stage of thermal runaway are obtained through Bayesian network calculation, such as Figure 1 As shown;

[0078] This embodiment performs data preprocessing on the collected temperature, voltage, combustible gas concentration and humidity information in the following manner:

[0079] The system collects the following data in real time through sensors installed in the lithium battery energy storage power station: voltage, combustible gas concentration, smoke concentration, temperature and humidity; and pre-processes the collected data in the following ways:

[0080] S1. Obtain historical voltage, temperature and gas data of the energy storage power station. The data should cover normal operation and abnormal conditions (such as battery failure, overheating, gas leakage, etc.);

[0081] S2. Clean the historical data, remove noise and outliers, format it into an input form suitable for machine learning algorithms, and provide annotations for the data;

[0082] S3, firstly perform denoising on the collected data, and use the moving average method to perform denoising on the temperature data and voltage data;

[0083]

[0084] Where T(t) is the temperature or voltage value at time t, and n is the window length, which can be adjusted according to the system's requirements for fluctuation sensitivity;

[0085] Exponential smoothing method is used to denoise the humidity and combustible gas concentration data. The formula is as follows:

[0086]

[0087] Where H(t) is the humidity value or combustible gas concentration value at time t, and α is the smoothing coefficient;

[0088] The denoised temperature and voltage data are normalized by Z-score so that the machine learning algorithm can effectively process the data. The formula is as follows:

[0089]

[0090] where x i is the temperature value at time i, μ is the data mean, and σ is the data standard deviation;

[0091] The collected gas concentration data is logarithmically normalized, and the formula is as follows:

[0092] x1=log(x+1)

[0093] Where x is the gas concentration;

[0094] The collected humidity data is processed by Min-Max standardization, and the formula is as follows:

[0095]

[0096] Where x is the current humidity value, xmin =0,x max =1;

[0097] After structural learning, the maximum likelihood estimation method is used to learn the parameters of the data. The steps are as follows:

[0098] Assume that the structure of the Bayesian network is known, the data set contains n samples, and each sample contains the observed values ​​of all variables;

[0099] For each node X i , divide the data according to the value of its parent node;

[0100] For each node X i For each parent node combination, calculate X i The conditional probability of each value:

[0101]

[0102] Among them, N(x,pa j ) is the parent node value pa j For and X i = the number of samples of x, N(pa j ) is the parent node with the value pa j The total number of samples;

[0103] After the Bayesian network is constructed, the present invention serializes the model and deploys it to the control module. The selected control module model can be an embedded device NVIDIA Jetson Nano with sufficient computing power. During deployment, the trained Bayesian network model is serialized into a file and saved to the device through the joblib tool. When the device is running, Jetson Nano connects to the sensor through the GPIO or I2C interface to collect key fire data in the energy storage power station, such as temperature, humidity, combustible gas concentration and voltage. The probability of fire occurrence and the stage of battery thermal runaway are judged through the reasoning results of the Bayesian network. When the probability exceeds the set threshold, the control module triggers a fire warning signal and automatically starts the corresponding fire extinguishing and waste liquid treatment system.

[0104] The fire identification probability formula is as follows:

[0105]

[0106] Among them, P(S,T,G,H|F) is the conditional probability of the currently detected smoke, temperature, combustible gas concentration and humidity when a fire occurs, and P(F|S,T,G,H) is the fire discrimination probability calculated by the Bayesian network based on the previously detected smoke, temperature, combustible gas concentration and humidity data.

[0107] P(F) is the prior probability of fire occurrence.

[0108] P(S,T,G,H) is the marginal probability of the sensor data.

[0109] By setting different fire probability thresholds, the various states of the battery are distinguished: when the fire probability is ≤0.1, the system determines that the battery is in normal working condition and does not trigger an early warning; when the fire probability is between 0.1 and 0.3, it enters the very early stage of thermal runaway and triggers a first-level warning to remind manual inspection; when the fire probability is between 0.3 and 1, the battery enters the middle and late stages of thermal runaway, the system triggers a second-level warning and starts the high-pressure water mist fire extinguishing system to control the fire with all efforts, and at the same time starts the waste liquid collection system to ensure safety.

[0110] The control module of the present invention realizes local reasoning of the fire warning model through the computing power of Jetson Nano, reduces the dependence on cloud computing, and improves the response speed. When the control module detects that the fire risk exceeds the set threshold, the system can automatically trigger the fire extinguishing system and the waste liquid treatment system, and upload the data to the remote monitoring center for subsequent accident analysis and system optimization.

[0111] The intelligent fire reasoning model of the present invention is based on the Bayesian network, which can calculate the probability of fire occurrence through multi-dimensional data (such as temperature, combustible gas concentration, humidity, etc.) collected by sensors in real time. This probabilistic reasoning method can not only provide early warning in the early stages of battery thermal runaway, but also dynamically adjust the fire warning threshold according to historical data, thereby improving the accuracy and timeliness of the warning.

[0112] The present invention is composed of a gas fire extinguishing system and a water-based fire extinguishing system, and is applicable to different fire stages of the lithium battery energy storage battery prefabricated cabin; in the fire warning stage, the system will remind the operator through the sound and light alarm equipment, and start the corresponding fire extinguishing device at the same time. The fire extinguishing system includes:

[0113] Gas fire extinguishing system: A prefabricated perfluorohexanone gas fire extinguishing device is used, which is suitable for the early stage of thermal runaway. When the Bayesian network determines that the lithium battery has entered the very early stage of thermal runaway, the system will start the gas fire extinguishing system and spray perfluorohexanone gas to locally suppress the fire source. The delay is 0-5s, the spray volume is 3L each time, the spray time is 20 seconds, the interval is 30 seconds, and it is repeated 15 times to ensure effective fire extinguishing.

[0114] Water-based fire extinguishing system: When the fire enters the late stage of thermal runaway or the smoke concentration and temperature sensors show that the fire has spread, the system will switch to the high-pressure water mist fire extinguishing system. The water-based fire extinguishing system controls the water flow through an electric ball valve, automatically opens and extinguishes the fire. The delay is 0-30s. According to the "Technical Specifications for Water Mist Fire Extinguishing Systems" GB50898-2013, the water pressure is not less than 10MPa. The high-pressure water mist system can quickly cool down and suppress the spread of fire, effectively controlling large-scale fires.

[0115] Traditional fire extinguishing systems can generally only use a single type of fire extinguishing medium, while the fire extinguishing system of the present invention can intelligently select fire extinguishing means according to the different stages of the fire. In the early stage of thermal runaway, a gas fire extinguishing system (such as perfluorohexanone) is used for local suppression, avoiding the problem of battery damage caused by water-based fire extinguishing; when the fire enters the middle and late stages, the system will automatically switch to a high-pressure fine water mist fire extinguishing system to achieve more efficient large-area fire control.

[0116] The waste liquid treatment system of the present invention includes waste liquid collection and real-time monitoring of waste liquid components; the waste liquid collection system includes a dedicated waste liquid collection pipeline and a waste liquid tank; when the water-based fire extinguishing system is started, the waste liquid collection system is activated at the same time, and the pipeline automatically guides the fire extinguishing waste liquid into the waste liquid tank. The waste liquid collection pipeline must have corrosion resistance and high temperature resistance. A stainless steel pipe is used, and an anti-corrosion paint is painted on the outer wall of the stainless steel pipe. Then bricks and concrete are used to build a groove and wrap the stainless steel pipe; an electric valve is installed on the waste liquid collection pipeline, and the pipeline valve automatically opens after receiving a start signal to transport the waste liquid generated during the fire extinguishing process to the waste liquid tank. The waste liquid tank is made of 316L stainless steel, and the inner wall of the waste liquid tank is treated with anti-seepage; the real-time monitoring waste liquid system, during the waste liquid collection process, the system monitors the waste liquid components (such as pH value, heavy metal ion concentration, etc.) in real time through sensors to ensure that the treated waste liquid meets environmental emission standards; if it is detected that the waste liquid treatment effect does not meet the standards, the system will trigger a secondary treatment process to ensure that the waste liquid can be discharged after safe treatment.

[0117] The waste liquid collection needs to consider the maximum estimated amount of waste liquid M in a fire accident to be at least 30%-50% more, M = N + L;

[0118] Where N is the estimated volume of the converted transfer chemicals, and L is the total water output within 1.0 hour of firefighting time.

[0119] The harmful waste liquid generated during the fire extinguishing process is often discharged directly, which may cause environmental pollution. The waste liquid treatment system of the present invention has corrosion resistance, and monitors and treats the composition of the waste liquid to ensure that the discharge meets environmental protection standards. This environmental protection design makes the system more suitable for application in rail transportation and other industrial fields with strict environmental requirements.

[0120] If the above functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program code.

[0121] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A fire handling system for lithium battery energy storage power stations used in rail transit. It is characterized in that The system includes: an energy storage battery, a sensor module, a fire extinguishing system, a waste liquid treatment system and a control module. The fire extinguishing system includes a gas fire extinguishing system and a water-based fire extinguishing system. The sensor module is connected to the control module, and the control module is connected to the gas fire extinguishing system, the water-based fire extinguishing system and the waste liquid treatment system. The control module is deployed with a Bayesian network, and the Bayesian network is used to predict the fire discrimination probability. The control module is used to judge the working state of the energy storage battery according to the fire discrimination probability, and control the corresponding fire extinguishing system and waste liquid treatment system to work according to the working state of the energy storage battery.

2. According to claim 1, a fire handling system for a lithium battery energy storage power station for rail transit, It is characterized in that The sensor module includes a voltage sensor, a smoke sensor, a temperature sensor, a combustible gas sensor, and a humidity sensor. The sensor module is used to collect fire data and transmit the fire data to the control module. The fire data is collected through various sensors of the sensor module, including the energy storage battery voltage, the indoor smoke concentration of the energy storage power station, the indoor temperature of the energy storage power station, the indoor combustible gas concentration of the energy storage power station, and the indoor humidity of the energy storage power station.

3. According to claim 1, a fire handling system for a lithium battery energy storage power station for rail transit, It is characterized in that The working state of the energy storage battery includes a normal working state, an early stage of thermal runaway and a late stage of thermal runaway.

4. According to claim 1, a fire handling system for a lithium battery energy storage power station for rail transit, It is characterized in that The functions of the control module include: determining whether the indoor smoke concentration and indoor temperature of the energy storage power station are greater than the first preset threshold and the second preset threshold respectively; if one of them is greater than the preset threshold, controlling the start of the water-based fire extinguishing system and the waste liquid treatment system; if neither is greater than the preset threshold, no treatment is performed.

5. According to claim 1, a fire handling system for a lithium battery energy storage power station for rail transit, It is characterized in that The gas fire extinguishing system is a prefabricated perfluorohexanone fire extinguishing system, which is connected and controlled by a control module and is used to start fire extinguishing when the energy storage battery is in the early stage of thermal runaway; the water-based fire extinguishing system is a high-pressure fine water mist fire extinguishing system, which is connected and controlled by a control module and is used for fire extinguishing. The water-based fire extinguishing system will produce fire extinguishing waste liquid when extinguishing a fire.

6. According to claim 1, a fire handling system for a lithium battery energy storage power station for rail transit, It is characterized in that The waste liquid treatment system includes a waste liquid collection module and a waste liquid component real-time monitoring module. The waste liquid treatment system and the water-based fire extinguishing system are started simultaneously. The waste liquid collection module includes a waste liquid collection pipe and a waste liquid tank. The waste liquid collection pipe is made of stainless steel pipe, and the outer wall of the stainless steel pipe is painted with anti-corrosion paint. The waste liquid collection pipe is installed with an electric valve, and the electric valve is connected and controlled by a control module. The waste liquid collection pipe is used to guide the fire extinguishing waste liquid into the waste liquid tank. The waste liquid tank is made of stainless steel, and the inner wall of the waste liquid tank is treated with anti-seepage. The waste liquid component real-time monitoring module includes a PH sensor and a heavy metal ion detection sensor, which are used to monitor the pH value and heavy metal ion concentration of the waste liquid in real time. The PH sensor and the heavy metal ion detection sensor are installed on the inner wall of the waste liquid tank.

7. A fire treatment method for a lithium battery energy storage power station in rail transit, It is characterized in that The following steps are involved: Step S1: construct a Bayesian network for predicting the probability of fire identification through historical data, and deploy the trained Bayesian network in the control module; Step S2: Collect fire data through the sensor module, including: energy storage battery voltage, indoor smoke concentration of the energy storage power station, indoor temperature of the energy storage power station, indoor combustible gas concentration of the energy storage power station and indoor humidity of the energy storage power station, and transmit the fire data to the control module; Step S3: the control module inputs the fire data into the Bayesian network, the Bayesian network outputs the fire discrimination probability, and the control module determines the working state of the energy storage battery according to the fire discrimination probability; Step S4: the control module selects a corresponding fire extinguishing system according to the working state of the energy storage battery, and sends a control signal to the corresponding fire extinguishing system and waste liquid treatment system; the control module determines whether the indoor smoke concentration and indoor temperature of the energy storage power station are greater than the first preset threshold and the second preset threshold respectively, and if one is greater than the preset threshold, the control module sends a control signal to the water-based fire extinguishing system and the waste liquid treatment system, and if neither is greater than the preset threshold, no processing is performed; Step S5: After receiving the control signal, the fire extinguishing system starts to start and extinguish the fire, and after receiving the control signal, the waste liquid treatment system starts to collect waste liquid.

8. A fire treatment method for a lithium battery energy storage power station for rail transit according to claim 7, It is characterized in that In step S1, a Bayesian network for predicting the probability of fire identification is constructed using historical data, including the following steps: Obtain historical voltage, temperature, smoke concentration, combustible gas concentration and indoor humidity data of the energy storage power station, including data under normal operation and abnormal conditions; Preprocessing the acquired data, including removing noise and outliers, formatting the data into an input form suitable for a machine learning algorithm, and labeling the data; The noise removal comprises the following steps: The temperature data and voltage data are subjected to noise reduction by using the moving average method; Where T(t) is the temperature or voltage value at time t, and n is the window length, which can be adjusted according to the system's requirements for fluctuation sensitivity; Exponential smoothing method is used to denoise the humidity, smoke concentration and combustible gas concentration data. The formula is as follows: Where H(t) is the humidity value or combustible gas concentration value at time t, and α is the smoothing coefficient; The input format suitable for machine learning algorithms includes: The temperature and voltage data are Z- score Normalized, the formula is: where x i is the temperature value at time i, μ is the data mean, and σ is the data standard deviation; The collected gas concentration data is logarithmically normalized, and the formula is as follows: x1=log(x+1) Where x is the gas concentration; The collected humidity data is Mi n -Ma x Standardization processing, the formula is as follows: Where x is the current humidity value, x min =0,x max =1; The parameters of the Bayesian network are learned by using the maximum likelihood estimation method through the preprocessed data to obtain the trained Bayesian network.

9. A fire treatment method for a lithium battery energy storage power station for rail transit according to claim 7, It is characterized in that In step S3, the control module inputs the fire data into the Bayesian network, and the Bayesian network outputs the fire discrimination probability. The fire discrimination probability formula is: Among them, P(S,T,G,H|F) is the conditional probability of the currently detected smoke, temperature, combustible gas concentration and humidity when a fire occurs, P(F) is the prior probability of fire occurrence, P(S,T,G,H) is the marginal probability of sensor data, and P(F|S,T,G,H) is the fire discrimination probability output by the Bayesian network.

10. A fire treatment method for a lithium battery energy storage power station for rail transit according to claim 7, It is characterized in that The control module determines the working state of the energy storage battery according to the fire identification probability, including: When the fire judgment probability is ≤0.1, the control module determines that the energy storage battery is in normal working condition and does not trigger the warning; When the fire identification probability is between 0.1 and 0.3, the control module determines that the energy storage battery is in the early stage of thermal runaway, and the control module sends a control signal to the gas fire extinguishing system; When the fire identification probability is between 0.3 and 1, the control module determines that the energy storage battery is in the late stage of thermal runaway, and the control module sends a control signal to the water-based fire extinguishing system and the waste liquid treatment system.

Images