New energy automobile fire monitoring and early warning emergency method and device based on Internet of Things

Abstract

The invention discloses a new energy automobile fire monitoring, early warning and emergency method and device based on the Internet of Things, and relates to the technical field of fire early warning and emergency. In the data acquisition stage, the charging state and environmental data of a charging area are monitored in real time through a multi-mode sensor, and data fusion processing is achieved through an Internet of Things platform; then, in a risk assessment stage, a deep learning model is introduced to dynamically generate a risk score in combination with historical data and real-time data, grading early warning conditions are further defined, and when data such as the temperature of the charging pile and the gas concentration are abnormal but do not reach a fire hazard standard, secondary early warning is triggered. Suspending power supply and ventilation equipment, and recording detailed abnormal data; when the smoke sensor and the temperature sensor jointly confirm the fire risk, primary early warning is triggered, linkage response is started, and the accuracy and response speed of early warning are greatly improved through intelligent evaluation and a linkage mechanism.

Description

Technical Field

[0001] The present invention relates to the field of fire early warning emergency technology, and in particular to a new energy vehicle fire monitoring early warning emergency method and device based on the Internet of Things. Background Art

[0002] The rapid popularization of new energy vehicles has profoundly changed people's travel methods, but the resulting safety hazards, especially fire risks, are gradually becoming challenges that cannot be ignored. In high-density usage scenarios such as charging stations and parking lots, traditional fire monitoring and early warning methods mainly rely on fixed sensor networks, which detect temperature, smoke or specific gases at a single point and combine video surveillance for manual confirmation to provide certain guarantees for the identification of fire risks and emergency response.

[0003] However, this traditional method has many limitations. First, the monitoring range of a single sensor is limited. It is difficult to achieve full coverage when multiple vehicles are charging at the same time or in large areas, and it is easy to form monitoring blind spots. Second, static sensors are difficult to perceive dynamic risks in real time. When battery thermal runaway causes rapid temperature rise or there is a trend of fire spread in adjacent vehicles, problems cannot be discovered in time, causing the fire to spread rapidly. In addition, the linkage and emergency response of traditional solutions are relatively delayed, which will further increase the risk of fire spread.

[0004] To address the above problems, some traditional solutions introduce multi-sensor systems, increase monitoring points and diversify data collection methods to supplement the shortcomings of a single sensor. However, such supplementary measures will also bring higher hardware costs and operation and maintenance complexity. At the same time, due to the lack of data fusion and comprehensive analysis capabilities, it is difficult to significantly improve the intelligence level of the system as a whole. Therefore, how to achieve dynamic risk perception and rapid response has become a key problem facing existing fire monitoring and emergency systems. Summary of the invention

[0005] In view of the above existing problems, the present invention is proposed.

[0006] The present invention provides a new energy vehicle fire monitoring and early warning emergency method and device based on the Internet of Things to solve the problems that the monitoring range of a single sensor is limited, it is difficult to achieve full coverage when multiple vehicles are charging at the same time or in a large area, and blind spots are easily formed; static sensors are difficult to perceive dynamic risks in real time, and when thermal runaway of the battery causes rapid temperature rise or there is a trend of fire spread in adjacent vehicles, the problem cannot be discovered in time, resulting in the fire possibly expanding rapidly; linkage and emergency response lags, which can easily lead to the rapid spread of the fire.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In a first aspect, an embodiment of the present invention provides a new energy vehicle fire monitoring and early warning emergency method based on the Internet of Things, which includes:

[0009] Step S1, collecting charging status and environmental data of the charging area in real time through a multimodal sensor, and transmitting the data to the Internet of Things platform for data fusion, the multimodal sensor includes a smoke sensor and a temperature and humidity sensor;

[0010] Step S2: The IoT platform uses the deep learning model and historical charging status and environmental data to conduct risk assessment based on the fused data. The assessment results are pushed to the management personnel in real time and synchronized to the cloud platform.

[0011] Risk assessment methods include:

[0012] If the temperature, gas concentration or charging status of the charging equipment is abnormal but does not meet the fire standard, a secondary warning will be triggered, the charging pile power supply and ventilation equipment will be suspended, and the abnormal temperature value, gas concentration and power fluctuation data will be recorded;

[0013] If the smoke sensor and the temperature sensor both confirm the fire risk, the first-level warning will be triggered immediately and the emergency linkage state will be entered;

[0014] Step S3, based on the early warning signal, the IoT platform coordinates the fire extinguishing system to carry out emergency treatment, and activates water spray, chemical fire extinguishing agent and emergency isolation chamber;

[0015] The emergency isolation chamber uses physical isolation devices to seal off the fire risk area, while cutting off the power supply to the charging piles and shutting down the ventilation equipment;

[0016] Step S4: After the emergency response is completed, all monitoring data are uploaded to the cloud platform for adjusting the early warning strategy.

[0017] As a preferred solution of the new energy vehicle fire monitoring and early warning emergency method based on the Internet of Things described in the present invention, wherein: the environmental data includes environmental temperature and humidity, gas concentration and smoke concentration, and the gas concentration includes CO2 and H2 concentration;

[0018] The charging status includes voltage, current, charging power and charging device temperature data.

[0019] As a preferred solution of the new energy vehicle fire monitoring and early warning emergency method based on the Internet of Things described in the present invention, wherein: the linkage state includes starting the warning light, the alarm device and pre-activating the fire extinguishing system on standby;

[0020] The emergency isolation chamber begins preparing for follow-up actions after receiving the early warning signal during the data collection phase.

[0021] As a preferred solution of the new energy vehicle fire monitoring and early warning emergency method based on the Internet of Things described in the present invention, the step of using the deep learning model and historical charging status and environmental data to perform risk assessment is as follows:

[0022] Perform data fusion and feature extraction, and the fused data is represented as D t :

[0023] D t ={T t ,G t ,C t},

[0024] Among them, T t ={t1,t2,…,t n}, t i represents the data collected by the i-th temperature sensor at time t, n is the number of temperature sensors, G t ={g1,g2,…,g m}, g j represents the data collected by the jth gas sensor at time t, m is the number of gas sensors, C t = {v t ,i t ,p t}, respectively representing the voltage v at time t t 、Current i t , power p t ,

[0025] The risk scoring function is used for risk assessment. The risk scoring function is based on a multi-layer neural network, and the specific structure is:

[0026] R t =f(D t ,H t )=σ(W (3) ·σ(W (2) ·σ(W (1) ·X+b (1) )+b (2) )+b (3) ),

[0027] Among them, R t represents the risk score at time t, H t represents the historical data set, and X represents the fused data D t and historical data H t splicing, σ(x) is the Sigmoid activation function, W (1) ,W (2) ,W (3) are the weight matrices of the first, second and third layers respectively, b (1) ,b(2) ,b (3) are the bias vectors of the corresponding layers respectively. The calculation goal of the risk score is to output R through the model. t Estimate the probability of fire occurrence, the value range is [0,1],

[0028] Perform model training and define the loss function of model training as mean square error MSE. The formula is:

[0029]

[0030] Among them, L represents the loss value of the model, k represents the number of training samples, represents the model prediction risk of the jth sample, R j Represents the true risk value of the jth sample.

[0031] As a preferred solution of the new energy vehicle fire monitoring and early warning emergency method based on the Internet of Things described in the present invention, the step of using the deep learning model and historical charging status and environmental data for risk assessment also includes:

[0032] Add alert conditions, including:

[0033] Secondary warning conditions: When one of the following conditions is met, the secondary warning is triggered:

[0034] max(T t ) <T th ,

[0035] max(G t ) <G th ,

[0036] ΔG t >C diff ,

[0037] Among them, max(T t ) represents the maximum temperature at time t, T th Indicates the second-level warning temperature threshold, max(G t ) represents the maximum gas concentration at time t, G th Indicates the secondary warning gas concentration threshold, ΔG t represents the power fluctuation at time t, C diff Indicates the power fluctuation threshold,

[0038] Level 1 warning conditions: When the following conditions are met at the same time, a level 1 warning is triggered:

[0039] s sm =1 and s tm =1,

[0040] Among them, s smIndicates smoke confirmation signal, value is 0 or 1, s tm Indicates the temperature confirmation signal, the value is 0 or 1;

[0041] After the warning is triggered, the abnormal data is recorded, including:

[0042] D abn ={max(T t ),max(G t ),ΔC t},

[0043] Among them, D abn represents an abnormal data set, max(T t )、max(G t ) and ΔC t They are abnormal maximum temperature, gas concentration and power fluctuations respectively.

[0044] As a preferred solution of the new energy vehicle fire monitoring and early warning emergency method based on the Internet of Things described in the present invention, the steps of the Internet of Things platform coordinating the fire extinguishing system to perform emergency treatment based on the early warning signal are as follows:

[0045] The IoT platform generates an emergency signal set U:

[0046] U={u1,u2,u3,u4,u5},

[0047] Among them, u1 represents the water sprinkler equipment start signal, the value is 0 or 1, u2 represents the chemical fire extinguishing agent equipment start signal, the value is 0 or 1, u3 represents the isolation device start signal, the value is 0 or 1, u4 represents the power supply cut-off signal, the value is 0 or 1, and u5 represents the ventilation equipment shutdown signal, the value is 0 or 1;

[0048] Define the linkage logic of the fire extinguishing system. The linkage conditions and operation rules of the fire extinguishing system are as follows:

[0049] If R t >R crit , then, u1=1,u2=1,

[0050] Among them, u1=1 means the water sprinkler equipment is activated, u2=1 means the chemical fire extinguishing agent equipment is activated, R t is the risk score at time t, R crit is the fire risk threshold;

[0051] Control the isolation device, which moves the physical plate of the enclosed area through the power system. The motion path equation is:

[0052]

[0053] Where x(t) is the position of the isolation plate at time t, x0 is the initial position, v is the initial velocity of the isolation plate, and a is the acceleration of the isolation plate.

[0054] As a preferred solution of the new energy vehicle fire monitoring and early warning emergency method based on the Internet of Things described in the present invention, the step of the Internet of Things platform coordinating the fire extinguishing system to perform emergency treatment based on the early warning signal also includes:

[0055] The operation shuts down the power supply and ventilation. The power cut-off signal and the ventilation equipment shut-down signal are controlled by the following conditions:

[0056] If u3=1, then u4=1, u5=1,

[0057] Among them, u4=1 means cutting off the power supply, u5=1 means closing the ventilation equipment,

[0058] Then record all emergency actions and operation data D act :

[0059] D act = {u1,u2,u3,u4,u5,V iso ,R t ,T t ,G t ,C t},

[0060] Among them, D act represents the emergency action data set, V iso is the closed volume of the isolated area, R t For risk scoring, T t , G t , C t They are temperature, gas concentration and charge status data respectively.

[0061] As a preferred solution of the new energy vehicle fire monitoring and early warning emergency method based on the Internet of Things described in the present invention, the monitoring data includes the temperature and gas concentration changes recorded by the sensor, the status logs of all devices and the event development process.

[0062] As a preferred solution of the new energy vehicle fire monitoring and early warning emergency method based on the Internet of Things described in the present invention, wherein: in step S4, the early warning strategy is adjusted based on the monitoring data, and the adjustment method is:

[0063] The monitoring data is cleaned and normalized before uploading to the cloud platform. The normalization formula is:

[0064]

[0065] Among them, X t is the original data at time t, Xmin , X max are the minimum and maximum values ​​of the data, respectively, and X' t is the normalized data,

[0066] Update the model parameters and optimize the deep learning model parameters through gradient descent. The optimization formula is:

[0067]

[0068] Among them, θ is the current model parameter, θ' is the updated model parameter, η is the learning rate,

[0069] is the loss function,

[0070] Combined with the analysis of the latest data, the warning threshold is adjusted dynamically. The adjustment formula is:

[0071] T′ th =T th +αΔT,G′ th =G th +βΔG,

[0072] Among them, T' th is the updated temperature threshold, ΔT is the temperature change rate, G' th is the updated gas concentration threshold, ΔG is the gas concentration change rate, α and β are adjustment coefficients,

[0073] Build a new early warning model, generate a new early warning scoring function based on normalized data and the latest historical data, and the function formula is:

[0074] R′ t =g(X′ t ,H t ),

[0075] Among them, R' t is the optimized risk score, g is the optimized scoring function, including new weights and features;

[0076] Finally, the adjustment record is stored. All adjustment records are stored as:

[0077] D adjust ={θ',T' th ,G' th ,R' t},

[0078] Among them, D adjust is the adjusted record, θ' is the updated model parameter, T' th and G' th are the updated thresholds, R' t The optimized risk score.

[0079] In a second aspect, the present invention provides a new energy vehicle fire monitoring and early warning emergency device based on the Internet of Things, comprising:

[0080] Emergency isolation cabins, fire extinguishing systems, multimodal sensors, IoT platforms, ventilation equipment, alarm devices, and cloud platforms;

[0081] The emergency isolation cabin includes a warning light, a lighting lamp, a camera, a charging socket, a physical isolation device and an electric control cabinet;

[0082] The fire extinguishing system includes water sprinkler equipment and chemical fire extinguishing equipment, which are used to cool down and suppress the fire in case of fire;

[0083] IoT platform, which collects and analyzes charging status and environmental data, executes warning and response instructions, and synchronizes with the cloud platform;

[0084] Ventilation equipment is used to maintain air circulation in the area when there is no fire, and will be closed after an early warning to prevent combustion;

[0085] Alarm device, used to notify surrounding personnel and remote management center of fire information in real time;

[0086] Multimodal sensors for monitoring charging status and environmental data;

[0087] Cloud platform for storing monitoring data, event logs, and adjusting learning models and early warning strategies;

[0088] During use, the fire monitoring and early warning emergency device:

[0089] In the monitoring stage of step S1, the emergency isolation cabin receives the warning signal in advance and stands by;

[0090] In the early warning stage of step S2 and step S3, the area is closed by physical isolation devices and independent power supply control;

[0091] In the fire extinguishing stage of step S3, the fire extinguishing system is activated to link the water spray, chemical fire extinguishing agent and isolation function.

[0092] The beneficial effects of the present invention are as follows: in the data collection stage, the present invention uses a multimodal sensor to monitor the charging status and environmental data of the charging area in real time, and uses the Internet of Things platform to realize data fusion processing, which effectively improves the comprehensiveness and real-time nature of the data. Subsequently, in the risk assessment stage, a deep learning model is introduced to dynamically generate risk scores in combination with historical data and real-time data, and hierarchical warning conditions are further defined. When data such as the temperature and gas concentration of the charging pile are abnormal but do not meet the fire standard, a secondary warning is triggered, the power supply and ventilation equipment are suspended, and detailed abnormal data are recorded; when the smoke sensor and the temperature sensor jointly confirm the fire risk, a primary warning is triggered, and a linkage response is started, including warning lights, alarm devices, and fire extinguishing systems on standby; the intelligent evaluation and linkage mechanism greatly improves the accuracy and response speed of the warning; after entering the emergency handling stage, the Internet of Things platform is used to generate emergency signals, coordinate water spray equipment, chemical fire extinguishing agents and emergency isolation chambers, cut off the fire spread path through physical isolation and multi-device linkage, and simultaneously close the power supply and ventilation of the area, significantly reducing the risk of fire spread; in the post-event data processing stage, all monitoring data are uploaded to the cloud platform, and the warning strategy is optimized in combination with normalization processing and deep learning models, and the threshold is dynamically adjusted and the model parameters are updated.

[0093] In summary, the present invention enhances the ability to predict future risks and improves the overall response efficiency and stability. In addition, the emergency isolation cabin, fire extinguishing system and alarm device of the present invention have clear division of labor, and form an organic coordination in the monitoring, early warning and emergency stages, thus solving the problems of linkage lag and functional separation of traditional systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0094] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0095] Figure 1 It is a schematic flow chart of the new energy vehicle fire monitoring and early warning emergency method of the present invention.

[0096] Figure 2 It is a structural schematic diagram of the new energy vehicle fire monitoring and early warning emergency device of the present invention.

[0097] Figure 3 This is a structural schematic diagram of the emergency isolation cabin of the new energy vehicle fire monitoring and early warning emergency device of the present invention. DETAILED DESCRIPTION

[0098] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0099] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0100] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0101] Example 1, reference Figure 1 , Figure 2 and Figure 3 This embodiment provides a new energy vehicle fire monitoring and early warning emergency method based on the Internet of Things, comprising the following steps:

[0102] Step S1, collecting charging status and environmental data of the charging area in real time through a multimodal sensor, and transmitting the data to the Internet of Things platform for data fusion, the multimodal sensor includes a smoke sensor and a temperature and humidity sensor;

[0103] Environmental data include ambient temperature and humidity, gas concentration and smoke concentration. Gas concentration includes CO2 and H2 concentration.

[0104] Charging status includes voltage, current, charging power and charging device temperature data;

[0105] Step S2: The IoT platform uses the deep learning model and historical charging status and environmental data to conduct risk assessment based on the fused data. The assessment results are pushed to the management personnel in real time and synchronized to the cloud platform.

[0106] Risk assessment methods include:

[0107] If the temperature, gas concentration or charging status of the charging equipment is abnormal but does not meet the fire standard, a secondary warning will be triggered, the charging pile power supply and ventilation equipment will be suspended, and the abnormal temperature value, gas concentration and power fluctuation data will be recorded;

[0108] If the smoke sensor and the temperature sensor both confirm the fire risk, the first-level warning will be triggered immediately and the emergency linkage state will be entered;

[0109] The linkage status includes starting the warning lights, alarm devices and pre-activating the fire extinguishing system on standby;

[0110] The emergency isolation chamber begins preparing for subsequent actions after receiving the warning signal during the data collection phase;

[0111] The steps for risk assessment using deep learning models and historical charging status and environmental data are:

[0112] Perform data fusion and feature extraction, and the fused data is represented as D t :

[0113] D t ={T t ,G t ,C t},

[0114] Among them, T t ={t1,t2,…,t n}, t i represents the data collected by the i-th temperature sensor at time t, n is the number of temperature sensors, G t ={g1,g2,…,g m}, g j represents the data collected by the jth gas sensor at time t, m is the number of gas sensors, C t = {v t ,i t ,p t}, respectively representing the voltage v at time t t 、Current i t , power p t ,

[0115] The risk scoring function is used for risk assessment. The risk scoring function is based on a multi-layer neural network, and the specific structure is:

[0116] R t =f(D t ),H t )=σ(W (3) ·σ(W (2) ·σ(W (1) ·X+b (1) )+b (2) )+b (3) ),

[0117] Among them, R t represents the risk score at time t, H t represents the historical data set, and X represents the fused data D t and historical data H t splicing, σ(x) is the Sigmoid activation function, W (1) ,W (2) ,W (3) are the weight matrices of the first, second and third layers respectively, b (1) ,b (2) ,b (3)are the bias vectors of the corresponding layers respectively. The calculation goal of the risk score is to output R through the model. t Estimate the probability of fire occurrence, the value range is [0,1],

[0118] Perform model training and define the loss function of model training as mean square error MSE. The formula is:

[0119]

[0120] Among them, L represents the loss value of the model, k represents the number of training samples, represents the model prediction risk of the jth sample, R j Represents the true risk value of the jth sample;

[0121] The steps of risk assessment using deep learning models and historical charging status and environmental data also include:

[0122] Add alert conditions, including:

[0123] Secondary warning conditions: When one of the following conditions is met, the secondary warning is triggered:

[0124] max(T t ) <T th ,

[0125] max(G t ) <G th ,

[0126] ΔC t >C diff ,

[0127] Among them, max(T t ) represents the maximum temperature at time t, T th Indicates the second-level warning temperature threshold, max(G t ) represents the maximum gas concentration at time t, G th Indicates the secondary warning gas concentration threshold, ΔC t represents the power fluctuation at time t, C diff Indicates the power fluctuation threshold,

[0128] Level 1 warning conditions: When the following conditions are met at the same time, a level 1 warning is triggered:

[0129] s sm =1 and s tm =1,

[0130] Among them, s sm Indicates smoke confirmation signal, value is 0 or 1, s tm Indicates the temperature confirmation signal, the value is 0 or 1;

[0131] After the warning is triggered, the abnormal data is recorded, including:

[0132] D abn ={max(T t ),max(G t ),ΔC t},

[0133] Among them, D abn represents an abnormal data set, max(T t )、max(G t ) and ΔC t They are abnormal maximum temperature, gas concentration, and power fluctuations;

[0134] Specifically, the risk scoring function is defined through a deep learning model, and the risk score is dynamically generated by combining fused data and historical data. The multi-layer structure model can efficiently extract data features and provide accurate early warnings through preset grading thresholds. At the same time, it supports subsequent analysis and strategy optimization by recording abnormal data, thereby enhancing the intelligence and responsiveness of the fire monitoring system.

[0135] Step S3, based on the early warning signal, the IoT platform coordinates the fire extinguishing system to carry out emergency treatment, and activates water spray, chemical fire extinguishing agent and emergency isolation chamber;

[0136] The emergency isolation chamber uses physical isolation devices to seal off the fire risk area, while cutting off the power supply to the charging piles and shutting down the ventilation equipment;

[0137] Based on the early warning signal, the IoT platform coordinates the fire extinguishing system to carry out emergency response steps as follows:

[0138] The IoT platform generates an emergency signal set U:

[0139] U={u1,u2,u3,u4,u5},

[0140] Among them, u1 represents the water sprinkler equipment start signal, the value is 0 or 1, u2 represents the chemical fire extinguishing agent equipment start signal, the value is 0 or 1, u3 represents the isolation device start signal, the value is 0 or 1, u4 represents the power supply cut-off signal, the value is 0 or 1, and u5 represents the ventilation equipment shutdown signal, the value is 0 or 1;

[0141] Define the linkage logic of the fire extinguishing system. The linkage conditions and operation rules of the fire extinguishing system are as follows:

[0142] If R t >R crit , then, u1=1,u2=1,

[0143] Among them, u1=1 means the water sprinkler equipment is activated, u2=1 means the chemical fire extinguishing agent equipment is activated, R tis the risk score at time t, R crit is the fire risk threshold;

[0144] Control the isolation device, which moves the physical plate of the enclosed area through the power system. The motion path equation is:

[0145]

[0146] Where x(t) is the position of the isolation plate at time t, x0 is the initial position, v is the initial velocity of the isolation plate, and a is the acceleration of the isolation plate;

[0147] Based on the early warning signal, the steps of the IoT platform coordinating the fire extinguishing system to carry out emergency response also include:

[0148] The operation shuts down the power supply and ventilation. The power cut-off signal and the ventilation equipment shut-down signal are controlled by the following conditions:

[0149] If u3=1, then u4=1, u5=1,

[0150] Among them, u4=1 means cutting off the power supply, u5=1 means closing the ventilation equipment,

[0151] Then record all emergency actions and operation data D act :

[0152] D act = {u1,u2,u3,u4,u5,V iso ,R t ,T t ,G t ,C t},

[0153] Among them, D act represents the emergency action data set, V iso is the closed volume of the isolated area, R t For risk scoring, T t , G t , C t They are temperature, gas concentration and charging status data respectively;

[0154] Specifically, this step coordinates the fire extinguishing system and isolation devices through the IoT platform, and combines dynamic risk scoring to achieve multi-device linkage operations, including the activation of the fire extinguishing system, the closure of the isolation area, and the response of the power supply and ventilation systems.

[0155] Step S4: After the emergency response is completed, all monitoring data are uploaded to the cloud platform for adjusting the early warning strategy;

[0156] Monitoring data includes temperature and gas concentration changes recorded by sensors, status logs of all equipment, and the development of events;

[0157] In step S4, the early warning strategy is adjusted based on the monitoring data, and the adjustment method is:

[0158] The monitoring data is cleaned and normalized before uploading to the cloud platform. The normalization formula is:

[0159]

[0160] Among them, X t is the original data at time t, X min , X max are the minimum and maximum values ​​of the data, respectively, and X' t is the normalized data,

[0161] Update the model parameters and optimize the deep learning model parameters through gradient descent. The optimization formula is:

[0162]

[0163] Among them, θ is the current model parameter, θ' is the updated model parameter, η is the learning rate,

[0164] is the loss function,

[0165] Combined with the analysis of the latest data, the warning threshold is adjusted dynamically. The adjustment formula is:

[0166] T′ th =T th +αΔT,G′ th =G th +βΔG,

[0167] Among them, T' th is the updated temperature threshold, ΔT is the temperature change rate, G' th is the updated gas concentration threshold, ΔG is the gas concentration change rate, α and β are adjustment coefficients,

[0168] Build a new early warning model, generate a new early warning scoring function based on normalized data and the latest historical data, and the function formula is:

[0169] R′ t =g(X′ t ,H t ),

[0170] Among them, R' t is the optimized risk score, g is the optimized scoring function, including new weights and features;

[0171] Finally, the adjustment record is stored. All adjustment records are stored as:

[0172] D adjust ={θ',T' th ,G' th ,R' t},

[0173] Among them, D adjust is the adjusted record, θ' is the updated model parameter, T' th and G' th are the updated thresholds, R' t The optimized risk score;

[0174] Specifically, the stability and reliability of the data are improved through cleaning and normalization. The model parameters are optimized and the warning threshold is adjusted to make the warning strategy more efficient and accurate. The new scoring model is used to effectively improve the intelligence of the fire monitoring system.

[0175] This embodiment also provides a new energy vehicle fire monitoring and early warning emergency device based on the Internet of Things, including: an emergency isolation cabin, a fire extinguishing system, a multimodal sensor, an Internet of Things platform, ventilation equipment, an alarm device and a cloud platform;

[0176] Emergency isolation cabin, including warning lights, lighting, cameras, charging sockets, physical isolation devices and electric control cabinets;

[0177] The fire extinguishing system includes water sprinkler equipment and chemical fire extinguishing equipment, which are used to cool down and suppress the fire in case of fire;

[0178] IoT platform, which collects and analyzes charging status and environmental data, executes warning and response instructions, and synchronizes with the cloud platform;

[0179] Ventilation equipment is used to maintain air circulation in the area when there is no fire, and will be closed after an early warning to prevent combustion;

[0180] Alarm device, used to notify surrounding personnel and remote management center of fire information in real time;

[0181] Multimodal sensors for monitoring charging status and environmental data;

[0182] Cloud platform for storing monitoring data, event logs, and adjusting learning models and early warning strategies;

[0183] Fire monitoring and early warning emergency device during use:

[0184] In the monitoring stage of step S1, the emergency isolation cabin receives the warning signal in advance and stands by;

[0185] In the early warning stage of step S2 and step S3, the area is closed by physical isolation devices and independent power supply control;

[0186] In the fire extinguishing stage of step S3, the fire extinguishing system is activated to link the water spray, chemical fire extinguishing agent and isolation function.

[0187] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A new energy vehicle fire monitoring and early warning emergency method based on the Internet of Things, characterized by: include, Step S1, collecting charging status and environmental data of the charging area in real time through a multimodal sensor, and transmitting the data to the Internet of Things platform for data fusion, the multimodal sensor includes a smoke sensor and a temperature and humidity sensor; Step S2: The IoT platform uses the deep learning model and historical charging status and environmental data to conduct risk assessment based on the fused data. The assessment results are pushed to the management personnel in real time and synchronized to the cloud platform. Risk assessment methods include: If the temperature, gas concentration or charging status of the charging equipment is abnormal but does not meet the fire standard, a secondary warning will be triggered, the charging pile power supply and ventilation equipment will be suspended, and the abnormal temperature value, gas concentration and power fluctuation data will be recorded; If the smoke sensor and the temperature sensor both confirm the fire risk, the first-level warning will be triggered immediately and the emergency linkage state will be entered; Step S3, based on the early warning signal, the IoT platform coordinates the fire extinguishing system to carry out emergency treatment, and activates water spray, chemical fire extinguishing agent and emergency isolation chamber; The emergency isolation chamber uses physical isolation devices to seal off the fire risk area, while cutting off the power supply to the charging piles and shutting down the ventilation equipment; Step S4: After the emergency response is completed, all monitoring data are uploaded to the cloud platform for adjusting the early warning strategy.

2. The method for monitoring and warning emergency response of new energy vehicles fire based on the Internet of Things as claimed in claim 1, characterized in that: The environmental data include environmental temperature and humidity, gas concentration and smoke concentration, and the gas concentration includes CO2 and H2 concentration; The charging status includes voltage, current, charging power and charging device temperature data.

3. A new energy vehicle fire monitoring and early warning emergency method based on the Internet of Things as claimed in claim 2, characterized in that: The linkage state includes starting the warning light, the alarm device and pre-activating the fire extinguishing system on standby; The emergency isolation chamber begins preparing for follow-up actions after receiving the early warning signal during the data collection phase.

4. The method for monitoring and warning fire of new energy vehicles based on the Internet of Things as claimed in claim 3 is characterized by: The steps of using the deep learning model and historical charging status and environmental data to perform risk assessment are: Perform data fusion and feature extraction, and the fused data is represented as D t : D t ={T t ,G t ,C t }, Among them, T t ={t1,t2,…,t n }, t i represents the data collected by the i-th temperature sensor at time t, n is the number of temperature sensors, G t ={g1,g2,…,g m }, g j represents the data collected by the jth gas sensor at time t, m is the number of gas sensors, C t = {v t ,i t ,p t }, respectively representing the voltage v at time t t 、Current i t , power p t , The risk scoring function is used for risk assessment. The risk scoring function is based on a multi-layer neural network, and the specific structure is: R t =f(D t ,H t )=σ(W (3) ·σ(W (2) ·σ(W (1) ·X+b (1) )+b (2) )+b (3) ), Among them, R t represents the risk score at time t, H t represents the historical data set, and X represents the fused data D t and historical data H t splicing, σ(x) is the Sigmoid activation function, W (1) ,W (2) ,W (3) are the weight matrices of the first, second and third layers respectively, b (1) ,b (2) ,b (3) are the bias vectors of the corresponding layers respectively. The calculation goal of the risk score is to output R through the model. t Estimate the probability of fire occurrence, the value range is [0,1], Perform model training and define the loss function of model training as mean square error MSE. The formula is: Among them, L represents the loss value of the model, k represents the number of training samples, represents the model prediction risk of the jth sample, R j Represents the true risk value of the jth sample.

5. A new energy vehicle fire monitoring and early warning emergency method based on the Internet of Things as claimed in claim 4, characterized in that: The step of using the deep learning model and historical charging status and environmental data to perform risk assessment also includes: Add alert conditions, including: Secondary warning conditions: When one of the following conditions is met, the secondary warning is triggered: max(T t )<T th , max(G t )<G th , ΔG t >C diff , Among them, max(T t ) represents the maximum temperature at time t, T th Indicates the second-level warning temperature threshold, max(G t ) represents the maximum gas concentration at time t, G th Indicates the secondary warning gas concentration threshold, ΔC t represents the power fluctuation at time t, C diff Indicates the power fluctuation threshold, Level 1 warning conditions: When the following conditions are met at the same time, a level 1 warning is triggered: s sm =1 and s tm =1, Among them, s sm Indicates smoke confirmation signal, value is 0 or 1, s tm Indicates the temperature confirmation signal, the value is 0 or 1; After the warning is triggered, the abnormal data is recorded, including: D abn ={max(T t ),max(G t ),ΔC t }, Among them, D abn represents an abnormal data set, max(T t )、max(G t ) and ΔC t They are abnormal maximum temperature, gas concentration and power fluctuations respectively.

6. A new energy vehicle fire monitoring and early warning emergency method based on the Internet of Things as claimed in claim 5, characterized in that: The steps of the IoT platform coordinating the fire extinguishing system to carry out emergency treatment based on the early warning signal are as follows: The IoT platform generates an emergency signal set U: U={u1,u2,u3,u4,u5}, Among them, u1 represents the water sprinkler equipment start signal, the value is 0 or 1, u2 represents the chemical fire extinguishing agent equipment start signal, the value is 0 or 1, u3 represents the isolation device start signal, the value is 0 or 1, u4 represents the power supply cut-off signal, the value is 0 or 1, and u5 represents the ventilation equipment shutdown signal, the value is 0 or 1; Define the linkage logic of the fire extinguishing system. The linkage conditions and operation rules of the fire extinguishing system are as follows: If R t >R crit , then, u1=1,u2=1, Among them, u1=1 means the water sprinkler equipment is activated, u2=1 means the chemical fire extinguishing agent equipment is activated, R t is the risk score at time t, R crit is the fire risk threshold; Control the isolation device, which moves the physical plate of the enclosed area through the power system. The motion path equation is: Where x(t) is the position of the isolation plate at time t, x0 is the initial position, v is the initial velocity of the isolation plate, and a is the acceleration of the isolation plate.

7. A new energy vehicle fire monitoring and early warning emergency method based on the Internet of Things as claimed in claim 6, characterized in that: The step of the IoT platform coordinating the fire extinguishing system to perform emergency response based on the early warning signal also includes: The operation shuts down the power supply and ventilation. The power cut-off signal and the ventilation equipment shut-down signal are controlled by the following conditions: If u3=1, then u4=1, u5=1, Among them, u4=1 means cutting off the power supply, u5=1 means closing the ventilation equipment, Then record all emergency actions and operation data D act : D act ={u1,u2,u3,u4,u5,V iso ,R t ,T t ,G t ,C t }, Among them, D act represents the emergency action data set, V iso is the closed volume of the isolated area, R t For risk scoring, T t , G t , C t They are temperature, gas concentration and charge status data respectively.

8. The method for monitoring and warning emergency response of new energy vehicles fire based on the Internet of Things as claimed in claim 7, characterized in that: The monitoring data includes temperature and gas concentration changes recorded by sensors, status logs of all equipment, and the development process of events.

9. A new energy vehicle fire monitoring and early warning emergency method based on the Internet of Things as claimed in claim 8, characterized in that: In step S4, the early warning strategy is adjusted based on the monitoring data, and the adjustment method is: The monitoring data is cleaned and normalized before uploading to the cloud platform. The normalization formula is: Among them, X t is the original data at time t, X min , X max are the minimum and maximum values ​​of the data, respectively, and X' t is the normalized data, Update the model parameters and optimize the deep learning model parameters through gradient descent. The optimization formula is: Among them, θ is the current model parameter, θ' is the updated model parameter, η is the learning rate, is the loss function, Combined with the analysis of the latest data, the warning threshold is adjusted dynamically. The adjustment formula is: T′ th =T th +αΔT,G′ th =G th +βΔG, Among them, T' th is the updated temperature threshold, ΔT is the temperature change rate, G' th is the updated gas concentration threshold, ΔG is the gas concentration change rate, α and β are adjustment coefficients, Build a new early warning model, generate a new early warning scoring function based on normalized data and the latest historical data, and the function formula is: R' t =g(X' t ,H t ), Among them, R' t is the optimized risk score, g is the optimized scoring function, including new weights and features; Finally, the adjustment record is stored. All adjustment records are stored as: D adjust ={θ',T' th ,G' th ,R' t }, Among them, D adjust is the adjusted record, θ' is the updated model parameter, T' th and G' th are the updated thresholds, R' t The optimized risk score.

10. A new energy vehicle fire monitoring and early warning emergency device based on the Internet of Things, based on a new energy vehicle fire monitoring and early warning emergency method based on the Internet of Things according to any one of claims 1 to 9, characterized in that: Including emergency isolation cabins, fire extinguishing systems, multimodal sensors, IoT platforms, ventilation equipment, alarm devices and cloud platforms; The emergency isolation cabin includes a warning light, a lighting lamp, a camera, a charging socket, a physical isolation device and an electric control cabinet; The fire extinguishing system includes water sprinkler equipment and chemical fire extinguishing equipment, which are used to cool down and suppress the fire in case of fire; IoT platform, which collects and analyzes charging status and environmental data, executes warning and response instructions, and synchronizes with the cloud platform; Ventilation equipment is used to maintain air circulation in the area when there is no fire, and will be closed after an early warning to prevent combustion; Alarm device, used to notify surrounding personnel and remote management center of fire information in real time; Multimodal sensors for monitoring charging status and environmental data; Cloud platform for storing monitoring data, event logs, and adjusting learning models and early warning strategies; During use, the fire monitoring and early warning emergency device: In the monitoring stage of step S1, the emergency isolation cabin receives the warning signal in advance and stands by; In the early warning stage of step S2 and step S3, the area is closed by physical isolation devices and independent power supply control; In the fire extinguishing stage of step S3, the fire extinguishing system is activated to link the water spray, chemical fire extinguishing agent and isolation function.

Images