CN115905998A - Cable tunnel fire extinguishing decision method combining multi-sensor data fusion with FWA-BP neural network algorithm - Google Patents
Cable tunnel fire extinguishing decision method combining multi-sensor data fusion with FWA-BP neural network algorithm Download PDFInfo
- Publication number
- CN115905998A CN115905998A CN202211421291.2A CN202211421291A CN115905998A CN 115905998 A CN115905998 A CN 115905998A CN 202211421291 A CN202211421291 A CN 202211421291A CN 115905998 A CN115905998 A CN 115905998A
- Authority
- CN
- China
- Prior art keywords
- data
- neural network
- sensor
- fire
- fusion
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Fire Alarms (AREA)
Abstract
The invention discloses a cable tunnel fire-extinguishing decision method combining multi-sensor data fusion with an FWA-BP neural network algorithm, wherein multi-parameter sensors adopted by the fire-extinguishing decision method are arranged at a plurality of point positions, each point position comprises at least one temperature sensor, one CO sensor and one smoke sensor, and the fire-extinguishing decision method comprises the following steps: A. fusing the same-parameter multi-point monitoring data by adopting an improved fuzzy support function to obtain temperature fusion data, CO concentration fusion data and smoke fusion data; B. performing principal component analysis noise reduction on the obtained multi-parameter fusion data, and creating and training a multi-parameter data fusion model based on an FWA-BP neural network algorithm, namely a cable fire FWA-BP neural network model; C. test data is input for verification, and a fire extinguishing decision and a fire extinguishing range are output. The invention can greatly reduce the false alarm rate of fire alarm and ensure the safe and stable operation of the high-voltage cable.
Description
Technical Field
The invention belongs to the field of on-line monitoring and fault diagnosis of power equipment, and particularly relates to a cable tunnel fire-extinguishing decision-making method combining multi-sensor data fusion with an FWA-BP neural network algorithm.
Background
In production sites where electrical equipment is concentrated, such as substations and power plants, the power cables and communication cables of the equipment are usually arranged in cable trenches. In actual production, the cable trench often has a narrow space and a dense cable, which are undesirable factors, due to field conditions, construction costs, and various other reasons. These factors are not favorable to the maintenance and repair of the routine line, and are easy to cause fire due to short circuit, overload, resistance increase and the like, thereby causing serious production accidents. According to statistics of relevant data of cable fire accidents, it can be known that fire accidents caused by cables occur for many times in China in recent 20 years, especially cable fires of thermal power plants and transformer substations, the fire accidents occur about 140 times, and more than 70% of cable fires have serious loss.
In most cases, however, a series of precursor events must occur before a cable fire can occur. These events include an abnormal rise in cable temperature and the generation of smoke particles, as well as other warning signals. If the fire risk can be positioned and the fault can be eliminated in the early stage of the fire, the occurrence of fire accidents can be effectively avoided, and the safe operation of industrial equipment is ensured. The automatic fire alarm system in the current cable tunnel has the advantages that the reliability of the device is not high or the adaptability to severe environments is not strong, the available period is short, great difficulty is caused to overhaul and maintenance, and false alarm is often generated, so that monitoring personnel lose alertness of alarm, and timely discovery and effective control are difficult to achieve when a fire initial stage occurs.
Disclosure of Invention
The invention aims to solve the problems in the prior art, and provides a cable tunnel fire-extinguishing decision-making method combining multi-sensor data fusion with an FWA-BP neural network algorithm.
The invention aims to solve the problems by the following technical scheme:
a cable tunnel fire extinguishing decision method combining multi-sensor data fusion with an FWA-BP neural network algorithm is characterized by comprising the following steps: the fire-extinguishing decision-making method adopts the multi-parameter sensors which are arranged at a plurality of point positions, the multi-parameter sensor at each point position comprises at least one temperature sensor, one CO sensor and one smoke sensor, and the fire-extinguishing decision-making method comprises the following steps:
A. fusing the same-parameter multi-point monitoring data by adopting an improved fuzzy support function to obtain temperature fusion data, CO concentration fusion data and smoke concentration fusion data;
B. performing principal component analysis and noise reduction on the obtained multi-parameter fusion data, and creating and training a multi-parameter data fusion model based on a FWA-BP neural network algorithm, namely a cable fire FWA-BP neural network model;
C. and inputting the test data into the cable fire FWA-BP neural network model for verification, and outputting a fire extinguishing decision and a fire extinguishing range.
The temperature sensor is attached to a cable tunnel cable metal support through heat-conducting silica gel to output a temperature signal for monitoring the condition of the cable, and the CO sensor and the smoke sensor are installed according to a sensing radiation range.
The improved fuzzy support function in the step A is to replace the absolute distance in the DTW method with the dynamic bending distance and calculate the support among the monitoring data from the same type of sensors.
The specific steps of improving the fuzzy support function are as follows:
a1, a calculation formula of the support degree among monitoring data of the sensors of the same type is as follows:
sup(U,V)＝2×(1+e K×dist(U,V) ) -1 (1)
in the formula (1), dist (U, V) is the dynamic bending distance of the time sequence data U and V, and the higher K is more than or equal to 0,K, the higher the discrimination degree of the mutual support degree is;
a2, a plurality of sensors of the same type respectively collect temperature signals, CO concentration signals and smoke concentration signals to obtain a data set X i ＝{X 1 ,X 2 ,...,X m And (i =1,2,.., m), wherein m is the number of sensors of the same type, and the support degree s between data of different sites of the sensors of the same type in the time period T can be obtained according to the formula (1) ij And further constructing a fuzzy support matrix S,
a3, in the time period T, the support degree of the sensors except the sensor to the ith sensor is shown in a formula (3), the value range of the support degree is [0,1],
in the formula (3), gamma i (T) the support degree of the sensor except the sensor for the ith sensor;
a4, if the data collected by the ith sensor in the time period T is x i (T)＝{x i (1),x i (2),...,x i (n), where n is the number of data, the mean and variance of the ith sensor node data are respectively shown in formula (4) and formula (5):
in the formulae (4) and (5), z i (T) is the mean value of the data of the ith sensor node,
variance of the data of the ith sensor node;
a5, completing data fusion among sensors of the same type through a weighted fusion algorithm, wherein the final weighted fusion expression is as follows:
in the formula (6), ω i (T) represents a weighted value corresponding to time-series data acquired by the ith sensor within the time period T, and the calculation method is as shown in formula (7):
the number of the sensors of the same type in the step A2 is at least 3.
The principal component analysis in the step B is reducedThe principal component analysis method adopted by noise can extract characteristic information from high-dimensional data, reduce dimensionality and retain the maximum information of the original high-dimensional data, and the specific process is as follows: in cable trench fire monitoring system n 1 Is m in total in unit time 1 N is formed by fused data acquired by different sensors at the same time 1 Line m 1 A column sample matrix X, wherein the sample matrix X is transformed by formula (8), formula (9) and formula (10) to obtain a standardized matrix U 1 ，
Solving a correlation coefficient matrix R of the normalized matrix U through a formula (11), solving a characteristic equation shown in a formula (12),
further obtaining m 1 A characteristic value according to
Determining p values, namely when the information contribution rate of the current p principal components is more than 85%, taking the first p principal components as sample characteristics, arranging the characteristic values in a descending order, and forming a transformation matrix A by taking characteristic vectors corresponding to the first p characteristic values T ：
A T ＝(u 1 ,u 2 ,...,u p ),p＜m (13)
Calculating Y = AX to obtain the first p principal components Y, so that the purpose of reducing the dimension of the fire monitoring fusion data can be realized; x is n 1 Line m 1 A sample matrix of columns.
B, the BP neural network in the FWA-BP neural network algorithm in the step B is a multilayer forward network, the multilayer comprises an input layer, a hidden layer and an output layer, the number of nodes of the input layer is 3, and the nodes respectively represent temperature fusion data, CO concentration fusion data and smoke concentration fusion data; the number of nodes of the output layer is 2, namely a fire state risk assessment value and a trigger value for triggering the fire extinguishing system or not, wherein the trigger value only comprises two states of 0 and 1, 0 represents no trigger and 1 represents trigger; the hidden layer of the BP neural network adopts a nonlinear Sigmoid function, while the output layer usually adopts a linear Purelin function, wherein the Sigmoid function is divided into logsig types, and expressions are respectively shown as an expression (14) and an expression (15):
purelin(x)＝x (15)。
before the FWA optimization of the FWA-BP neural network algorithm in the step B is started, the population size, the upper limit and the lower limit of firework dimension, the explosion spark radius adjusting constant, the explosion spark number adjusting constant, the gaussian spark number and the parameter of the maximum number of generations need to be determined, then the fitness value of each firework is calculated, and the square error and the SSE of the neural network are used as fitness functions:
in the formula (16), t is the expected output of the BP neural network, s is the neuron number of the output layer of the BP neural network, and y is the actual output value of the BP neural network;
finally, carrying out high-precision training optimization on the weight and the threshold value of the BP neural network by using a Levenberg-Marquardt algorithm, and setting a training target error function as a square error and an SSE; when the training reaches the target error or the maximum algebra, the establishment of the ideal cable fire FWA-BP neural network model is completed, otherwise, the training is continued by returning to the previous step.
Compared with the prior art, the invention has the following advantages:
the invention provides a cable tunnel fire extinguishing decision method combining multi-sensor data fusion with an FWA-BP neural network algorithm, which adopts a multi-parameter multi-point sensor to monitor fire characteristic parameters, carries out multi-parameter data fusion based on the FWA-BP neural network algorithm, solves the problems of abnormal values, incompleteness and even inconsistency of observation targets of measurement data of different sensors due to complexity and uncertainty of an observation environment through multi-parameter multi-point two-dimensional data fusion, displays a fire at a terminal in time after obvious abnormal faults are identified, triggers a fire extinguishing device, greatly reduces the false alarm rate of fire alarm, avoids severe accidents, and effectively ensures safe and stable operation of a high-voltage cable.
Drawings
FIG. 1 is a general flow chart of an algorithm of the cable tunnel fire extinguishing decision method of the present invention;
FIG. 2 is a flow chart of building a FWA-BP neural network model of a cable fire according to the present invention;
FIG. 3 is a diagram showing a prediction result of the FWA-BP neural network model for a cable fire according to the present invention after test data is input.
Detailed Description
The invention is further described with reference to the following figures and examples.
As shown in fig. 1: the invention provides a cable tunnel fire extinguishing decision method combining multi-sensor data fusion with an FWA-BP neural network algorithm; the fire extinguishing decision method for the cable tunnel comprises a multi-parameter multi-point sensing device, a same-parameter multi-point data fusion method and a multi-parameter data fusion method based on a FWA-BP neural network algorithm. In the multi-parameter multi-point sensing device, the multi-parameter sensor of each point position comprises at least one temperature sensor, a CO sensor and a smoke sensor, the temperature sensor is attached to a cable tunnel cable metal support by adopting heat-conducting silica gel to output a temperature signal for monitoring the condition of the cable, and the CO sensor and the smoke sensor are installed according to a sensing radiation range.
The same-parameter multi-point data fusion method is used for carrying out data fusion on the same-parameter sensing data by a data fusion method for improving fuzzy support degree. The method improves a fuzzy support function, replaces absolute distances in a DTW method with dynamic bending distances by improving the fuzzy support function, and calculates the support among monitoring data from the sensors of the same type, as shown in formula (1):
sup(U,V)＝2×(1+e K×dist(U,V) ) -1 (1)
in the formula (1), dist (U, V) is the dynamic bending distance of the time sequence data U and V, and the higher K is more than or equal to 0,K, the higher the discrimination degree of the mutual support degree is;
according to the invention, a plurality of sensors of the same type are used for respectively acquiring a temperature signal, a CO concentration signal and a smoke concentration signal to obtain a data set X i ＝{X 1 ,X 2 ,...,X m The method comprises the following steps of (i =1,2,.., m), wherein m is the number of the sensors of the same type, the larger the value of m is, the better the data fusion effect is, but the sensor resources are excessively wasted, and the cost of the fire extinguishing decision method is increased, so that the number of the sensors of the same type is recommended to be 3. The support degree s between data of different sites of the same type of sensor in the time period T can be obtained according to the formula (1) ij And further constructing a fuzzy support matrix S,
in the time period T, the support degree of the sensors except the sensor to the ith sensor is shown in a formula (3), the value range of the support degree is [0,1],
in the formula (3), gamma i (T) is other than itselfThe support degree of the outer sensor to the ith sensor;
if the data collected by the ith sensor in the time period T is x i (T)＝{x i (1),x i (2),...,x i (n), where n is the number of data, the mean and the variance of the ith sensor node data are respectively shown in formula (4) and formula (5):
in the formulae (4) and (5), z i (T) is the mean value of the data of the ith sensor node,
variance of the data of the ith sensor node;
data fusion among the sensors of the same type is completed through a weighted fusion algorithm, and the final weighted fusion expression is as follows:
in the formula (6), ω i (T) represents a weighted value corresponding to time-series data acquired by the ith sensor in the time period T, and the calculation method is as shown in formula (7):
the multi-parameter data fusion method based on the FWA-BP neural network algorithm comprises the steps of data acquisition, principal component analysis and dimensionality reduction, creation and training of a multi-parameter data fusion model based on the FWA-BP neural network algorithm, prediction of a fire result according to detection data, and decision of whether to trigger a fire extinguishing system and a fire extinguishing range.
The method can extract characteristic information from high-dimensional data, reduce dimensionality and retain maximum information of original high-dimensional data as a principal component analysis method widely used in a data processing technology; the specific process is as follows: in cable trench fire monitoring system n 1 Is m in total in unit time 1 N is formed by fused data acquired by different sensors at the same time 1 Line m 1 A column sample matrix X, wherein the sample matrix X is transformed by formula (8), formula (9) and formula (10) to obtain a standardized matrix U 1 。
Solving a correlation coefficient matrix R of the normalized matrix U through a formula (11), and solving a characteristic equation shown in a formula (12),
further obtaining m 1 A characteristic value according to
Determining p values, namely when the information contribution rate of the current p principal components is more than 85%, taking the first p principal components as sample characteristics, arranging the characteristic values in a descending order, and forming a transformation matrix A by taking characteristic vectors corresponding to the first p characteristic values T ：
A T ＝(u 1 ,u 2 ,...,u p ),p＜m (13)
Calculating Y = AX to obtain the first p principal components Y, so that the purpose of reducing the dimension of the fire monitoring fusion data can be realized; x is n 1 Line m 1 A sample matrix of columns.
As shown in the flow chart of fig. 2 for establishing the FWA-BP neural network model of the cable fire, the FWA-BP neural network algorithm firstly establishes a BP neural network, which is a multilayer forward network, the multilayer comprises an input layer, a hidden layer and an output layer, the number of nodes of the input layer is 3, and the nodes are respectively temperature fusion data, CO concentration fusion data and smoke concentration fusion data; the number of nodes of the output layer is 2, namely a fire state risk assessment value and a trigger value for triggering the fire extinguishing system or not, wherein the trigger value only comprises two states of 0 and 1, 0 represents no trigger and 1 represents trigger; the hidden layer of the BP neural network adopts a nonlinear Sigmoid function, while the output layer usually adopts a linear Purelin function, wherein the Sigmoid function is divided into logsig types, and expressions are respectively shown as an expression (14) and an expression (15):
purelin(x)＝x (15)。
before the FWA optimization is started, the population size, the upper limit and the lower limit of firework dimensionality, an explosion spark radius adjusting constant, an explosion spark number adjusting constant, a Gaussian spark number and a maximum iteration number parameter are determined, then the fitness value of each firework is calculated, and the square error and SSE of a neural network are used as fitness functions:
in the formula (16), t is the expected output of the BP neural network, s is the neuron number of the output layer of the BP neural network, and y is the actual output value of the BP neural network;
finally, carrying out high-precision training optimization on the weight and the threshold value of the BP neural network by using a Levenberg-Marquardt algorithm, and setting a training target error function as a square error sum SSE; when the training reaches the target error or the maximum algebra, the establishment of the ideal cable fire FWA-BP neural network model is completed, otherwise, the training is continued by returning to the previous step.
As shown in fig. 3, 380 sets of data are selected as training data for the prediction result of the FWA-BP neural network model of the cable fire for test data input through published papers, field research and the cable fire test according to the present invention, wherein 9 data points are randomly extracted as test data, and the test data is tested. As can be seen from fig. 3, the predicted result substantially matches the expected result, and 3 fire sample data of the elliptical region in the figure can be successfully identified.
Through the model, the prediction and judgment of the cable tunnel fire can be finally realized. When the fire is identified by the monitored parameters, the superfine dry powder fire extinguishing system is triggered to extinguish the fire, and in addition, the fire extinguishing range is determined by the arrangement method of the multipoint sensors.
According to the invention, the ground fire monitoring terminal can check the monitoring data of the temperature, the CO concentration and the smoke concentration in real time, and can also check the running states of each temperature sensor, each CO concentration sensor and each smoke concentration sensor, when the data are found to be disconnected, the system fault can be checked according to the prompt information data, and the later maintenance is facilitated.
What needs to be supplemented is: typically, the fuzzy support function is measured by absolute distance, and if the absolute distance between two sensor data points at any given time is small, the fuzzy support between them is large; if the absolute distance between two data points is large, it is determined that the degree of blur support between them is small. However, in processing time series sensor data, this approach ignores the correlation information that exists between the data before and after time, and it is impractical to estimate the mutual support provided by the two sensors based only on the proximity of the data. Therefore, in order to analyze the problems, the invention adopts an improved fuzzy support function, the improved fuzzy support function replaces the absolute distance in the DTW method with the dynamic bending distance, and the support between the monitoring data from the same type of sensor is calculated.
The invention provides a cable tunnel fire extinguishing decision method combining multi-sensor data fusion with an FWA-BP neural network algorithm, which adopts a multi-parameter multi-point sensor to monitor fire characteristic parameters, carries out multi-parameter data fusion based on the FWA-BP neural network algorithm, solves the problems of abnormal values, incompleteness, even inconsistency of observation targets and the like of measurement data of different sensors due to complexity and uncertainty of an observation environment through multi-parameter multi-point two-dimensional data fusion, displays a fire at a terminal in time after an obvious abnormal fault is identified, triggers a fire extinguishing device, greatly reduces the false alarm rate of fire alarm, avoids the occurrence of severe accidents, and effectively ensures safe and stable operation of a high-voltage cable.
The cable tunnel fire-extinguishing decision-making method combining multi-sensor data fusion and the FWA-BP neural network algorithm is suitable for complex field implementation environments of cable tunnels, is low in cost, can be used as an effective supplementary means of a traditional temperature measurement optical fiber method, and ensures safe and stable operation of high-voltage cables.
The above embodiments are only for illustrating the technical idea of the present invention, and the protection scope of the present invention should not be limited thereby, and any modification made on the basis of the technical solution according to the technical idea proposed by the present invention is within the protection scope of the present invention; the technology not related to the invention can be realized by the prior art.
Claims (8)
1. A cable tunnel fire extinguishing decision method combining multi-sensor data fusion with an FWA-BP neural network algorithm is characterized by comprising the following steps: the multi-parameter sensor adopted by the fire-extinguishing decision-making method is arranged at a plurality of point positions, the multi-parameter sensor at each point position comprises at least one temperature sensor, one CO sensor and one smoke sensor, and the fire-extinguishing decision-making method comprises the following steps:
A. fusing the same-parameter multi-point monitoring data by adopting an improved fuzzy support function to obtain temperature fusion data, CO concentration fusion data and smoke concentration fusion data;
B. performing principal component analysis and noise reduction on the obtained multi-parameter fusion data, and creating and training a multi-parameter data fusion model based on a FWA-BP neural network algorithm, namely a cable fire FWA-BP neural network model;
C. and inputting the test data into the cable fire FWA-BP neural network model for verification, and outputting a fire extinguishing decision and a fire extinguishing range.
2. The cable tunnel fire-extinguishing decision method combining multi-sensor data fusion and FWA-BP neural network algorithm according to claim 1, wherein: the temperature sensor is attached to a cable tunnel cable metal support through heat-conducting silica gel to output a temperature signal for monitoring the condition of the cable, and the CO sensor and the smoke sensor are installed according to a sensing radiation range.
3. The cable tunnel fire-extinguishing decision method combining multi-sensor data fusion and FWA-BP neural network algorithm according to claim 1, wherein: and the improved fuzzy support function in the step A is to replace the absolute distance in the DTW method with the dynamic bending distance and calculate the support among the monitoring data from the same type of sensors.
4. The cable tunnel fire-extinguishing decision method combining multi-sensor data fusion and FWA-BP neural network algorithm according to claim 3, wherein: the specific steps of improving the fuzzy support function are as follows:
a1, a calculation formula of the support degree among the monitoring data of the sensors of the same type is as follows:
sup(U,V)＝2×(1+e K×dist(U,V) ) -1 (1)
in the formula (1), dist (U, V) is the dynamic bending distance of the time series data U and V, and the higher K is not less than 0,K, the higher the discrimination degree of the mutual support degree is;
a2, a plurality of sensors of the same type respectively collect temperature signals, CO concentration signals and smoke concentration signals to obtain a data set X i ＝{X 1 ,X 2 ,...,X m }(i=1,2, ·, m), m is the number of the sensors of the same type, and the support degree s between the data of different sites of the sensors of the same type in the time period T can be obtained according to the formula (1) ij And further constructing a fuzzy support matrix S,
a3, in the time period T, the support degree of the sensors except the sensor to the ith sensor is shown in a formula (3), the value range of the support degree is [0,1],
in the formula (3), gamma i (T) the support degree of the sensor except the sensor for the ith sensor;
a4, if the data collected by the ith sensor in the time period T is x i (T)＝{x i (1),x i (2),...,x i (n), where n is the number of data, the mean and variance of the ith sensor node data are respectively shown in formula (4) and formula (5):
in the formulae (4) and (5), z i (T) is the mean value of the ith sensor node data, δ i 2 (T) is the variance of the ith sensor node data;
a5, completing data fusion among sensors of the same type through a weighted fusion algorithm, wherein the final weighted fusion expression is as follows:
in the formula (6), ω i (T) represents a weighted value corresponding to time-series data acquired by the ith sensor within the time period T, and the calculation method is as shown in formula (7):
5. the fire extinguishing decision method for the cable tunnel by combining multi-sensor data fusion and FWA-BP neural network algorithm according to claim 1, wherein: the number of the sensors of the same type in the step A2 is at least 3.
6. The cable tunnel fire-extinguishing decision method combining multi-sensor data fusion and FWA-BP neural network algorithm according to claim 1, wherein: the principal component analysis method adopted for principal component analysis noise reduction in the step B can extract characteristic information from high-dimensional data, reduce dimensionality and retain maximum information of original high-dimensional data, and the specific process is as follows: in cable trench fire monitoring system n 1 Is m in total in unit time 1 N is formed by fusion data acquired by different sensors at the same time 1 Line m 1 A column sample matrix X, wherein the sample matrix X is transformed by formula (8), formula (9) and formula (10) to obtain a standardized matrix U 1 ，
Solving a correlation coefficient matrix R of the normalized matrix U through a formula (11), solving a characteristic equation shown in a formula (12),
further obtaining m 1 A characteristic value according to
Determining p values, namely when the information contribution rate of the current p principal components is more than 85%, taking the first p principal components as sample characteristics, arranging the characteristic values in a descending order, and forming a transformation matrix A by taking characteristic vectors corresponding to the first p characteristic values T ：
A T ＝(u 1 ,u 2 ,...,u p ),p＜m (13)
Y = AX is calculated to obtain the first p principal components Y, and the purpose of reducing the dimension of the fire monitoring fusion data can be achieved; x is n 1 Line m 1 A sample matrix of columns.
7. The fire extinguishing decision method for the cable tunnel by combining multi-sensor data fusion and FWA-BP neural network algorithm according to claim 1 or 6, wherein: b, the BP neural network in the FWA-BP neural network algorithm in the step B is a multilayer forward network, the multilayer comprises an input layer, a hidden layer and an output layer, the number of nodes of the input layer is 3, and the nodes respectively represent temperature fusion data, CO concentration fusion data and smoke concentration fusion data; the number of nodes of the output layer is 2, namely a fire state risk assessment value and a trigger value for triggering the fire extinguishing system or not, wherein the trigger value only comprises two states of 0 and 1, 0 represents no trigger and 1 represents trigger; the hidden layer of the BP neural network adopts a nonlinear Sigmoid function, while the output layer usually adopts a linear Purelin function, wherein the Sigmoid function is divided into logsig types, and expressions are respectively shown as an expression (14) and an expression (15):
purelin(x)＝x (15)。
8. the fire extinguishing decision method for the cable tunnel by combining multi-sensor data fusion and FWA-BP neural network algorithm according to claim 1 or 7, wherein: before the FWA optimization of the FWA-BP neural network algorithm in the step B is started, the population size, the upper limit and the lower limit of firework dimension, the explosion spark radius adjusting constant, the explosion spark number adjusting constant, the gaussian spark number and the parameter of the maximum number of generations need to be determined, then the fitness value of each firework is calculated, and the square error and the SSE of the neural network are used as fitness functions:
in the formula (16), t is the expected output of the BP neural network, s is the neuron number of the output layer of the BP neural network, and y is the actual output value of the BP neural network;
finally, carrying out high-precision training optimization on the weight and the threshold value of the BP neural network by using a Levenberg-Marquardt algorithm, and setting a training target error function as a square error and an SSE; when the training reaches the target error or the maximum algebra, the establishment of the ideal cable fire FWA-BP neural network model is completed, otherwise, the training is continued by returning to the previous step.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211421291.2A CN115905998A (en) | 2022-11-14 | 2022-11-14 | Cable tunnel fire extinguishing decision method combining multi-sensor data fusion with FWA-BP neural network algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211421291.2A CN115905998A (en) | 2022-11-14 | 2022-11-14 | Cable tunnel fire extinguishing decision method combining multi-sensor data fusion with FWA-BP neural network algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115905998A true CN115905998A (en) | 2023-04-04 |
Family
ID=86496615
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211421291.2A Pending CN115905998A (en) | 2022-11-14 | 2022-11-14 | Cable tunnel fire extinguishing decision method combining multi-sensor data fusion with FWA-BP neural network algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115905998A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116882268A (en) * | 2023-06-15 | 2023-10-13 | 重庆大学 | Data-driven tunnel fire smoke development prediction method and intelligent control system |
-
2022
- 2022-11-14 CN CN202211421291.2A patent/CN115905998A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116882268A (en) * | 2023-06-15 | 2023-10-13 | 重庆大学 | Data-driven tunnel fire smoke development prediction method and intelligent control system |
| CN116882268B (en) * | 2023-06-15 | 2024-02-06 | 重庆大学 | Data-driven tunnel fire smoke development prediction method and intelligent control system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wu et al. | An intelligent tunnel firefighting system and small-scale demonstration | |
| CN103886374B (en) | A kind of cable connector conductor temperature Forecasting Methodology based on RBF neural | |
| CN112464813B (en) | Mountain fire monitoring method and device | |
| CN109284938A (en) | A kind of comprehensive estimation method and device of power cable line state | |
| CN112002095A (en) | Fire early warning method in mine tunnel | |
| CN110533294B (en) | Nuclear power plant operation fault alarm method based on artificial intelligence technology | |
| CN110793616A (en) | All-fiber distributed cable safety and reliability monitoring system | |
| CN111986436B (en) | Comprehensive flame detection method based on ultraviolet and deep neural networks | |
| CN114169237B (en) | Power cable joint temperature abnormality early warning method combining EEMD-LSTM and isolated forest algorithm | |
| CN115905998A (en) | Cable tunnel fire extinguishing decision method combining multi-sensor data fusion with FWA-BP neural network algorithm | |
| CN109767352A (en) | A kind of power information physics emerging system safety situation evaluation method | |
| CN113516837B (en) | Urban fire judgment method and system based on multi-source information fusion and storage medium thereof | |
| CN116052360B (en) | Fire alarm system | |
| CN112016835A (en) | Power distribution network cable line monitoring method, computer equipment and storage medium | |
| CN117275206A (en) | Electrical fire monitoring and early warning system based on Internet of things | |
| CN111400114A (en) | Deep recursion network-based big data computer system fault detection method and system | |
| Wang et al. | Forest fire detection system based on Fuzzy Kalman filter | |
| CN116777088B (en) | Power supply emergency repair environment monitoring method and system for guaranteeing life safety | |
| CN116546043A (en) | Cable tunnel inspection personnel environment safety monitoring device and method | |
| CN114897084A (en) | Tower crane structure safety monitoring method based on graph convolution neural network | |
| CN114997375A (en) | Insulator surface metal pollution model building system in air type switch cabinet | |
| CN202961637U (en) | Intelligent cable duct real-time fireproof monitoring and automatic fire extinguishing system | |
| CN111199254A (en) | High-voltage electrical equipment real-time detection method based on RBF neural network and Bayesian network | |
| CN115494353A (en) | Switch cabinet safety monitoring method based on holographic data and ultrasonic signal diagnosis | |
| CN117706258B (en) | Fault detection system based on big data processing |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |