Fire-fighting joint control system for multiple energy storage devices

Abstract

The invention discloses a fire-fighting joint control system for multiple energy storage devices, and belongs to the technical field of fire-fighting control of the energy storage devices, an information acquisition module comprises a basic sensing layer for judging the situation after a fire behavior occurs and a prediction sensing layer data analysis decision module for predicting before the fire behavior occurs, comprising a fire situation assessment unit used for assessing the fire development situation and the potential spreading path in real time and a thermal runaway risk assessment unit used for performing risk index calculation after an improved entropy weight-TOPSIS model is adopted for assessment. The cooperative control module is used for optimizing fire extinguishing resource scheduling based on a dynamic game theory and generating a multi-device linkage strategy through a priority weight algorithm; and the fire-fighting execution module integrates a multi-channel spraying system, an inert gas injection module and a physical isolation mechanism and executes graded response fire fighting. Intelligent linkage control over multiple energy storage devices can be achieved, the thermal runaway risk is accurately predicted, and fire extinguishing resource allocation is optimized.

Description

Technical Field

[0001] This invention belongs to the field of fire protection control technology for energy storage equipment, and specifically relates to a fire protection joint control system for multiple energy storage devices. Background Technology

[0002] With the large-scale application of renewable energy, electrochemical energy storage power stations have developed rapidly as important energy regulation facilities. However, the risk of thermal runaway of energy storage batteries (especially lithium-ion batteries) has always been a major hidden danger restricting their safe application. Most existing fire protection systems for energy storage power stations adopt a passive response design, that is, the fire extinguishing procedure is only activated after an open flame or high temperature is detected. This delayed fire protection measure is often unable to effectively control the chain reaction of battery thermal runaway.

[0003] Especially in large-scale energy storage power plants, when multiple energy storage devices experience thermal runaway simultaneously or successively, existing fire protection systems lack effective global coordination and control capabilities, which may lead to problems such as uneven distribution of fire extinguishing resources and untimely response. In addition, different types of energy storage devices (such as lithium-ion batteries, sodium-sulfur batteries, and flow batteries) have different requirements for fire extinguishing agents, while existing systems often use a single fire extinguishing agent, making it difficult to achieve optimal fire extinguishing effects. Summary of the Invention

[0004] The present invention aims to at least partially solve one of the technical problems in the related art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: This invention provides a fire-fighting joint control system for multiple energy storage devices, including an information acquisition module comprising a basic sensing layer and a predictive sensing layer. The basic sensing layer is used to acquire temperature signals, smoke signals, VOC gas signals, and pressure signals in real time to determine the situation after a fire occurs. The predictive sensing layer is used to acquire sound wave signals, gas composition, thermodynamic parameters, and image features in real time and is used to predict fires before they occur. The data analysis and decision-making module includes a fire situation assessment unit and a thermal runaway risk assessment unit. The fire situation assessment unit is used to assess the real-time development of the fire and potential spread paths; the thermal runaway risk assessment unit, based on a dynamic weighted fusion model, integrates features... Mapped to the key parameter space of thermal runaway, the risk index is calculated after evaluation using the improved entropy weight-TOPSIS model; The collaborative control module optimizes the scheduling of firefighting resources based on dynamic game theory and generates multi-device linkage strategies through a priority weight algorithm. The fire-fighting execution module integrates a multi-channel sprinkler system, an inert gas injection module, and a physical isolation mechanism to perform graded response fire-fighting.

[0006] Furthermore, the fire-fighting methods of the fire control system include the following steps: S1: Create a virtual game agent for each energy storage device and initialize the strategy space. ,in, To maximize available firefighting resources, calculations are made based on the firefighting execution compartment volume of the firefighting execution module; the registered firefighting execution module includes a sprinkler unit mounted on top of the energy storage device. and the inert gas injection unit surrounding the energy storage device. and fireproof partition units installed between two adjacent energy storage devices. And set the upper limit of the fire protection execution module capacity for each unit. ; S2: Define the revenue function, including the revenue function for energy storage devices. and the revenue function of the fire protection execution module :

[0007]

[0008] In the formula, As a risk index, The severity of the fire on equipment i. The amount of fire extinguishing resources requested by device i. Let be the critical threshold for thermal runaway of device i. These are the initial values ​​for the weights. This is the resource consumption penalty coefficient. For the execution cost coefficient, This represents the total number of associated devices. The transmission loss coefficient for allocating fire extinguishing resources from execution unit j to device i. Let be the fire extinguishing efficiency coefficient of execution unit j for device i.

[0009] S3: Real-time game strategy process: When the prediction and perception layer detects the risk of thermal runaway, the risk index R is evaluated by the thermal runaway risk assessment unit, and the payoff function is adjusted according to the risk index R to output the optimal combination strategy of each unit in the fire protection execution module. ; Implement precise resource allocation based on the game outcome A closed-loop detection is performed once within a preset time period to provide feedback and adjust strategies; when the basic sensing layer detects a fire, the fire-fighting execution module outputs that all units adopt the combined strategies. Firefighting effectiveness is assessed through a fire situation assessment unit.

[0010] Furthermore, in step S3, the strategy optimization method is as follows:

[0011] In the formula, For learning rate, For instant rewards, As a discount factor, Let be the action value function for taking action a in state s. For the next state All possible actions The largest in value.

[0012] Furthermore, the assessment method of the fire situation assessment unit includes the following steps: A1: Temperature Smoke concentration VOC concentration Pressure changes Sensor data is normalized and weighted to generate a comprehensive fire index F:

[0013] In the formula, These are the weighting coefficients for the temperature parameter. The weighting coefficient for smoke concentration is... The weighting coefficient for VOC concentration. The weighting coefficient for pressure. The initial ambient temperature, Atmospheric pressure , , , These are the upper limits of the sensor ranges for temperature sensors, smoke sensors, VOC gas sensors, and pressure sensors, respectively. A2: Confirm the occurrence of a fire and classify its intensity using threshold judgment and Bayesian probability model:

[0014] In the formula, Let be the posterior probability of a fire occurring. This represents the conditional probability that the data from each sensor satisfy the comprehensive fire index F when a fire occurs. Let be the prior probability of a fire occurring. The marginal probability that the data from each sensor satisfy the comprehensive fire index F when a fire occurs; A3: Quantifying fire energy based on heat release rate model:

[0015] In the formula, The rate of heat release from the fire. This is a comprehensive coefficient. The temperature of the fire source; A4: Calculation of heat flux based on Stefan-Boltzmann's law:

[0016] In the formula, For heat flux density, This is the Stefan-Boltzmann constant. For the emissivity of battery materials, The absolute temperature of the surface of the fire source. The absolute temperature of the environment. The convective heat transfer coefficient; A5: Predicting gas diffusion using a Gaussian plume model:

[0017] In the formula, The concentration of heat released by the fire in three-dimensional space. For three-dimensional coordinate axes, The rate of heat release from the fire. The average wind speed, This represents the length of the heat release from the fire diffused along the y-axis. This represents the diffusion length of heat release materials from the fire along the z-axis. The height of the exhaust duct for the energy storage equipment.

[0018] Furthermore, the assessment method for the thermal runaway risk assessment unit includes the following steps: B1: Constructing a fusion model to perform real-time feature extraction and fusion computation:

[0019] In the formula, As a feature of fusion, For the sound wave time-frequency matrix, This is the gas concentration gradient matrix. For thermodynamic temperature field, The image pixel matrix It is a nonlinear mapping function. For modal identifiers, These are the fusion weighting coefficients for each modality. for Norm normalization, For modality-specific trainable parameter matrices, To modify the activation function of the linear unit, Extract convolution kernels for modal features. For raw multimodal data, Modal bias parameters; This is a convolution operation used for cross-modal feature extraction; B2: Mapping the fused feature F to the thermal runaway key parameter space:

[0020] In the formula, For the rate of temperature change, The gradient represents the change in hydrogen concentration. This is the second derivative of thermal imaging; B3: Risk Index calculate

[0021] B4: Dynamic Threshold Adjustment

[0022] B5: Risk Level Mapping and Decision-Making

[0023] In the formula, This is the proportionality coefficient. This is the weight matrix. This is a bias term.

[0024] Furthermore, the basic sensing layer includes distributed temperature sensors, smoke detectors, VOC gas sensors, and pressure sensors, which are redundantly deployed to cover key monitoring points of the energy storage device.

[0025] Furthermore, the acoustic signal acquisition of the predictive sensing layer employs an ultrasonic detection device to identify abnormal acoustic features such as lithium plating or separator rupture inside the battery.

[0026] Furthermore, the multi-channel sprinkler system of the fire-fighting execution module includes the following spray modes: high-pressure water mist mode for rapid cooling; fine water mist mode for suppressing reignition; and directional spray mode for precisely covering the core area of ​​the fire source.

[0027] Furthermore, the physical isolation mechanism includes a fireproof partition, an explosion-proof door, and a fuse protection device, wherein the fuse protection device actively cuts off the electrical connection when it detects an abnormal voltage between battery packs.

[0028] The beneficial effects of this invention are as follows: This invention utilizes Nash equilibrium solution based on dynamic game theory to generate the optimal resource allocation strategy within 100ms. Based on the entropy-weighted TOPSIS model, strategy space optimization ensures a non-linear match between sprinkler intensity and fire severity, achieving a sprinkler coverage rate of 98% in the core fire zone, a 35% improvement over the fixed flow mode. Furthermore, by establishing a penalty mechanism in the payoff function, excessive fire suppression is effectively suppressed, and by dynamically adjusting the payoff function weights, high-risk equipment receives greater resource protection.

[0029] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0030] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a structural schematic diagram of a fire-fighting joint control system for multiple energy storage devices provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the execution flow of a fire-fighting joint control system for multiple energy storage devices, provided in an embodiment of the present invention. Detailed Implementation

[0031] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0032] like Figure 1 As shown, the present invention provides a fire control system for multiple energy storage devices, comprising: The information acquisition module 100 includes a basic sensing layer and a predictive sensing layer. The basic sensing layer includes distributed temperature sensors, smoke sensors, VOC gas sensors, and pressure sensors, and is used to determine whether a fire has occurred. The predictive sensing layer includes a multi-modal sensor array set at key nodes of each energy storage device, which is used to collect sound wave signals, gas composition, thermodynamic parameters, and image features in real time. The predictive sensing layer is used to predict whether a fire is about to occur. The data analysis and decision-making module 200 includes a fire situation assessment unit and a thermal runaway risk assessment unit. The fire situation assessment unit is used to assess the real-time development of the fire and potential spread paths; the thermal runaway risk assessment unit, based on a dynamic weighted fusion model, integrates features... Mapped to the key parameter space of thermal runaway, the risk index is calculated after evaluation using the improved entropy weight-TOPSIS model; The collaborative control module 300 optimizes the scheduling of firefighting resources based on dynamic game theory and generates multi-device linkage strategies through a priority weight algorithm. The fire-fighting execution module 400 integrates a multi-channel sprinkler system, an inert gas injection module, and a physical isolation mechanism to perform graded response fire-fighting.

[0033] Further, refer to Figure 2The fire-fighting methods of the fire control system include the following steps: S1: Create a virtual game agent for each energy storage device and initialize the strategy space. ,in, To maximize available firefighting resources, calculations are made based on the firefighting execution compartment volume of the firefighting execution module; the registered firefighting execution module includes a sprinkler unit mounted on top of the energy storage device. and the inert gas injection unit surrounding the energy storage device. and fireproof partition units installed between two adjacent energy storage devices. And set the upper limit of the fire protection execution module capacity for each unit. ; Among them, energy storage device strategy set :

[0034] Firefighting Execution Module Strategy Set :

[0035] In the formula, To maximize available firefighting resources; This represents the upper limit of the fire protection execution module capacity. S2: Define the revenue function, including the revenue function for energy storage devices. and the revenue function of the fire protection execution module :

[0036]

[0037] In the formula, As a risk index, The severity of the fire on equipment i. The amount of fire extinguishing resources requested by device i. Let be the critical threshold for thermal runaway of device i. These are the initial values ​​for the weights. This is the resource consumption penalty coefficient. For the execution cost coefficient, This represents the total number of associated devices. The transmission loss coefficient for allocating fire extinguishing resources from execution unit j to device i. Let be the fire extinguishing efficiency coefficient of execution unit j for device i.

[0038] S3: Real-time game strategy process: When the prediction and perception layer detects the risk of thermal runaway, the risk index R is evaluated by the thermal runaway risk assessment unit, and the payoff function is adjusted according to the risk index R to output the optimal combination strategy of each unit in the fire protection execution module. ; Implement precise resource allocation based on the game outcome A closed-loop detection is performed once within a preset time period to provide feedback and adjust strategies; when the basic sensing layer detects a fire, the fire-fighting execution module outputs that all units adopt the combined strategies. Firefighting effectiveness is assessed through a fire situation assessment unit.

[0039] The fire situation assessment unit is used to assess the development of the fire and its potential spread in real time. Specific steps include: A1: Temperature Smoke concentration VOC concentration Pressure changes Sensor data is normalized and weighted to generate a comprehensive fire index F:

[0040] In the formula, These are the weighting coefficients for the temperature parameter. The weighting coefficient for smoke concentration is... The weighting coefficient for VOC concentration. The weighting coefficient for pressure. The initial ambient temperature, Atmospheric pressure , , , These are the upper limits of the sensor ranges for temperature sensors, smoke sensors, VOC gas sensors, and pressure sensors, respectively. A2: Confirm the occurrence of a fire and classify its intensity using threshold judgment and Bayesian probability model:

[0041] In the formula, Let be the posterior probability of a fire occurring. This represents the conditional probability that the data from each sensor satisfy the comprehensive fire index F when a fire occurs. Let be the prior probability of a fire occurring. The marginal probability that the data from each sensor satisfy the comprehensive fire index F when a fire occurs; A3: Quantifying fire energy based on heat release rate model:

[0042] In the formula, The rate of heat release from the fire. This is a comprehensive coefficient. The temperature of the fire source; A4: Calculation of heat flux based on Stefan-Boltzmann's law:

[0043] In the formula, For heat flux density, This is the Stefan-Boltzmann constant. For the emissivity of battery materials, The absolute temperature of the surface of the fire source. The absolute temperature of the environment. The convective heat transfer coefficient; A5: Predicting gas diffusion using a Gaussian plume model:

[0044] In the formula, The concentration of heat released by the fire in three-dimensional space. For three-dimensional coordinate axes, The rate of heat release from the fire. The average wind speed, This represents the length of the heat release from the fire diffused along the y-axis. This represents the diffusion length of heat release materials from the fire along the z-axis. The height of the exhaust duct for the energy storage equipment.

[0045] The thermal runaway risk assessment unit, based on a dynamic weighted fusion model, integrates features... Mapped to the key parameter space of thermal runaway, the risk index is calculated after evaluation using the improved entropy weight-TOPSIS model; B1: Constructing a fusion model to perform real-time feature extraction and fusion computation:

[0046] In the formula, As a feature of fusion, For the sound wave time-frequency matrix, This is the gas concentration gradient matrix. For thermodynamic temperature field, The image pixel matrix It is a nonlinear mapping function. For modal identifiers, These are the fusion weighting coefficients for each modality. for Norm normalization, For modality-specific trainable parameter matrices, To modify the activation function of the linear unit, Extract convolution kernels for modal features. For raw multimodal data, Modal bias parameters; This is a convolution operation used for cross-modal feature extraction; B2: Mapping the fused feature F to the thermal runaway key parameter space:

[0047] In the formula, For the rate of temperature change, The gradient represents the change in hydrogen concentration. This is the second derivative of thermal imaging; B3: Risk Index calculate

[0048] B4: Dynamic Threshold Adjustment

[0049] B5: Risk Level Mapping and Decision-Making

[0050] In the formula, This is the proportionality coefficient. This is the weight matrix. This is a bias term.

[0051] This solution uses Nash equilibrium from dynamic game theory to generate the optimal resource allocation strategy within 100ms. Based on the entropy weight-TOPSIS model, the strategy space optimization makes the sprinkler intensity non-linearly matched with the fire severity, achieving a sprinkler coverage rate of 98% in the core fire zone, which is 35% higher than the fixed flow mode. By establishing a penalty mechanism in the payoff function, excessive fire suppression is effectively suppressed, and by dynamically adjusting the weight of the payoff function, high-risk equipment receives more resource protection.

[0052] Finally, it should be noted that the above preferred embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail through the above preferred embodiments, those skilled in the art should understand that various changes can be made to it in form and detail without departing from the scope defined by the claims of the present invention.

Claims

1. A fire-fighting joint control system for multiple energy storage devices, characterized in that, include: The information acquisition module includes a basic sensing layer and a predictive sensing layer. The basic sensing layer is used to acquire temperature signals, smoke signals, VOC gas signals, and pressure signals in real time to determine the situation after a fire occurs. The predictive sensing layer is used to acquire sound wave signals, gas composition, thermodynamic parameters, and image features in real time and is used to predict fires before they occur. The data analysis and decision-making module includes a fire situation assessment unit and a thermal runaway risk assessment unit. The fire situation assessment unit is used to assess the real-time development of the fire and potential spread paths; the thermal runaway risk assessment unit, based on a dynamic weighted fusion model, integrates features... Mapped to the key parameter space of thermal runaway, the risk index is calculated after evaluation using the improved entropy weight-TOPSIS model; The collaborative control module optimizes the scheduling of firefighting resources based on dynamic game theory and generates multi-device linkage strategies through a priority weight algorithm. The fire-fighting execution module integrates a multi-channel sprinkler system, an inert gas injection module, and a physical isolation mechanism to perform graded response fire-fighting.

2. The fire control system for multiple energy storage devices according to claim 1, characterized in that: The fire-fighting methods of the fire control system include the following steps: S1: Create a virtual game agent for each energy storage device and initialize the strategy space. ,in, To maximize available firefighting resources, calculations are made based on the firefighting execution compartment volume of the firefighting execution module; the registered firefighting execution module includes a sprinkler unit mounted on top of the energy storage device. and the inert gas injection unit surrounding the energy storage device. and fireproof partition units installed between two adjacent energy storage devices. And set the upper limit of the fire protection execution module capacity for each unit. ; S2: Define the revenue function, including the revenue function for energy storage devices. and the revenue function of the fire protection execution module : In the formula, As a risk index, The severity of the fire on equipment i. The amount of fire extinguishing resources requested by device i. Let be the critical threshold for thermal runaway of device i. These are the initial values ​​for the weights. This is the resource consumption penalty coefficient. For the execution cost coefficient, This represents the total number of associated devices. The transmission loss coefficient for allocating fire extinguishing resources from execution unit j to device i. Let be the fire extinguishing efficiency coefficient of execution unit j for device i. S3: Real-time game strategy process: When the prediction and perception layer detects the risk of thermal runaway, the risk index R is evaluated by the thermal runaway risk assessment unit, and the payoff function is adjusted according to the risk index R to output the optimal combination strategy of each unit in the fire protection execution module. ; Implement precise resource allocation based on the game outcome A closed-loop detection is performed once within a preset time period to provide feedback and adjust strategies; when the basic sensing layer detects a fire, the fire-fighting execution module outputs that all units adopt the combined strategies. Firefighting effectiveness is assessed through a fire situation assessment unit.

3. The fire control system for multiple energy storage devices according to claim 2, characterized in that: In step S3, the strategy optimization method is as follows: In the formula, For learning rate, For instant rewards, As a discount factor, Let be the action value function for taking action a in state s. For the next state All possible actions The largest in value.

4. The fire control system for multiple energy storage devices according to claim 3, characterized in that: The assessment method of the fire situation assessment unit includes the following steps: A1: Temperature Smoke concentration VOC concentration Pressure changes Sensor data is normalized and weighted to generate a comprehensive fire index F: In the formula, These are the weighting coefficients for the temperature parameter. The weighting coefficient for smoke concentration is... The weighting coefficient for VOC concentration. The weighting coefficient for pressure. The initial ambient temperature, Atmospheric pressure , , , These are the upper limits of the sensor ranges for temperature sensors, smoke sensors, VOC gas sensors, and pressure sensors, respectively. A2: Confirm the occurrence of a fire and classify its intensity using threshold judgment and Bayesian probability model: In the formula, Let be the posterior probability of a fire occurring. This represents the conditional probability that the data from each sensor satisfy the comprehensive fire index F when a fire occurs. Let be the prior probability of a fire occurring. The marginal probability that the data from each sensor satisfy the comprehensive fire index F when a fire occurs; A3: Quantifying fire energy based on heat release rate model: In the formula, The rate of heat release from the fire. This is a comprehensive coefficient. The temperature of the fire source; A4: Calculation of heat flux based on Stefan-Boltzmann's law: In the formula, For heat flux density, This is the Stefan-Boltzmann constant. For the emissivity of battery materials, The absolute temperature of the surface of the fire source. The absolute temperature of the environment. The convective heat transfer coefficient; A5: Predicting gas diffusion using a Gaussian plume model: In the formula, The concentration of heat released by the fire in three-dimensional space. For three-dimensional coordinate axes, The rate of heat release from the fire. The average wind speed, This represents the length of the heat release from the fire diffused along the y-axis. This represents the diffusion length of heat release materials from the fire along the z-axis. The height of the exhaust duct for the energy storage equipment.

5. The fire control system for multiple energy storage devices according to claim 4, characterized in that: The assessment method for the thermal runaway risk assessment unit includes the following steps: B1: Constructing a fusion model to perform real-time feature extraction and fusion computation: In the formula, As a feature of fusion, For the sound wave time-frequency matrix, This is the gas concentration gradient matrix. For thermodynamic temperature field, The image pixel matrix It is a nonlinear mapping function. For modal identifiers, These are the fusion weighting coefficients for each modality. for Norm normalization, For modality-specific trainable parameter matrices, To modify the activation function of the linear unit, Extract convolution kernels for modal features. For raw multimodal data, Modal bias parameters; This is a convolution operation used for cross-modal feature extraction; B2: Mapping the fused feature F to the thermal runaway key parameter space: In the formula, For the rate of temperature change, The gradient represents the change in hydrogen concentration. This is the second derivative of thermal imaging; B3: Risk Index calculate B4: Dynamic Threshold Adjustment B5: Risk Level Mapping and Decision-Making In the formula, This is the proportionality coefficient. This is the weight matrix. This is a bias term.

6. The fire control system for multiple energy storage devices according to claim 1, characterized in that: The basic sensing layer includes distributed temperature sensors, smoke detectors, VOC gas sensors, and pressure sensors, which are redundantly deployed to cover key monitoring points of the energy storage device.

7. The fire control system for multiple energy storage devices according to claim 1, characterized in that: The acoustic signal acquisition of the predictive sensing layer uses an ultrasonic detection device to identify abnormal acoustic features such as lithium plating or separator rupture inside the battery.

8. The fire control system for multiple energy storage devices according to claim 1, characterized in that: The multi-channel sprinkler system of the fire-fighting execution module includes the following spray modes: high-pressure water mist mode for rapid cooling; fine water mist mode for suppressing reignition; and directional spray mode for precisely covering the core area of ​​the fire source.

9. The fire control system for multiple energy storage devices according to claim 1, characterized in that: The physical isolation mechanism includes a fireproof partition, an explosion-proof door, and a fuse protection device, wherein the fuse protection device actively cuts off the electrical connection when it detects an abnormal voltage between battery packs.

Images