Internet of Things accurate fire-fighting monitoring and fire extinguishing system based on intelligent cloud platform management

Abstract

The invention provides an Internet of Things accurate fire-fighting monitoring and fire extinguishing system based on smart cloud platform management, and relates to the technical field of accurate fire-fighting monitoring and fire extinguishment. Historical power consumption data and fault types of a power distribution cabinet are acquired through an electrical acquisition module to form time sequence data, and fast Fourier transform is performed on the time sequence data to obtain frequency spectrum features; collected current frequency spectrum features are input into the electrical judgment module to establish a fault recognition model, a current fault type is recognized, historical environment data in the power distribution cabinet are collected through the environment collection module and smoothed, and stable environment data are formed. And the fire-fighting monitoring module is used for establishing a fire-fighting monitoring model based on the Bayesian theorem, identifying a fire type according to a fault type and the characteristics of each environment type, and sending the fire-fighting monitoring model to the fire-fighting monitoring module, so that the fire-fighting monitoring module can carry out fire-fighting monitoring according to the fault type and the characteristics of each environment type. And selecting a fire extinguishing mode and alarming.

Description

Technical Field

[0001] The present invention relates to the technical field of precise fire monitoring and fire extinguishing, and specifically to an Internet of Things precise fire monitoring and fire extinguishing system based on smart cloud platform management. Background Art

[0002] At present, fire monitoring and fire extinguishing systems for electrical equipment, especially distribution cabinets, mostly rely on simple temperature and smoke detection technologies, which cannot accurately identify the type of fire and accurately determine whether fire extinguishing is needed. Because traditional fire warning systems are based only on single environmental indicators such as temperature and smoke concentration, there are often false alarms and missed alarms, which leads to the failure to take timely and accurate fire extinguishing measures, and even causes equipment damage or fire spread.

[0003] In the prior art, the publication number CN11580848A discloses a comprehensive protection system for timely warning and flame extinguishing of electrical fires in distribution cabinets, which includes: (1) a distribution cabinet electrical fire monitoring system, which collects and monitors the signal changes of temperature, current, voltage and residual current in the distribution cabinet circuit; (2) a distribution cabinet electrical fire warning system: predicts the fire without fire, smoldering and open flame conditions; (3) a distribution cabinet automatic fire extinguishing system: triggers the automatic flame extinguishing device in the distribution cabinet to extinguish the flame. However, this scheme is overly dependent on electrical parameters and historical data, lacks the collection and comprehensive analysis of environmental data, and the neural network training relies on historical data. There may be problems such as insufficient training data and poor adaptability, and it is easily affected by the fluctuation of the collected data.

[0004] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0005] The purpose of the present invention is to provide an IoT precise fire monitoring and extinguishing system based on smart cloud platform management to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions: An IoT precise fire monitoring and extinguishing system based on smart cloud platform management, including: An electrical collection module is used to collect historical power consumption data of the power distribution cabinet and print a timestamp on the fault type to form time series data, divide the time series data into interval data at equal time intervals, form a sliding window with multiple interval data, perform fast Fourier transform on the time series data in the window, and obtain the corresponding spectrum features in each window; An electrical judgment module, which is used to establish a fault identification model, train the model by taking the corresponding frequency spectrum features in the previous window as the input of the model and the fault type in the previous interval data as the output of the model, and output the corresponding fault type in the current interval data according to the collected frequency spectrum features in the current window; An environment collection module is used to collect historical environmental data in the power distribution cabinet, perform discrete sampling at the same time interval, obtain discrete environmental data, and smooth the discrete environmental data set through an exponentially weighted moving average algorithm to form stable environmental data; An environment judgment module, which is used to obtain the safety threshold of each environment type, and calculate the exceeding coefficient and the duration of exceeding the standard for each environment type according to the stable environment data, calculate the change trend coefficient of each environment type according to the acquisition timestamp of the stable environment data, and calculate the characteristics of each environment type according to the exceeding coefficient, the change trend coefficient and the duration of exceeding the standard; The fire monitoring module is used to classify the fire types of the distribution cabinet, establish a fire monitoring model according to the Bayesian theorem, identify the fire type through the characterization factors composed of the fault type and the characteristics of each environmental type, and select the fire extinguishing method according to the fire type, and at the same time issue an alarm and send the alarm information to the monitoring personnel.

[0007] Furthermore, the power consumption data includes total incoming line current, incoming line voltage, outgoing line current, and outgoing line voltage; The fault types include current overload, short circuit fault, and poor contact: The frequency spectrum characteristics include the number of frequency components and the amplitude and phase of each frequency component.

[0008] Furthermore, according to the time sequence, the time series data is processed at equal time intervals. Split into interval data to ensure that all continuous interval data completely cover the time series data, and when the end of the time series data is not enough to form a complete interval data, zero padding is used to mark the time series data fragments corresponding to each generated time interval data as: in, Indicates the time series data The interval data corresponding to the time interval, is a positive integer, and ; Bundle A continuous interval of data forms a window width , and when the end of the continuous interval data is not enough to form a complete window width, zero padding is used, and the continuous windows are calibrated as follows: in, Indicates The time series data corresponding to the window, is the width of the window, Indicates the time series data The interval data corresponding to the time interval.

[0009] Furthermore, the calculation formula for performing fast Fourier transform on the time series data in the window is: in, is the frequency domain signal of the time series data frequency components, is the time domain signal of the time series data samples, is the total number of samples, is an imaginary unit, Represents the frequency domain frequency components, .

[0010] Furthermore, the fault identification model is based on a long short-term memory neural network, and the mathematical calculation formula of the model is: The long short-term memory neural network includes an input gate, a forget gate and an output gate; The equation of the forget gate is: in, is the current switch state of the forget gate, represents the sigmoid function, is the forget gate weight, is the forget gate bias parameter, is the hidden state at the previous moment, is the corresponding spectrum feature in the previous window; Determine how much of the unit state at the previous moment needs to be retained to the current moment; The equation for the input gate is: in, is the current switch state of the input gate, represents the sigmoid function, is the input gate weight, is the input gate bias parameter, is the output hidden state of the previous moment, is the corresponding spectrum feature in the previous window; in, Candidate cell states, To represent the hyperbolic tangent function, is the candidate cell weight, is the candidate cell state bias parameter, is the hidden state at the previous moment, is the corresponding spectrum feature in the previous window; Determine how much of the network's input data needs to be saved to the unit state at the current moment; The update equation is: in, Current cell state, is the cell state at the previous moment; The output gate equation is: in, is the current switch state of the output gate, represents the sigmoid function, where is the output gate weight, is the output gate bias parameter, is the hidden state at the previous moment, is the corresponding spectrum feature in the previous window, is the fault type corresponding to the current interval data; Controls how much of the current unit state needs to be output to the current output value; Among them, the mathematical expressions of the sigmoid function and the hyperbolic tangent function are: in, is the sigmoid function, is the input of the sigmoid function, is the hyperbolic tangent function, Input to the function.

[0011] Furthermore, the environmental data includes four environmental types, namely, smoke, temperature, CO2, and VOC; In chronological order, the environmental data is sorted at equal time intervals. Perform discrete sampling to obtain discrete environmental data, which is marked as: in, For Environmental data at all times, Respectively in The concentration of smoke, temperature, CO2, and VOC in the environmental data at the moment, For environmental data, For the Environmental data collected at time intervals.

[0012] Furthermore, the calculation formula for smoothing the discrete environmental data using the exponentially weighted moving average algorithm is: in, For Stable environmental data at all times, For Environmental data at all times, is the smoothing factor, ; The expression of stable environment data is: in, For Stable environmental data at all times, Respectively in The concentration of smoke, temperature, CO2, and VOC in the stable environmental data at all times. To stabilize environmental data, For the time interval points.

[0013] Furthermore, the calculation formula for calculating the excess coefficient of each environment type is: in, They are the excess coefficients of smoke, temperature, CO2, and VOC in the stable environmental data. They are the safety thresholds for smoke, temperature, CO2, and VOC. They are the concentrations of smoke, temperature, CO2, and VOC in the environmental data; The calculation formula for calculating the change trend coefficient of each environment type is: in, They are the change trend coefficients of smoke, temperature, CO2 and VOC respectively. is the data collection time interval; The characteristics of each environment type are calculated based on the exceeding coefficient, changing trend coefficient and exceeding duration, and the calculation formula for the environmental characteristic data set is: in, They are the characteristics of smoke, temperature, CO2, and VOC. They are the smoke exceeding standard coefficient, change trend coefficient, and weight of exceeding standard duration. They are the temperature exceeding standard coefficient, change trend coefficient, and weight of exceeding standard duration. They are the CO2 exceeding standard coefficient, changing trend coefficient, and weight of exceeding standard duration. They are the VOC exceeding standard coefficient, changing trend coefficient, and weight of exceeding standard duration. They are the duration of smoke, temperature, CO2 and VOC exceeding the standard. .

[0014] Further, the fire types include smoldering fire and open fire; The calculation formula for establishing the fire monitoring model based on Bayesian theorem is: According to the fire types in the historical data, the probability of each fire type occurring is calculated, that is, the prior probability of each fire type: in, For each fire type The prior probability of is the number of fires in historical data, is the indicative function, For the historical data Type of accident, , is the fire type in the historical data, Indicates a smoldering fire. Represents bright fire, is the smoothing coefficient; The formula of the indicative function is: ; Calculate the likelihood probability of the characterization factors under each fire type: in, For a given fire type Characteristic factors observed below The probability of For a given fire type The following observation Characteristic Factors The probability of is the number of influencing factors, is a positive integer; The posterior probability of each fire type occurring is calculated according to Bayes' theorem: in, Fire type The posterior probability of is a normalization constant; Establish a decision model for fire type occurrence: in, The type of fire that occurred was finally identified; The logic for selecting a fire extinguishing method is as follows: When an open fire is identified, immediately open the perfluoroacetylene fire extinguishing device to extinguish the fire; When a smoldering fire is identified, immediately use the perfluoroacetylene fire extinguishing device intermittently to extinguish the fire.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: the historical power consumption data and fault types of the power distribution cabinet are collected by the electrical acquisition module to form time series data, and the time series data is subjected to fast Fourier transform to obtain spectrum characteristics, and the current spectrum characteristics collected are input into the electrical judgment module to establish a fault identification model, and the current fault type is identified; the historical environmental data in the power distribution cabinet is collected by the environment acquisition module and smoothed to form stable environmental data, and the exceeding coefficient, changing trend coefficient and statistical exceeding duration of each environmental type are calculated based on the stable environmental data, and the characteristics of each environmental type are calculated; the fire monitoring module establishes a fire monitoring model based on the Bayesian theorem, and the fire type is identified according to the fault type and the characteristics of each environmental type, and the fire extinguishing method and alarm are selected; Based on the power consumption data of the electrical acquisition module, the present invention also includes environmental data in the analysis scope, and obtains stable environmental data through smoothing. By combining historical power consumption data, fault type and environmental data, the system can realize multi-dimensional monitoring and identification of fire occurrence, provide more accurate fire warning, establish a fire monitoring model through Bayesian theorem, and comprehensively consider multiple factors such as fault type and environmental characteristics to identify the fire type. In the case of complex interweaving of multiple factors, more accurate fire prediction and risk assessment are provided; The present invention also extracts spectrum features from electrical data through fast Fourier transform, which can effectively identify frequency features caused by faults or anomalies in the electrical system, and help quickly identify the type of fault through spectrum analysis. It can achieve dynamic and real-time monitoring, and timely discover potential faults in the electrical system. By calculating the exceeding coefficient, changing trend coefficient, and statistical exceeding duration of each environment type, it provides a deep understanding of environmental changes. It can help the system dynamically evaluate the abnormality of the current environment, so as to make more accurate fire predictions and firefighting decisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a schematic diagram of the overall system structure of the present invention. DETAILED DESCRIPTION

[0017] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0018] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0019] Example

[0020] See also Figure 1 , the present invention provides a technical solution: An IoT precise fire monitoring and extinguishing system based on smart cloud platform management includes an electrical acquisition module, an electrical judgment module, an environmental acquisition module, an environmental judgment module and a fire monitoring module, wherein: The electrical acquisition module is used to collect historical power consumption data and fault types of the distribution cabinet and print timestamps to form time series data. The time series data is divided into interval data at equal time intervals. Multiple interval data form a sliding window, and the time series data in the window is fast Fourier transformed to obtain the corresponding frequency spectrum features in each window.

[0021] In this embodiment, the power consumption data includes the total incoming line current, the incoming line voltage, the outgoing line current, and the outgoing line voltage; The fault types include current overload, short circuit fault, and poor contact: The frequency spectrum characteristics include the number of frequency components and the amplitude and phase of each frequency component.

[0022] Current overload means that the current in the distribution cabinet or electrical equipment exceeds its rated carrying capacity. When the current exceeds the maximum rated value of the equipment, it may cause the electrical equipment to overheat and cause a fire. The total incoming current refers to the current in the incoming line of the distribution cabinet. If the incoming current exceeds the maximum carrying value designed for the distribution cabinet, it can be determined that the system is overloaded. This overload may cause the internal components of the distribution cabinet, such as wires, switches, and contactors, to overheat, and ultimately cause a fire. The outgoing current is the current distributed from the distribution cabinet to downstream electrical equipment. If the current of a certain outgoing line is higher than the load capacity designed for the line, this may also indicate that a branch circuit may be overloaded, resulting in local heating and causing a fire.

[0023] A short circuit is when the positive and negative electrodes in a circuit are in direct contact or a low-resistance path appears, causing an instantaneous increase in current. Since the current increases sharply during a short circuit, the heat generated can instantly cause a fire. When a short circuit occurs, due to the rapid increase in current, the total incoming current and the current of each outgoing line may change abnormally. In particular, the incoming current will soar rapidly, while the change in the outgoing current may be uneven, usually accompanied by a sudden drop in the voltage of the incoming and outgoing lines.

[0024] Poor contact refers to the loose, corroded, worn, etc. contact points in electrical equipment, such as wiring terminals and switch contacts, which lead to poor electrical connections, increased resistance, and local heating. If this continues for a long time, it will cause a fire. Poor contact usually leads to unstable current flow, and the incoming or outgoing current may fluctuate, which is manifested as unstable current or periodic rise and fall. This is because the local resistance generated by the contact point affects the smoothness of the current flow. At the same time, because poor contact will increase the contact resistance of the circuit, the voltage may also be unstable or fluctuate. When the resistance increases, the voltage may rise or fall abnormally, which is usually reflected in the fluctuation of the incoming and outgoing voltages.

[0025] The four key electrical parameters, total incoming current, incoming voltage, outgoing current and outgoing voltage, play a very important role in the fire monitoring system. They can judge abnormal phenomena such as current overload, short circuit fault and poor contact in real time by reflecting the load status, current change and voltage fluctuation of electrical equipment, so as to more accurately identify the type of fire, because current overload usually causes smoldering fire, short circuit fault almost always causes open fire, because the current increases rapidly, high temperature and arc are generated, directly igniting surrounding objects, and poor contact usually causes smoldering fire, but if the fault persists and overheats, it may turn into an open fire.

[0026] In this embodiment, the time series data is processed in time sequence at equal time intervals. Split into interval data to ensure that all continuous interval data completely cover the time series data, and when the end of the time series data is not enough to form a complete interval data, zero padding is used to mark the time series data fragments corresponding to each generated time interval data as: in, Indicates the time series data The interval data corresponding to the time interval, is a positive integer, and ; Bundle A continuous interval of data forms a window width , and when the end of the continuous interval data is not enough to form a complete window width, zero padding is used, and the continuous windows are calibrated as follows: in, Indicates The time series data corresponding to the window, is the width of the window, Indicates the time series data The interval data corresponding to the time interval.

[0027] Fast Fourier transform can convert time domain signals into frequency domain signals, revealing the frequency components of the signal. The normal operation of electrical equipment will show certain frequency characteristics, and when a fault occurs, the frequency components of the signal will often change. For example, faults such as current overload, short circuit fault, and poor contact will cause specific frequency changes in the current signal. Fast Fourier transform can help capture these changes. For example, short circuit faults usually produce large instantaneous current fluctuations, which will appear as a sudden increase in low-frequency components in the frequency domain; while current overload or poor contact may lead to relatively stable signal changes, but there will also be frequency offsets or increased noise.

[0028] In this embodiment, the calculation formula for performing fast Fourier transform on the time series data in the window is: in, is the frequency domain signal of the time series data frequency components, is the time domain signal of the time series data samples, is the total number of samples, is an imaginary unit, Represents the frequency domain frequency components, .

[0029] The electrical judgment module is used to establish a fault identification model. The model is trained by taking the corresponding frequency spectrum features in the previous window as the input of the model and the fault type in the previous interval data as the output of the model. The corresponding fault type in the current interval data is output based on the collected frequency spectrum features in the current window.

[0030] Long short-term memory neural network is a deep learning model that is good at processing time series data. It has a strong memory capacity and can capture long-term dependencies in time series data. Long short-term memory neural network is very suitable for processing such continuous and time-varying signals in power systems. It can process time series signals such as current and voltage, and accurately identify system state changes by learning the patterns of historical data. By training the long short-term memory neural network model, it can identify different types of faults such as current overload, short circuit fault and poor contact from the spectrum characteristics.

[0031] In this embodiment, the fault identification model is based on a long short-term memory neural network, and the mathematical calculation formula of the model is: The long short-term memory neural network includes an input gate, a forget gate and an output gate; The equation of the forget gate is: in, is the current switch state of the forget gate, represents the sigmoid function, is the forget gate weight, is the forget gate bias parameter, is the hidden state at the previous moment, is the corresponding spectrum feature in the previous window; Determine how much of the unit state at the previous moment needs to be retained to the current moment; The equation for the input gate is: in, is the current switch state of the input gate, represents the sigmoid function, is the input gate weight, is the input gate bias parameter, is the output hidden state of the previous moment, is the corresponding spectrum feature in the previous window; in, Candidate cell states, To represent the hyperbolic tangent function, is the candidate cell weight, is the candidate cell state bias parameter, is the hidden state at the previous moment, is the corresponding spectrum feature in the previous window; Determine how much of the network's input data needs to be saved to the unit state at the current moment; The update equation is: in, Current cell state, is the cell state at the previous moment; The output gate equation is: in, is the current switch state of the output gate, represents the sigmoid function, is the output gate weight, is the output gate bias parameter, is the hidden state at the previous moment, is the corresponding spectrum feature in the previous window, is the fault type corresponding to the current interval data; Controls how much of the current unit state needs to be output to the current output value; Among them, the mathematical expressions of the sigmoid function and the hyperbolic tangent function are: in, is the sigmoid function, Input for the sigmoid function, is the hyperbolic tangent function, Input to the function.

[0032] The environment acquisition module is used to collect historical environmental data in the distribution cabinet, perform discrete sampling at the same time interval, obtain discrete environmental data, and smooth the discrete environmental data set through an exponentially weighted moving average algorithm to form stable environmental data.

[0033] In this embodiment, the environmental data includes four environmental types, namely smoke, temperature, CO2, and VOC.

[0034] Environmental data such as smoke, temperature, CO2, and VOC are often interfered with by various factors in the distribution cabinet, such as equipment switch operation, air flow changes, local pollutant concentration fluctuations, etc. This may cause the data to be discrete and volatile, resulting in delays or errors in judgment and response, seriously affecting the accuracy of the judgment of the fire type.

[0035] The exponentially weighted moving average algorithm smoothes historical data by assigning greater weights to recent data points. This can effectively reduce accidental or short-term fluctuations and highlight long-term trends and patterns. Especially when environmental data fluctuates greatly, the exponentially weighted moving average algorithm can effectively eliminate large fluctuations. This smoothing process can help the system reduce overreaction to instantaneous abnormal fluctuations, making the monitoring data more stable, thereby improving the accuracy of fire warnings.

[0036] In this embodiment, the environmental data is sorted in time order at equal time intervals. Perform discrete sampling to obtain discrete environmental data, which is marked as: in, For Environmental data at all times, Respectively in The concentration of smoke, temperature, CO2, and VOC in the environmental data at the moment, For environmental data, For the Environmental data collected at time intervals.

[0037] In this embodiment, the calculation formula for smoothing discrete environmental data using the exponentially weighted moving average algorithm is: in, For Stable environmental data at all times, For Environmental data at all times, is the smoothing factor, ; The expression of stable environment data is: in, For Stable environmental data at all times, Respectively in The concentration of smoke, temperature, CO2, and VOC in the stable environmental data at all times. To stabilize environmental data, For the time interval points.

[0038] The environmental judgment module is used to obtain the safety threshold of each environment type, and calculate the excess coefficient and statistical excess duration of each environment type based on the stable environment data, calculate the change trend coefficient of each environment type based on the collection timestamp of the stable environment data, and calculate the characteristics of each environment type based on the excess coefficient, change trend coefficient and excess duration.

[0039] By introducing comprehensive indicators such as the excess coefficient, the change trend coefficient and the duration of the excess, a more detailed analysis of environmental data can be performed. The system can track the changing trend of environmental data in real time and identify sudden anomalies and chronic changes. This means that when parameters such as smoke concentration and temperature begin to rise, the system can not only detect these changes, but also identify whether these changes are fast or slow. For example, rapidly rising temperatures and smoke concentrations may indicate an open fire, while slower temperature changes and longer durations may indicate a smoldering fire. Such detailed analysis can help the system to alarm in advance and reduce losses caused by fire.

[0040] By calculating smoke characteristics, temperature characteristics, CO2 characteristics and VOC characteristics separately, this method can comprehensively analyze the impact of different environmental parameters on fire. This allows the system to adapt to different fire scenarios more flexibly. For example, in some scenarios, the smoke concentration may vary greatly, while in other scenarios, the concentration of CO2 and VOC may be better fire indicators. The flexibility of this method enables it to provide more accurate judgments in changing environments.

[0041] By analyzing the duration of exceeding the standard and the coefficient of change trend, this method can not only identify the occurrence of fire, but also determine the different stages of the fire. For example, when the smoke concentration and temperature rise rapidly, the system can determine that it is an open fire; if the temperature rises slowly and lasts for a long time, the system can identify it as a smoldering fire. This refined fire stage judgment can help the fire protection system adopt different response strategies.

[0042] In this embodiment, the calculation formula for calculating the excess coefficient of each environment type is: in, They are the excess coefficients of smoke, temperature, CO2, and VOC in the stable environmental data. They are the safety thresholds for smoke, temperature, CO2, and VOC. They are the concentrations of smoke, temperature, CO2, and VOC in the environmental data respectively.

[0043] The excess coefficient measures the deviation between the current environmental data value and the safety threshold. It reflects the degree of deviation of the current environmental state from the normal value. The excess coefficient helps to determine the stage of fire. If an environmental factor such as smoke, temperature, CO2 or VOC exceeds the set safety threshold, the excess coefficient will become larger. This coefficient helps the system quickly identify whether the environment is in a dangerous state. The higher the excess coefficient, the more serious the environmental abnormality and the greater the possibility of fire.

[0044] The calculation formula for calculating the change trend coefficient of each environment type is: in, They are the change trend coefficients of smoke, temperature, CO2 and VOC respectively. is the data collection time interval.

[0045] The change trend coefficient is used to measure the trend of a certain environmental factor over time, and is usually calculated by the change rate of the data collected by the timestamp. It can reflect whether the environmental data is rising rapidly, rising slowly, or tending to be stable or falling. The larger the change trend coefficient, the faster the environment changes, which may mean that the fire develops rapidly, especially the fast-burning fire such as the open fire.

[0046] The characteristics of each environment type are calculated based on the exceeding coefficient, changing trend coefficient and exceeding duration, and the calculation formula for the environmental characteristic data set is: in, They are the characteristics of smoke, temperature, CO2, and VOC. They are the smoke exceeding standard coefficient, change trend coefficient, and weight of exceeding standard duration. They are the temperature exceeding standard coefficient, change trend coefficient, and weight of exceeding standard duration. They are the CO2 exceeding standard coefficient, changing trend coefficient, and weight of exceeding standard duration. They are the VOC exceeding standard coefficient, changing trend coefficient, and weight of exceeding standard duration. They are the duration of smoke, temperature, CO2 and VOC exceeding the standard. .

[0047] Environmental features conduct a comprehensive analysis of environmental data to more accurately characterize the current environmental status, predict the occurrence of fires, and identify fire types. Environmental features reflect the relationship between environmental types such as smoke density, temperature, CO2 concentration, VOC concentration, and normal safety status, helping the system determine whether there is a fire and whether the fire type is open fire or smoldering fire.

[0048] The fire monitoring module is used to classify the fire types of the distribution cabinet, establish a fire monitoring model according to the Bayesian theorem, identify the fire type through the characterization factors composed of the fault type and the characteristics of each environmental type, and select the fire extinguishing method according to the fire type, and at the same time issue an alarm and send the alarm information to the monitoring personnel.

[0049] One of the biggest advantages of Bayesian theorem is its ability to handle incomplete information or uncertainty. Sensor data is often affected by various noises and environmental interferences. For example, in the early stages of a fire, temperature rise may be accompanied by strong signal interference, or sensor errors may cause data fluctuations. Bayesian models can smooth observed data based on prior probabilities, that is, the probability of fire occurrence established based on historical data and expert experience, and give the best fire identification results under uncertainty. The identification of fire types usually requires the combination of data from multiple sensors, such as temperature, smoke concentration, CO2, VOC, and these data often have complex correlations. Bayesian networks can jointly reason about these multidimensional data and analyze the conditional dependencies between them. For example, when a distribution cabinet catches fire, the temperature and smoke concentration will change at the same time, and other power consumption factors, such as current and voltage, may also affect the type and development of the fire. Bayesian networks can effectively combine the information of these multiple variables, thereby improving the accuracy of fire type identification.

[0050] The Bayesian model provides an efficient, accurate and flexible solution for the IoT precision fire monitoring and extinguishing system managed by the smart cloud platform, which can improve the fire monitoring and response capabilities in complex and changing environments.

[0051] Open fire refers to a fire with visible flames. This usually occurs when electrical equipment has a short circuit, overload, arc, etc. The high temperature generated will cause the combustible materials around the electrical equipment, such as cables, plastic casings, insulating materials, etc. to burn rapidly and form open flames.

[0052] Smoldering fire refers to a fire that occurs without an open flame and is characterized by low temperature and slow burning. Smoldering fire is usually caused by overheating of electrical equipment and slow burning of insulating materials. There is no obvious flame during the process, which is mainly manifested as the oxidation process of substances or the internal combustion of certain objects.

[0053] In this embodiment, the fire types include smoldering fire and open fire; The calculation formula for establishing the fire monitoring model based on Bayesian theorem is: According to the fire types in the historical data, the probability of each fire type occurring is calculated, that is, the prior probability of each fire type: in, For each fire type The prior probability of is the number of fires in historical data, is the indicative function, For the historical data Type of accident, , is the fire type in the historical data, Indicates a smoldering fire. Represents bright fire, is the smoothing coefficient; The formula of the indicative function is: ; Calculate the likelihood probability of the characterization factors under each fire type: in, For a given fire type Characteristic factors observed below The probability of For a given fire type The following observation Characteristic Factors The probability of is the number of influencing factors, is a positive integer; The posterior probability of each fire type occurring is calculated according to Bayes' theorem: in, Fire type The posterior probability of is a normalization constant; Establish a decision model for fire type occurrence: in, The type of fire that occurred was finally identified; The logic for selecting a fire extinguishing method is as follows: When an open fire is identified, immediately open the perfluoroacetylene fire extinguishing device to extinguish the fire; When a smoldering fire is identified, immediately use the perfluoroacetylene fire extinguishing device intermittently to extinguish the fire.

[0054] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0055] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product. Those skilled in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0056] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0057] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. An IoT precise fire monitoring and extinguishing system based on smart cloud platform management, characterized in that: include: An electrical collection module is used to collect historical power consumption data of the power distribution cabinet and print a timestamp on the fault type to form time series data, divide the time series data into interval data at equal time intervals, form a sliding window with multiple interval data, perform fast Fourier transform on the time series data in the window, and obtain the corresponding spectrum features in each window; An electrical judgment module, which is used to establish a fault identification model, train the model by taking the corresponding frequency spectrum features in the previous window as the input of the model and the fault type in the previous interval data as the output of the model, and output the corresponding fault type in the current interval data according to the collected frequency spectrum features in the current window; An environment collection module is used to collect historical environmental data in the power distribution cabinet, perform discrete sampling at the same time interval, obtain discrete environmental data, and smooth the discrete environmental data set through an exponentially weighted moving average algorithm to form stable environmental data; An environment judgment module, which is used to obtain the safety threshold of each environment type, and calculate the exceeding coefficient and the duration of exceeding the standard for each environment type according to the stable environment data, calculate the change trend coefficient of each environment type according to the acquisition timestamp of the stable environment data, and calculate the characteristics of each environment type according to the exceeding coefficient, the change trend coefficient and the duration of exceeding the standard; The fire monitoring module is used to classify the fire types of the distribution cabinet, establish a fire monitoring model according to the Bayesian theorem, identify the fire type through the characterization factors composed of the fault type and the characteristics of each environmental type, and select the fire extinguishing method according to the fire type, and at the same time issue an alarm and send the alarm information to the monitoring personnel.

2. According to claim 1, an IoT precise fire monitoring and extinguishing system based on smart cloud platform management is characterized in that: The electricity consumption data includes total incoming line current, incoming line voltage, outgoing line current, and outgoing line voltage; The fault types include current overload, short circuit fault, and poor contact: The frequency spectrum characteristics include the number of frequency components and the amplitude and phase of each frequency component.

3. According to claim 1, the IoT precise fire monitoring and fire extinguishing system based on smart cloud platform management is characterized in that: In chronological order, the time series data is processed at equal time intervals. Split into interval data to ensure that all continuous interval data completely cover the time series data, and when the end of the time series data is not enough to form a complete interval data, zero padding is used to mark the time series data fragments corresponding to each generated time interval data as: in, Indicates the time series data The interval data corresponding to the time interval, is a positive integer, and ; Bundle A continuous interval of data forms a window width , and when the end of the continuous interval data is not enough to form a complete window width, zero padding is used, and the continuous windows are calibrated as follows: in, Indicates The time series data corresponding to the window, is the width of the window, Indicates the time series data The interval data corresponding to the time interval.

4. According to claim 1, the IoT precise fire monitoring and fire extinguishing system based on smart cloud platform management is characterized in that: The calculation formula for performing fast Fourier transform on the time series data in the window is: in, is the frequency domain signal of the time series data frequency components, is the time domain signal of the time series data samples, is the total number of samples, is an imaginary unit, Represents the frequency domain frequency components, .

5. According to claim 1, the IoT precise fire monitoring and fire extinguishing system based on smart cloud platform management is characterized in that: The fault identification model is based on a long short-term memory neural network, and the mathematical calculation formula of the model is: The long short-term memory neural network includes an input gate, a forget gate and an output gate; The equation of the forget gate is: in, is the current switch state of the forget gate, represents the sigmoid function, is the forget gate weight, is the forget gate bias parameter, is the hidden state at the previous moment, is the corresponding spectrum feature in the previous window; Determine how much of the unit state at the previous moment needs to be retained to the current moment; The equation for the input gate is: in, is the current switch state of the input gate, represents the sigmoid function, is the input gate weight, is the input gate bias parameter, is the output hidden state of the previous moment, is the corresponding spectrum feature in the previous window; in, Candidate cell states, To represent the hyperbolic tangent function, is the candidate cell weight, is the candidate cell state bias parameter, is the hidden state at the previous moment, is the corresponding spectrum feature in the previous window; Determine how much of the network's input data needs to be saved to the unit state at the current moment; The update equation is: in, Current cell state, is the cell state at the previous moment; The output gate equation is: in, is the current switch state of the output gate, represents the sigmoid function, is the output gate weight, is the output gate bias parameter, is the hidden state at the previous moment, is the corresponding spectrum feature in the previous window, is the fault type corresponding to the current interval data; Controls how much of the current unit state needs to be output to the current output value; Among them, the mathematical expressions of the sigmoid function and the hyperbolic tangent function are: in, is the sigmoid function, is the input of the sigmoid function, is the hyperbolic tangent function, Input to the function.

6. According to claim 1, the IoT precise fire monitoring and fire extinguishing system based on smart cloud platform management is characterized in that: The environmental data includes four environmental types, namely smoke, temperature, CO2, and VOC; In chronological order, the environmental data is sorted at equal time intervals. Perform discrete sampling to obtain discrete environmental data, which is marked as: in, For Environmental data at all times, Respectively in The concentration of smoke, temperature, CO2, and VOC in the environmental data at the moment, For environmental data, For the Environmental data collected at time intervals.

7. According to claim 1, the IoT precise fire monitoring and fire extinguishing system based on smart cloud platform management is characterized in that: The calculation formula for smoothing discrete environmental data using the exponentially weighted moving average algorithm is: in, For Stable environmental data at all times, For Environmental data at all times, is the smoothing factor, ; The expression of stable environment data is: in, For Stable environmental data at all times, Respectively in The concentration of smoke, temperature, CO2, and VOC in the stable environmental data at all times. To stabilize environmental data, For the time interval points.

8. The IoT precise fire monitoring and extinguishing system based on smart cloud platform management according to claim 1 is characterized by: The calculation formula for calculating the excess coefficient of each environment type is: in, They are the excess coefficients of smoke, temperature, CO2, and VOC in the stable environmental data. They are the safety thresholds for smoke, temperature, CO2, and VOC. They are the concentrations of smoke, temperature, CO2, and VOC in the environmental data; The calculation formula for calculating the change trend coefficient of each environment type is: in, They are the change trend coefficients of smoke, temperature, CO2 and VOC respectively. is the data collection time interval; The characteristics of each environment type are calculated based on the exceeding coefficient, changing trend coefficient and exceeding duration, and the calculation formula for the environmental characteristic data set is: in, They are the characteristics of smoke, temperature, CO2, and VOC. They are the smoke exceeding standard coefficient, change trend coefficient, and weight of exceeding standard duration. They are the temperature exceeding standard coefficient, change trend coefficient, and weight of exceeding standard duration. They are the CO2 exceeding standard coefficient, changing trend coefficient, and weight of exceeding standard duration. They are the VOC exceeding standard coefficient, changing trend coefficient, and weight of exceeding standard duration. They are the duration of smoke, temperature, CO2 and VOC exceeding the standard. .

9. The IoT precise fire monitoring and extinguishing system based on smart cloud platform management according to claim 1 is characterized in that: The fire types include smoldering fire and open fire; The calculation formula for establishing the fire monitoring model based on Bayesian theorem is: According to the fire types in the historical data, the probability of each fire type occurring is calculated, that is, the prior probability of each fire type: in, For each fire type The prior probability of is the number of fires in historical data, is the indicative function, For the historical data Type of accident, , is the fire type in the historical data, Indicates a smoldering fire. Represents bright fire, is the smoothing coefficient; The formula of the indicative function is: ; Calculate the likelihood probability of the characterization factors under each fire type: in, For a given fire type Characteristic factors observed below The probability of For a given fire type Next, observe Characteristic Factors The probability of is the number of influencing factors, is a positive integer; The posterior probability of each fire type occurring is calculated according to Bayes' theorem: in, Fire type The posterior probability of is a normalization constant; Establish a decision model for fire type occurrence: in, The type of fire that occurred was finally identified; The logic for selecting a fire extinguishing method is as follows: When an open fire is identified, immediately open the perfluoroacetylene fire extinguishing device to extinguish the fire; When a smoldering fire is identified, immediately use the perfluoroacetylene fire extinguishing device intermittently to extinguish the fire.

Images