Fire control system and method for thermal power plant based on image processing

Abstract

The invention belongs to the technical field of fire-fighting systems, and particularly relates to a fire-fighting control system and method for a thermal power plant based on image processing, and the system comprises a central control room, a detection alarm system, and a fire-fighting equipment inspection system. The detection alarm system is composed of a fire detection system, an alarm and linkage control system, an emergency broadcast and telephone system and a fire-fighting area integrated cabinet. The alarm and linkage control system is internally provided with an information acquisition unit, an evacuation plan unit, an information sending unit, a navigation generation unit, a disaster area fire extinguishing / fire fighting facility starting unit and a blocking plan unit, and the evacuation plan unit is internally provided with an evacuation information generation unit, an evacuation route generation unit and an edge updating feedback unit. According to the invention, the problems of false alarm and missing alarm caused by equipment failure or aging can be reduced, a scheme for hindering fire spreading and an emergency operation scheme for emergency treatment are provided, and the emergency operation scheme can guarantee the passing of safety officers as much as possible.

Description

Technical Field

[0001] The invention belongs to the technical field of fire protection systems, and in particular relates to a fire protection control system and method for a thermal power plant based on image processing. Background Art

[0002] The fire protection configuration within a thermal power plant requires comprehensive consideration of multiple aspects and levels, including fire lanes, fire rescue facilities, automatic fire alarm systems, fire extinguishing facilities and other fire protection measures; these configurations should cooperate and work together to ensure that the power plant can quickly and effectively extinguish fires and evacuate personnel in the event of a fire.

[0003] Problems with existing technologies: High risk of failure of fire protection and alarm equipment: Due to the complex production environment of thermal power plants, fire protection and alarm equipment are exposed to high temperature, high pressure, dusty and other harsh environments for a long time, which makes them prone to failure. These failures may include false alarms or missed alarms of detectors, failure of fire extinguishing systems, etc., which seriously affect the reliability of fire protection systems. Incomplete emergency plans: Although fire control systems can prevent and respond to emergencies such as fires to a certain extent, a complete emergency plan is equally important; the emergency plans of some thermal power plants are not detailed enough or lack practical operability, resulting in the inability to quickly and effectively organize rescue and evacuation work in emergency situations. Specifically, there is no plan to prevent the spread of fire. In addition, it is impossible to intuitively and quickly understand the status information of important valves and equipment switches, and it is impossible to generate emergency operation plans. Summary of the invention

[0004] The purpose of the present invention is to provide a fire control system and method for a thermal power plant based on image processing, which can reduce the problems of false alarms and missed alarms caused by equipment failure or aging, and has a solution to hinder the spread of fire, as well as an emergency operation plan for emergency treatment, and the emergency operation plan can also ensure the passage of safety personnel as much as possible.

[0005] The technical solution adopted by the present invention is as follows: A fire control system for a thermal power plant based on image processing, comprising: Central control room, which consists of the fire alarm main control panel, CRT graphics station, broadcasting station and communication host; Detection alarm system, wherein the detection alarm system is composed of a fire detection system, an alarm and linkage control system, an emergency broadcast and telephone system, and a fire protection area integrated cabinet. The fire detection system is used to monitor the working status of all fire detection equipment in the factory area, and the fire detection equipment includes flame detectors, smoke detectors, temperature detectors, temperature sensing cables, and cameras. An intelligent alarm system is also provided inside the fire detection system. The emergency broadcast and telephone system is used to execute alarm commands and send alarm information through a broadcast station and a communication host. The fire protection area integrated cabinet is used to control the fire protection facilities in the area it is responsible for. The fire protection facilities include fire doors, gas fire extinguishing systems, and water fire protection systems; Gas fire extinguishing system, wherein the gas fire extinguishing system consists of an emergency start / stop button, an audible and visual alarm, a gas discharge indicator light and gas fire extinguishing facilities; Water fire fighting system, where the water fire fighting system consists of automatic water fire extinguishing facilities, fire hydrant system, foam system and water supply system; Firefighting facility inspection system, wherein the firefighting facility inspection system is composed of a fire extinguishing facility inspection and recording unit, an alarm facility inspection and recording unit, a fire door inspection and recording unit, a detection facility inspection record, and a fire passage inspection record. The firefighting facility inspection system is used to transmit the firefighting facility inspection data information to the intelligent alarm system, and the intelligent alarm system accurately processes the alarm commands through big data processing and analysis, artificial intelligence algorithms, signal processing and data analysis technology, machine learning algorithms, and time series analysis algorithms.

[0006] The big data processing and analysis is used to analyze and process massive data, and determine the authenticity and possible development trend of fire alarm signals through preset fire models and artificial intelligence algorithms; The artificial intelligence algorithm involves: Linear algebra: vector and matrix operations are the basis of artificial intelligence algorithms and are used to represent and process high-dimensional data; eigenvalues ​​and eigenvectors are mainly used in dimensionality reduction technology; matrix decomposition is mainly used in recommendation systems, signal processing, and machine learning model initialization; Probability theory and statistics: Probability distribution is used to model uncertainty; Bayes' theorem is used in Bayesian networks, probabilistic programming and probabilistic machine learning methods to update prior probabilities to obtain posterior probabilities; Maximum likelihood estimation and maximum a posteriori estimation are used for parameter estimation and model fitting; hypothesis testing is used to evaluate whether data conform to a certain hypothetical model; Optimization theory: Gradient descent, Newton method, and quasi-Newton method optimization algorithms are used to train neural networks and other machine learning models by adjusting model parameters to minimize the loss function; Calculus and gradient calculation: Differentiation and gradient calculation are the basis of the neural network back propagation algorithm, which is used to calculate the gradient of the loss function with respect to the model parameters and then update the model parameters; Information theory: information entropy, mutual information, and KL divergence are used to measure the amount of information, measure model uncertainty, and construct loss functions; Logistic regression: The hypothesis function is usually expressed as a linear equation, such as ,in is the parameter vector, is the input feature vector; the cost function (or loss function) is used to measure the quality of the model prediction. For logistic regression, the commonly used cost function is the negative value of the log-likelihood function; the cost function gradient is used to guide parameter iteration to minimize the cost function; Neural Networks: Artificial neurons are defined as functions , which contains the weight , Bias And the activation function ; The neural network is composed of multiple layers of artificial neurons, forming a complex function mapping relationship.

[0007] The time series analysis algorithm is used to predict future states and detect anomalies based on the time series data in the inspection records, which specifically involves: Moving Average (MA); Exponential smoothing; Autoregressive model (AR); Autoregressive Moving Average (ARIMA) model; Vector Autoregressive Moving Average Model (VARMA); Deep Learning Methods: Recurrent Neural Network (RNN): Suitable for predicting long-term dependencies; One-dimensional convolutional neural network (1D-CNN): extracts features through convolutional layers and is suitable for predicting short-term data; Transformer: Based on the attention mechanism, it can capture long-distance dependencies.

[0008] The alarm and linkage control system is internally provided with an information acquisition unit, an evacuation plan unit, an information sending unit, a navigation generation unit, a disaster area fire extinguishing / firefighting facility activation unit and a blocking plan unit. The disaster area fire extinguishing / firefighting facility activation unit activates the fire extinguishing and firefighting facilities in the fire area by controlling the corresponding fire area integrated cabinet.

[0009] The information acquisition unit is data-connected to the factory personnel information big data platform, and the information acquisition unit is internally provided with a factory personnel information linking unit, a safety officer information linking unit and a positioning tracking unit; Among them, the factory personnel information link unit is mainly used to obtain the communication information of factory employees, the safety officer information link unit is mainly used to obtain the communication information of safety officers. The safety officers are evaluated by the factory manager and are mainly responsible for fire protection matters in the designated area. The positioning and tracking unit is used to track and locate the location information of the safety officers.

[0010] The evacuation plan unit is internally provided with an evacuation information generation unit, an evacuation route generation unit and an edge update feedback unit; The evacuation information generation unit is used to generate evacuation information, the evacuation route generation unit is used to generate an evacuation route, and cooperates with the navigation generation unit to generate an evacuation navigation interface. The edge update feedback unit determines the fire area according to the status of the detection equipment, and at the same time feeds back the fire area change information to the evacuation route generation unit.

[0011] The blocking plan unit is internally provided with a disaster area edge blocking unit, a valve / switch blocking unit, a blocking edge updating unit, a valve / switch status information unit, a blocking route generating unit and a fire extinguishing / fire fighting facility intermittent unit; The disaster area edge blocking unit starts the fire extinguishing and fire fighting facilities near the edge of the fire area by controlling the corresponding fire area integrated cabinet. The valve / switch blocking unit is used to close the dangerous pipeline valves and dangerous equipment switches in and around the fire area. The blocking edge update unit starts the fire extinguishing and fire fighting facilities outside the edge of the new fire area according to the fire area change information and in cooperation with the disaster area edge blocking unit. The valve / switch status information unit is used to obtain the information of unclosed valves and equipment switches near the fire area. The blocking route generation unit is used to generate safe route information for closing the above valves and equipment switches. The fire extinguishing / firefighting facility intermittent unit is used to intermittently close the fire extinguishing and firefighting facilities in the direction of the above safe route.

[0012] The blocking route generation unit and the evacuation route generation unit both utilize path planning algorithms, obstacle avoidance algorithms, potential field methods, multi-objective optimization, and real-time dynamic planning to generate safe routes that are as short as possible and as far away from the fire area as possible; The path planning algorithm includes the Dijkstra algorithm, which finds the final shortest path by continuously updating the shortest path from the starting point to each vertex; Obstacle avoidance algorithm: Based on path planning, an obstacle avoidance mechanism needs to be added to ensure that the route is away from dangerous areas; Potential field method: Potential field method is a method that simulates physical force fields, in which the target position produces attractive forces and obstacles produce repulsive forces. By calculating the resultant of these forces, a smooth path can be provided for a robot or other navigation system. Real-time dynamic planning: The fire area is changing dynamically, and the real-time dynamic planning technology is used to update the path planning in real time.

[0013] A fire control method for a thermal power plant based on image processing, the specific steps are as follows: Step 1: Find out where the fire is and where it is located, and notify the fire brigade; Step 2: Close the fire doors, valves of dangerous pipes and switches of dangerous equipment in and around the fire area; Step 3: Issue a fire command to the fire area integrated cabinet in the designated area and activate the fire extinguishing facilities through it; Step 4: Activate fire-fighting facilities near the fire area to generate the first blocking layer; 4.1): According to the status of the detection equipment, the spread of the fire is detected and it is determined that the first blocking layer has been breached by the fire; 4.2): Start the fire-fighting facilities near the newly generated fire area to generate the second blocking layer; 4.3): Repeat the above operations to continuously generate a blocking layer outside the fire area; Step 5: The factory alarm device sounds an alarm, and sends alarm and evacuation information to the factory staff communication devices through the communication host; 5.1): Generate fire alarm information and generate evacuation route plan based on personnel location information; 5.2): Send to the communication devices of factory employees via SMS / email; 5.3): Generate evacuation route navigation interface; Step 6: Send emergency blocking operation information to the safety officer's communication equipment to assist the safety officer in emergency handling.

[0014] In step 6, the specific process of emergency handling is as follows: Step 6.1: Determine the positions D1, D2, etc. of the valves / switches that are not closed; Step 6.2: Determine the location information of the safety officers near the locations D1, D2, etc.; Step 6.3: Generate a route for guiding the closing of valves / switches and send it to the communication device of the safety officer, and generate a blocking route navigation interface at the same time; Step 6.4: According to the location information of the safety officer, temporarily open the fire door near him and close the fire extinguishing facilities near him to allow the safety officer to pass through; Step 6.5: After the safety officer has passed through, close the temporarily opened fire door and restart the fire extinguishing facilities; Step 6.6: After confirming that the safety officer has closed the valve / switch, the best evacuation route information is drawn up based on the fire spread information and sent to the safety officer's communication equipment, while tracking the blocking layer generation information; Step 6.7: Temporarily open the fire door on the safety officer's evacuation route, temporarily close the fire extinguishing facilities on the evacuation route, and generate the evacuation route navigation interface; Step 6.8: After the safety officer evacuates, close the corresponding fire doors and restart the fire extinguishing facilities.

[0015] The technical effects achieved by the present invention are: The present invention can ensure that the equipment is in good working condition by regularly inspecting the fire alarm equipment and fire-fighting facilities. By comprehensively managing and statistically analyzing the collected data, it helps managers to better understand the operating status of the equipment, discover problems in a timely manner and deal with them, thereby reducing false alarms and missed alarms caused by equipment failure or aging, enhancing emergency response capabilities, and thus improving the overall fire safety level.

[0016] In the present invention, when a fire occurs, the fire detection system detects the occurrence, location and scope of the fire, and the evacuation information generation unit edits such information into text messages, and uses the communication host and the broadcasting station to send alarm and evacuation information to the communication equipment of the factory employees, so as to quickly notify the factory employees and facilitate rapid evacuation.

[0017] The present invention activates fire extinguishing and firefighting facilities near the edge of the fire area according to the location and scope of the fire, and implements fire extinguishing operations at the edge of the fire area in advance, thereby reducing the temperature of edge objects in advance and generating a defense line to block the spread of fire. In addition, when the fire extinguishing facilities at the edge of the fire area are activated for the first time, a first blocking layer is generated. When it is determined that the first blocking layer has been broken through by the fire, the fire extinguishing facilities near the newly generated fire area are activated to generate a second blocking layer. This process is repeated to continuously generate blocking layers outside the fire area to continuously hinder the spread of fire.

[0018] The present invention determines the position of valves and equipment switches that are not closed near the fire area, and compiles the information into information and sends it to the communication equipment of nearby safety officers, thereby issuing an alarm message that there are still valves and equipment switches that are not closed, and at the same time generates a route and a navigation interface for guiding the closing of valves / switches. When the safety officer moves along the route, the fire doors near the officer are intermittently opened according to the change of his position, and the fire extinguishing facilities near the officer are closed, so as to facilitate the safety officer to pass quickly and avoid the safety officer's line of sight being obstructed by fire water. After the safety officer passes through, the temporarily opened fire doors are closed and the fire extinguishing facilities are restarted.

[0019] In the present invention, the route generated by the blocking route generation unit can be as short as possible and as far away from the fire area as possible, and the route can be flexibly changed according to the changes in the fire area. In addition, after determining that the safety officer closes the valve / switch, the optimal evacuation route information is formulated according to the fire spread information. The evacuation route can also be flexibly changed according to the changes in the fire area. In addition, the fire door on the safety officer's evacuation route is temporarily opened, and the fire extinguishing facilities on the evacuation route are temporarily closed for the safety officer to evacuate. After the safety officer evacuates, the corresponding fire door is closed and the fire extinguishing facilities are restarted. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is a system diagram of a fire control system provided by an embodiment of the present invention; Figure 2 is a system diagram of an alarm and linkage control system provided by an embodiment of the present invention; Figure 3 It is a factory employee evacuation and safety officer blocking route diagram provided by an embodiment of the present invention; Figure 4 It is a schematic diagram of changes in fire-fighting facilities when a safety officer blocks an emergency operation provided by an embodiment of the present invention; Figure 5 is a schematic diagram of changes in the fire area provided by an embodiment of the present invention; Figure 6 is a flow chart of a fire control method provided by an embodiment of the present invention; Figure 7 is a flow chart of emergency handling by a safety officer provided by an embodiment of the present invention; Figure 8 It is a data processing schematic diagram of the intelligent alarm system provided by an embodiment of the present invention.

[0021] In the accompanying drawings, the components represented by the reference numerals are listed as follows: 1. Central control room; 101. Fire alarm main control panel; 102. CRT graphics station; 103. Broadcasting station; 104. Communication host; 2. Detection alarm system; 201. Fire detection system; 202. Alarm and linkage control system; 203. Emergency broadcast and telephone system; 204. Fire area integrated cabinet; 205. Information acquisition unit; 206. Plant personnel information link unit; 207. Safety officer information link unit; 208. Positioning and tracking unit; 209. Evacuation plan unit; 210. Evacuation information generation unit; 211. Evacuation route generation unit; 212. Edge update feedback unit; 213. Information sending unit; 214. Navigation generation unit; 215. Disaster area fire extinguishing / firefighting facility activation unit; 216. Blocking plan unit; 217. Disaster area edge blocking unit; 218. Valve / switch blocking unit; 219. Blocking edge update unit; 220. Valve / switch status information unit; 221. Blocking route generation unit; 222. Fire extinguishing / firefighting facility intermittent unit; 3. Gas fire extinguishing system; 301. Emergency start / stop button; 302. Sound and light alarm; 303. Deflation indicator light; 304. Gas fire extinguishing facility; 4. Water fire fighting system; 401. Automatic water fire extinguishing facility; 402. Fire hydrant system; 403. Foam system; 404. Water supply system; 5. Fire fighting facility inspection system; 501. Fire extinguishing facility inspection unit; 502. Alarm facility inspection unit; 503. Fire door inspection unit; 504. Detection facility inspection record; 505. Fire passage inspection record. DETAILED DESCRIPTION

[0022] In order to make the purpose and advantages of the present invention more clearly understood, the present invention is specifically described below in conjunction with embodiments. It should be understood that the following text is only used to describe one or several specific embodiments of the present invention, and does not strictly limit the scope of protection of the specific claims of the present invention.

[0023] like Figure 1-Figure 2 As shown, a fire control system for a thermal power plant based on image processing comprises: The central control room 1 is composed of a fire alarm main control panel 101, a CRT graphic station 102, a broadcasting station 103 and a communication host 104.

[0024] Detection alarm system 2, wherein the detection alarm system 2 is composed of a fire detection system 201, an alarm and linkage control system 202, an emergency broadcast and telephone system 203 and a fire area integrated cabinet 204; the emergency broadcast and telephone system 203 is used to execute the alarm command and send the alarm information through the broadcast station 103 and the communication host 104. The emergency broadcast and telephone system 203 can also promptly notify the nearest fire brigade when a fire occurs. The fire area integrated cabinet 204 is used to control the fire-fighting facilities in its responsible area, and the fire-fighting facilities include fire doors, a gas fire extinguishing system 3 and a water fire-fighting system 4.

[0025] The gas fire extinguishing system 3 is composed of an emergency start / stop button 301 , an audible and visual alarm 302 , a deflation indicator light 303 and a gas fire extinguishing facility 304 .

[0026] The water fire fighting system 4 is composed of an automatic water fire extinguishing facility 401 , a fire hydrant system 402 , a foam system 403 and a water supply system 404 .

[0027] The firefighting facility inspection system 5, wherein the firefighting facility inspection system 5 is composed of a fire extinguishing facility inspection and recording unit 501, an alarm facility inspection and recording unit 502, a fire door inspection and recording unit 503, a detection facility inspection record 504 and a fire passage inspection record 505. The firefighting facility inspection system 5 is used to input the inspection data of various firefighting facilities, including fire extinguishing facilities, alarm facilities, fire doors, detection facilities and fire passages.

[0028] Embodiment 1: Refer to the attached Figure 1 and Figure 8 The fire detection system 201 is used to monitor the working status of all fire detection equipment in the factory, and the fire detection equipment includes flame detectors, smoke detectors, temperature detectors, temperature sensing cables and cameras. An intelligent alarm system is also set up inside the fire detection system 201.

[0029] The firefighting facility inspection system 5 is used to transmit the firefighting facility inspection data information to the intelligent alarm system, and the intelligent alarm system accurately processes the alarm commands through big data processing and analysis, artificial intelligence algorithms, signal processing and data analysis technology, machine learning algorithms and time series analysis algorithms.

[0030] Artificial intelligence algorithms involve: Linear algebra: vector and matrix operations are the basis of artificial intelligence algorithms and are used to represent and process high-dimensional data. For example, in neural networks, weights and inputs and outputs are usually represented as vectors or matrices. Eigenvalues ​​and eigenvectors are mainly used in dimensionality reduction techniques (such as principal component analysis PCA). Matrix decomposition (such as singular value decomposition SVD, eigendecomposition) is mainly used in recommendation systems, signal processing, and machine learning model initialization. Probability theory and statistics: Probability distributions (such as Gaussian distribution, Bernoulli distribution, and multinomial distribution) are used to model uncertainty; Bayes' theorem is used in probabilistic machine learning methods such as Bayesian networks and probabilistic programming to update prior probabilities to obtain posterior probabilities; maximum likelihood estimation and maximum a posteriori estimation are used for parameter estimation and model fitting; hypothesis testing is used to evaluate whether data conform to a certain hypothetical model; Optimization theory: Gradient descent, Newton method, and quasi-Newton method optimization algorithms are used to train neural networks and other machine learning models by adjusting model parameters to minimize the loss function; Calculus and gradient calculation: Differentiation and gradient calculation are the basis of the neural network back propagation algorithm, which is used to calculate the gradient of the loss function with respect to the model parameters and then update the model parameters; Information theory: information entropy, mutual information, and KL divergence are used to measure the amount of information, measure model uncertainty, and construct loss functions; Logistic regression: The hypothesis function is usually expressed as a linear equation, such as ,in is the parameter vector, is the input feature vector; the cost function (or loss function) is used to measure the quality of the model prediction. For logistic regression, the commonly used cost function is the negative value of the log-likelihood function; the cost function gradient is used to guide parameter iteration to minimize the cost function; Neural Networks: Artificial neurons are defined as functions , which contains the weight , Bias And the activation function ; The neural network is composed of multiple layers of artificial neurons, forming a complex function mapping relationship.

[0031] The time series analysis algorithm is used to predict future states and detect anomalies based on the time series data in the inspection records. Specifically, it involves: 1) Moving Average (MA) Simple moving average method: take the average of the data of the past several periods. in, It's at the time Calculate the moving average, is the first data points, is the length of the moving average; Weighted moving average method: different weights are assigned to data of different periods. in, It's at the time The weighted moving average calculated, It is The weight of the period, is the first data points.

[0032] 2) Exponential smoothing method Single exponential smoothing: Applicable to univariate time series without obvious trend or seasonality. in, It's time point The smoothed value of is the smoothing coefficient , It's time point The actual observed value of Holt-Winters method: Applicable to data with trend and seasonality. in, The time points The horizontal component, trend component and seasonal component of are three smoothing parameters, is the season length.

[0033] 3) Autoregressive model (AR) Autoregressive model: Use the observed values ​​of the previous periods to predict future values, which is suitable for stationary time series. in, It's at the time The predicted data points, are the parameters of the autoregressive model, is the previous data points, It is a random error.

[0034] 4) Autoregressive Moving Average Model (ARIMA) ARIMA model: It combines autoregression (AR), difference (I) and moving average (MA) methods and is applied to the modeling of non-stationary time series. in, is the backshift operator (i.e. ), are the parameters of the autoregressive part, are the parameters of the moving average part, is the number of differences, is a constant term, is a white noise process.

[0035] 5) Vector Autoregressive Moving Average Model (VARMA) VARMA model: used for the prediction of multivariate time series, taking into account the mutual influence between multiple variables.

[0036] 6) Deep Learning Methods Recurrent Neural Network (RNN): Suitable for predicting long-term dependencies. The mathematical expression of RNN involves weight matrix, bias vector and activation function, etc. The specific form depends on the network structure. One-dimensional convolutional neural network (1D-CNN): extracts features through convolutional layers and is suitable for predicting short-term data. The mathematical expression of 1D-CNN includes the setting of parameters such as convolution kernel, step size, and padding. Transformer: Based on the attention mechanism, it can capture long-distance dependencies. The mathematical expression of Transformer involves mathematical formulas of components such as self-attention mechanism, encoder and decoder.

[0037] The working principle of the present invention is as follows: big data processing and analysis are used to analyze and process massive data, and the authenticity and possible development trend of fire alarm signals are judged through preset fire models and artificial intelligence algorithms. Signal processing and data analysis technology pre-processes the raw data collected by sensors, including filtering, denoising, etc., to improve the quality of data. Feature extraction is to extract meaningful features from the pre-processed data so that subsequent algorithms can process them. Machine learning algorithms can identify normal and abnormal states, and use clustering and other algorithms to discover patterns and abnormal points in inspection data. Time series analysis algorithms are used to predict future states and detect abnormalities based on time series data in inspection records. In the above process, by regularly inspecting fire alarm equipment and fire-fighting facilities, it can ensure that the equipment is in good working condition. By comprehensively managing and statistically analyzing the collected data, this helps managers better understand the operating status of the equipment, discover problems in a timely manner and deal with them, thereby reducing false alarms and missed alarms caused by equipment failure or aging, enhancing emergency response capabilities, and thus improving the overall fire safety level.

[0038] Embodiment 2: See attached Figure 2 The alarm and linkage control system 202 is internally provided with an information acquisition unit 205, an evacuation plan unit 209, an information sending unit 213, a navigation generation unit 214, a disaster area fire extinguishing / firefighting facility activation unit 215 and a blocking plan unit 216. The disaster area fire extinguishing / firefighting facility activation unit 215 activates the fire extinguishing and firefighting facilities in the fire area by controlling the corresponding fire area integrated cabinet 204.

[0039] The information acquisition unit 205 is data-connected with the plant personnel information big data platform, and the information acquisition unit 205 is internally provided with a plant personnel information linking unit 206, a safety officer information linking unit 207 and a positioning tracking unit 208; The factory personnel information link unit 206 is mainly used to obtain the communication information of factory employees, the safety officer information link unit 207 is mainly used to obtain the communication information of safety officers. Safety officers are evaluated by the factory manager and are mainly responsible for fire protection matters in designated areas. The positioning and tracking unit 208 is used to track and locate the location information of safety officers.

[0040] See attached Figure 2 The evacuation plan unit 209 is internally provided with an evacuation information generation unit 210, an evacuation route generation unit 211 and an edge update feedback unit 212; The evacuation information generation unit 210 is used to generate evacuation information, the evacuation route generation unit 211 is used to generate an evacuation route, and cooperates with the navigation generation unit 214 to generate an evacuation navigation interface, and the edge update feedback unit 212 determines the fire area according to the status of the detection equipment, and at the same time feeds back the fire area change information to the evacuation route generation unit 211.

[0041] The working principle of the present invention is as follows: when a fire occurs, the fire detection system 201 detects the occurrence, location and scope of the fire, and the evacuation information generation unit 210 edits such information into text messages, and uses the communication host 104 and the broadcasting station 103 to send alarm and evacuation information to the factory employees' communication equipment, so as to quickly notify the factory employees and facilitate rapid evacuation.

[0042] See attached Figure 2 The blocking plan unit 216 is internally provided with a disaster area edge blocking unit 217, a valve / switch blocking unit 218, a blocking edge updating unit 219, a valve / switch state information unit 220, a blocking route generating unit 221 and a fire extinguishing / fire fighting facility intermittent unit 222; The disaster area edge blocking unit 217 starts the fire extinguishing and fire fighting facilities near the edge of the fire area by controlling the corresponding fire area integrated cabinet 204; the valve / switch blocking unit 218 is used to close the dangerous pipeline valves and dangerous equipment switches in and around the fire area; the blocking edge update unit 219 starts the fire extinguishing and fire fighting facilities outside the edge of the new fire area based on the fire area change information and in cooperation with the disaster area edge blocking unit 217.

[0043] The working principle of the present invention is as follows: according to the location and scope of the fire, the disaster area fire extinguishing / firefighting facilities activation unit 215 activates the fire extinguishing and firefighting facilities in the fire area by controlling the corresponding fire area integrated cabinet 204, and implements fire extinguishing operations in the fire area. In addition, the disaster area edge blocking unit 217 activates the fire extinguishing and firefighting facilities near the edge of the fire area by controlling the corresponding fire area integrated cabinet 204, and implements fire extinguishing operations at the edge of the fire area in advance, thereby reducing the temperature of the edge objects in advance and generating a defense line to block the spread of the fire. In addition, when the fire extinguishing facilities at the edge of the fire area are activated for the first time, a first blocking layer is generated; Figure 5 As shown, according to the status of the detection equipment, the spread of the fire is detected. When it is determined that the first blocking layer is broken by the fire, the fire extinguishing facilities near the newly generated fire area are activated to generate a second blocking layer. This cycle can continuously generate blocking layers outside the fire area to continuously hinder the spread of the fire.

[0044] The valve / switch status information unit 220 is used to obtain the information of valves and equipment switches that are not closed near the fire area, the blocking route generation unit 221 is used to generate safe route information for closing the above-mentioned valves and equipment switches, and the fire extinguishing / firefighting facility intermittent unit 222 is used to intermittently close the fire extinguishing and firefighting facilities in the direction of the above-mentioned safe route.

[0045] The working principle of the present invention is: through the information of valves and equipment switches that are not closed near the fire area, their positions are determined, and the information is edited and sent to the communication equipment of the nearby safety officers, so as to issue an alarm message that there are still valves and equipment switches that are not closed, and at the same time generate a route and a navigation interface to guide the closing of valves / switches. When the safety officer moves along the route, the fire doors near him are intermittently opened according to his position changes, and the fire extinguishing facilities near him are closed, so as to facilitate the safety officer to pass quickly and avoid fire water and other obstructions to the safety officer's line of sight. After the safety officer passes through, the temporarily opened fire doors are closed and the fire extinguishing facilities are restarted.

[0046] The blocking route generation unit 221 and the evacuation route generation unit 211 both utilize path planning algorithms, obstacle avoidance algorithms, potential field methods, multi-objective optimization, and real-time dynamic planning to generate safe routes that are as short as possible and as far away from the fire area as possible; The path planning algorithm includes the Dijkstra algorithm, which finds the final shortest path by continuously updating the shortest path from the starting point to each vertex; Its mathematical expression mainly involves the updating of distance and the determination of the shortest path: Initialization: For each vertex in the graph ,set up (indicates that the initial distance from the starting point to the vertex is infinite), except for the starting point itself, let ; Distance update: for the current vertex , traverse all its adjacent vertices ,if , then update ,in represents the edge weight from vertex u to vertex v; Select the next vertex: Select the vertex with the smallest distance from the unprocessed vertices as the next vertex to be processed; Repeat: Repeat the above steps of distance update and selecting the next vertex until all vertices are processed or the target vertex is visited.

[0047] Obstacle avoidance algorithm: Based on path planning, an obstacle avoidance mechanism needs to be added to ensure that the route is away from dangerous areas; this can be achieved by increasing the weight of dangerous areas so that the algorithm tends to choose paths that avoid these areas; for example, in the Dijkstra algorithm, the weight of the edge represents the cost of moving from one grid to another.

[0048] Potential field method: Potential field method is a method to simulate physical force field, in which the target position produces attraction and obstacles produce repulsion. By calculating the resultant force of these forces, a smooth path can be provided for a robot or other navigation system; it can be expressed as the resultant force on the robot ,in is the gravitational force generated by the target point, are the repulsive forces exerted by obstacles; these forces are usually calculated based on the distance between the robot, the target point, and the obstacle.

[0049] Real-time dynamic planning: The fire area is dynamically changing, and the real-time dynamic planning technology is used to update the path planning in real time; this can be achieved through techniques such as rolling horizon control (RCPC) or dynamic window method (DWA).

[0050] The working principle of the present invention is: the route generated by the blocking route generation unit 221 can be as short as possible and as far away from the fire area as possible under the above-mentioned technology, and the route can be flexibly changed according to the changes in the fire area. In addition, after determining that the safety officer has closed the valve / switch, the best evacuation route information is formulated according to the fire spread information and sent to the safety officer's communication equipment. The evacuation route can also be flexibly changed according to the changes in the fire area. In addition, the fire door on the safety officer's evacuation route is temporarily opened, and the fire extinguishing facilities on the evacuation route are temporarily closed for the safety officer to evacuate. After the safety officer has evacuated, the corresponding fire door is closed and the fire extinguishing facilities are restarted.

[0051] like Figure 6 As shown, a fire control method for a thermal power plant based on image processing, the specific steps are as follows: Step 1: Find out where the fire is and where it is located, and notify the fire brigade; Step 2: Close the fire doors, valves of dangerous pipes and switches of dangerous equipment in and around the fire area; Step 3: Send a fire command to the fire area integrated cabinet 204 of the designated area, and activate the fire extinguishing facilities through it; Step 4: Activate fire-fighting facilities near the fire area to generate the first blocking layer; 4.1): According to the status of the detection equipment, the spread of the fire is detected and it is determined that the first blocking layer has been breached by the fire; 4.2): Start the fire-fighting facilities near the newly generated fire area to generate the second blocking layer; 4.3): Repeat the above operations to continuously generate a blocking layer outside the fire area; Step 5: The factory alarm device sounds an alarm, and sends an alarm and evacuation message to the factory staff communication devices through the communication host 104; 5.1): Generate fire alarm information and generate evacuation route plan based on personnel location information; 5.2): Send to the communication devices of factory employees via SMS / email; 5.3): Generate evacuation route navigation interface; Step 6: Send emergency blocking operation information to the safety officer's communication equipment to assist the safety officer in emergency handling.

[0052] like Figure 7 As shown, in step 6, the specific process of emergency handling is as follows: Step 6.1: Determine the positions D1, D2, etc. of the valves / switches that are not closed; Step 6.2: Determine the location information of the safety officers near the locations D1, D2, etc.; Step 6.3: Generate a route for guiding the closing of valves / switches and send it to the communication device of the safety officer, and generate a blocking route navigation interface at the same time; Step 6.4: According to the location information of the safety officer, temporarily open the fire door near him and close the fire extinguishing facilities near him to allow the safety officer to pass through; Step 6.5: After the safety officer has passed through, close the temporarily opened fire door and restart the fire extinguishing facilities; Step 6.6: After confirming that the safety officer has closed the valve / switch, the best evacuation route information is drawn up based on the fire spread information and sent to the safety officer's communication equipment, while tracking the blocking layer generation information; Step 6.7: Temporarily open the fire door on the safety officer's evacuation route, temporarily close the fire extinguishing facilities on the evacuation route, and generate the evacuation route navigation interface; Step 6.8: After the safety officer evacuates, close the corresponding fire doors and restart the fire extinguishing facilities.

[0053] The above is only a preferred embodiment of the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made without departing from the principles of the present invention, and these improvements and modifications should also be considered as the protection scope of the present invention. The structures, devices and operating methods not specifically described and explained in the present invention shall be implemented according to the conventional means in the art unless otherwise specified and limited.

Claims

1. A fire control system for a thermal power plant based on image processing, characterized in that: include: A central control room (1), wherein the central control room (1) is composed of a fire alarm main control panel (101), a CRT graphics station (102), a broadcasting station (103) and a communication host (104); A detection alarm system (2), wherein the detection alarm system (2) is composed of a fire detection system (201), an alarm and linkage control system (202), an emergency broadcast and telephone system (203) and a fire protection area integrated cabinet (204); the fire detection system (201) is used to monitor the working status of all fire detection equipment in the factory area, and the fire detection equipment includes a flame detector, a smoke detector, a temperature detector, a temperature sensing cable and a camera; an intelligent alarm system is also provided inside the fire detection system (201); the emergency broadcast and telephone system (203) is used to execute an alarm command and send alarm information through a broadcast station (103) and a communication host (104); the fire protection area integrated cabinet (204) is used to control the fire protection facilities in the area under its responsibility, and the fire protection facilities include fire doors, a gas fire extinguishing system (3) and a water fire protection system (4); A gas fire extinguishing system (3), wherein the gas fire extinguishing system (3) is composed of an emergency start / stop button (301), an audible and visual alarm (302), a gas deflation indicator light (303), and a gas fire extinguishing facility (304); A water fire fighting system (4), wherein the water fire fighting system (4) is composed of an automatic water fire extinguishing facility (401), a fire hydrant system (402), a foam system (403) and a water supply system (404); A firefighting facility inspection system (5), wherein the firefighting facility inspection system (5) is composed of a fire extinguishing facility inspection unit (501), an alarm facility inspection unit (502), a fire door inspection unit (503), a detection facility inspection record (504), and a fire passage inspection record (505); the firefighting facility inspection system (5) is used to transmit firefighting facility inspection data information to an intelligent alarm system, and the intelligent alarm system accurately processes alarm commands through big data processing and analysis, artificial intelligence algorithms, signal processing and data analysis technology, machine learning algorithms, and time series analysis algorithms.

2. The fire control system for a thermal power plant based on image processing according to claim 1, characterized in that: The big data processing and analysis is used to analyze and process massive data, and determine the authenticity and possible development trend of fire alarm signals through preset fire models and artificial intelligence algorithms; The artificial intelligence algorithm involves: Linear algebra: vector and matrix operations are the basis of artificial intelligence algorithms and are used to represent and process high-dimensional data; eigenvalues ​​and eigenvectors are mainly used in dimensionality reduction technology; matrix decomposition is mainly used in recommendation systems, signal processing, and machine learning model initialization; Probability theory and statistics: Probability distribution is used to model uncertainty; Bayes' theorem is used in Bayesian networks, probabilistic programming and probabilistic machine learning methods to update prior probabilities to obtain posterior probabilities; Maximum likelihood estimation and maximum a posteriori estimation are used for parameter estimation and model fitting; hypothesis testing is used to evaluate whether data conform to a certain hypothetical model; Optimization theory: Gradient descent, Newton method, and quasi-Newton method optimization algorithms are used to train neural networks and other machine learning models by adjusting model parameters to minimize the loss function; Calculus and gradient calculation: Differentiation and gradient calculation are the basis of the neural network back propagation algorithm, which is used to calculate the gradient of the loss function with respect to the model parameters and then update the model parameters; Information theory: information entropy, mutual information, and KL divergence are used to measure the amount of information, measure model uncertainty, and construct loss functions; Logistic regression: The hypothesis function is usually expressed as a linear equation, such as ,in is the parameter vector, is the input feature vector; the cost function (or loss function) is used to measure the quality of the model prediction. For logistic regression, the commonly used cost function is the negative value of the log-likelihood function; the cost function gradient is used to guide parameter iteration to minimize the cost function; Neural Networks: Artificial neurons are defined as functions , which contains the weight , Bias And the activation function ; The neural network is composed of multiple layers of artificial neurons, forming a complex function mapping relationship.

3. The fire control system for a thermal power plant based on image processing according to claim 1, characterized in that: The time series analysis algorithm is used to predict future states and detect anomalies based on the time series data in the inspection records, which specifically involves: Moving Average (MA); Exponential smoothing; Autoregressive model (AR); Autoregressive Moving Average (ARIMA) model; Vector Autoregressive Moving Average Model (VARMA); Deep Learning Methods: Recurrent Neural Network (RNN): Suitable for predicting long-term dependencies; One-dimensional convolutional neural network (1D-CNN): extracts features through convolutional layers and is suitable for predicting short-term data; Transformer: Based on the attention mechanism, it can capture long-distance dependencies.

4. The fire control system for a thermal power plant based on image processing according to claim 1, characterized in that: The alarm and linkage control system (202) is internally provided with an information acquisition unit (205), an evacuation plan unit (209), an information sending unit (213), a navigation generation unit (214), a disaster area fire extinguishing / firefighting facility activation unit (215), and a blocking plan unit (216). The disaster area fire extinguishing / firefighting facility activation unit (215) activates the fire extinguishing and firefighting facilities in the fire area by controlling the corresponding fire area integrated cabinet (204).

5. The fire control system for a thermal power plant based on image processing according to claim 4 is characterized in that: The information acquisition unit (205) is data-connected to the plant personnel information big data platform, and the information acquisition unit (205) is internally provided with a plant personnel information linking unit (206), a safety officer information linking unit (207) and a positioning tracking unit (208); The factory personnel information link unit (206) is mainly used to obtain the communication information of factory employees, the safety officer information link unit (207) is mainly used to obtain the communication information of the safety officer. The safety officer is evaluated by the factory manager and is mainly responsible for fire protection matters in the designated area. The positioning tracking unit (208) is used to track and locate the location information of the safety officer.

6. The fire control system for a thermal power plant based on image processing according to claim 4, characterized in that: The evacuation plan unit (209) is internally provided with an evacuation information generation unit (210), an evacuation route generation unit (211) and an edge update feedback unit (212); The evacuation information generation unit (210) is used to generate evacuation information, the evacuation route generation unit (211) is used to generate an evacuation route, and cooperates with the navigation generation unit (214) to generate an evacuation navigation interface, and the edge update feedback unit (212) determines the fire area according to the state of the detection equipment, and simultaneously feeds back the fire area change information to the evacuation route generation unit (211).

7. The fire control system for a thermal power plant based on image processing according to claim 4, characterized in that: The blocking plan unit (216) is internally provided with a disaster area edge blocking unit (217), a valve / switch blocking unit (218), a blocking edge updating unit (219), a valve / switch state information unit (220), a blocking route generating unit (221) and a fire extinguishing / fire fighting facility intermittent unit (222); The disaster area edge blocking unit (217) activates the fire extinguishing and fire fighting facilities near the edge of the fire area by controlling the corresponding fire area integrated cabinet (204); the valve / switch blocking unit (218) is used to close the dangerous pipeline valves and dangerous equipment switches in and around the fire area; the blocking edge updating unit (219) activates the fire extinguishing and fire fighting facilities outside the edge of the new fire area according to the fire area change information and in cooperation with the disaster area edge blocking unit (217); The valve / switch status information unit (220) is used to obtain information about valves and equipment switches that are not closed near the fire area, the blocking route generation unit (221) is used to generate safe route information for closing the above-mentioned valves and equipment switches, and the fire extinguishing / firefighting facility intermittent unit (222) is used to intermittently close the fire extinguishing and firefighting facilities in the direction of the above-mentioned safe route.

8. The fire control system for a thermal power plant based on image processing according to claim 7 is characterized in that: The blocking route generation unit (221) and the evacuation route generation unit (211) both utilize a path planning algorithm, an obstacle avoidance algorithm, a potential field method, multi-objective optimization, and real-time dynamic planning to generate a safe route that is as short as possible and as far away from the fire area as possible; The path planning algorithm includes the Dijkstra algorithm, which finds the final shortest path by continuously updating the shortest path from the starting point to each vertex; Obstacle avoidance algorithm: Based on path planning, an obstacle avoidance mechanism needs to be added to ensure that the route is away from dangerous areas; Potential field method: Potential field method is a method that simulates physical force fields, in which the target position produces attractive forces and obstacles produce repulsive forces. By calculating the resultant of these forces, a smooth path can be provided for a robot or other navigation system. Real-time dynamic planning: The fire area is changing dynamically, and the real-time dynamic planning technology is used to update the path planning in real time.

9. A fire control method for a thermal power plant based on image processing, used for using the fire control system for a thermal power plant based on image processing according to any one of claims 1 to 8, characterized in that: The specific steps are as follows: Step 1: Find out where the fire is and where it is located, and notify the fire brigade; Step 2: Close the fire doors, valves of dangerous pipes and switches of dangerous equipment in and around the fire area; Step 3: Sending a fire command to the fire area integrated cabinet (204) of the designated area, and activating the fire extinguishing facilities therethrough; Step 4: Activate fire-fighting facilities near the fire area to generate the first blocking layer; 4.1): According to the status of the detection equipment, the spread of the fire is detected and it is determined that the first blocking layer has been breached by the fire; 4.2): Start the fire-fighting facilities near the newly generated fire area to generate the second blocking layer; 4.3): Repeat the above operations to continuously generate a blocking layer outside the fire area; Step 5: The factory alarm device sounds an alarm, and sends an alarm and evacuation message to the factory staff communication devices via the communication host (104); 5.1): Generate fire alarm information and generate evacuation route plan based on personnel location information; 5.2): Send to the communication devices of factory employees via SMS / email; 5.3): Generate evacuation route navigation interface; Step 6: Send emergency blocking operation information to the safety officer's communication equipment to assist the safety officer in emergency handling.

10. The fire control method for a thermal power plant based on image processing according to claim 9, characterized in that: In step 6, the specific process of emergency handling is as follows: Step 6.1: Determine the positions D1, D2, etc. of the valves / switches that are not closed; Step 6.2: Determine the location information of the safety officers near the locations D1, D2, etc.; Step 6.3: Generate a route for guiding the closing of valves / switches and send it to the communication device of the safety officer, and generate a blocking route navigation interface at the same time; Step 6.4: According to the location information of the safety officer, temporarily open the fire door near him and close the fire extinguishing facilities near him to allow the safety officer to pass through; Step 6.5: After the safety officer has passed through, close the temporarily opened fire door and restart the fire extinguishing facilities; Step 6.6: After confirming that the safety officer has closed the valve / switch, the best evacuation route information is drawn up based on the fire spread information and sent to the safety officer's communication equipment, while tracking the blocking layer generation information; Step 6.7: Temporarily open the fire door on the safety officer's evacuation route, temporarily close the fire extinguishing facilities on the evacuation route, and generate the evacuation route navigation interface; Step 6.8: After the safety officer evacuates, close the corresponding fire doors and restart the fire extinguishing facilities.

Images