Intelligent control system and method of deluge valve based on fire prediction

Through the intelligent control system of rain valve based on fire prediction, the flow parameters are optimized using digital twin models and multi-party prediction channels, and real-time regulation is carried out in combination with long and short-term memory networks, the accuracy of substation fire prediction and fire extinguishing control is solved, and the accuracy of substation fire prediction and fire extinguishing control is reduced. Waste of water resources has improved the efficiency and quality of fire treatment.

CN118662832BActive Publication Date: 2024-11-15DALIAN DONGFANGYU TECH
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202411142082.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-20
Publication Date
2024-11-15
Estimated Expiration
2044-08-20

AI Technical Summary

Technical Problem

The existing rain valve control method used in substations is poor in accuracy of fire analysis and prediction, which makes it impossible to set appropriate fire extinguishing control parameters based on the fire prediction results, resulting in low accuracy and reliability of fire extinguishing control. There is a problem of waste of water resources.

Method used

Adopt the intelligent control system of rain valve based on fire prediction, and build a fire extinguishing analysis model through a digital twin model, combine multi-source sensing arrays and multi-party prediction channels for fire monitoring and prediction, optimize traffic parameters, and use long-term memory networks to perform trends. Analysis and real-time regulation ensure the accuracy and efficiency of fire extinguishing operations.

Benefits of technology

It improves the accuracy of fire situation prediction and the flexibility, accuracy and reliability of fire extinguishing control, reduces water resources and improves the efficiency and quality of fire situation treatment in substations.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118662832B_ABST
    Figure CN118662832B_ABST
Patent Text Reader

Abstract

The present application provides a deluge valve intelligent control system and method based on fire prediction, which relates to the field of intelligent control technology, including: a fire extinguishing analysis twin model construction module; a correlation feature extraction module; a simulated fire extinguishing analysis module, which is used to input the fire feature set into the multi-party prediction channel, output the fire prediction information, and perform simulated fire extinguishing analysis on the fire prediction information through the fire extinguishing analysis twin model to determine the simulated flow parameters; a simulated flow compensation module; a fire trend prediction module; a flow optimization and control module, which is used to optimize and control the flow parameters according to the fire trend prediction information. This application can solve the existing deluge valve control method. Due to the poor accuracy of fire analysis and prediction, it is impossible to set the corresponding fire extinguishing parameters according to the fire prediction results, resulting in poor accuracy and reliability of fire extinguishing control. The effect of reducing water resource waste and improving the efficiency and quality of fire handling can be achieved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of intelligent control technology, and in particular to an intelligent control system and method for a deluge valve based on fire prediction. Background Art

[0002] Deluge valve is an automatic fire-fighting device, mainly used in large-scale fire-fighting systems, especially in places where rapid and large-area water spraying is required, such as chemical plants, substations, etc. The main function of the deluge valve is to quickly open and release a large amount of water when a fire occurs, and evenly spray it around the fire source through the nozzle to achieve rapid and effective fire extinguishing.

[0003] At present, the existing deluge valve control method used in substations has poor accuracy in analyzing and predicting substation fire conditions, which makes it impossible to set appropriate fire extinguishing control parameters according to the fire prediction results. As a result, the accuracy and reliability of fire extinguishing control are poor, resulting in low efficiency and quality in substation fire handling, and there is also a technical problem of wasting water resources. Summary of the invention

[0004] The purpose of this application is to provide a deluge valve intelligent control system and method based on fire prediction, in order to solve the existing deluge valve control method used in substations. Due to the poor accuracy of substation fire analysis and prediction, it is impossible to set appropriate fire extinguishing control parameters according to the fire prediction results, resulting in poor accuracy and reliability of fire extinguishing control, resulting in low efficiency and quality of substation fire handling, and there is also a technical problem of waste of water resources.

[0005] In view of the above problems, the present application provides a deluge valve intelligent control system and method based on fire prediction.

[0006] In the first aspect, the present application provides a deluge valve intelligent control system based on fire prediction, which is used to execute a deluge valve intelligent control method based on fire prediction, including: a fire extinguishing analysis twin model construction module, which is used to simulate based on digital twins according to substation structure information, deluge valve attribute characteristics and fire extinguishing equipment information to construct a fire extinguishing analysis twin model; an associated feature extraction module, which is used to monitor the fire condition of the target substation based on a multi-source sensor array, extract multidimensional monitoring data according to predetermined associated features, and obtain a fire condition feature set; a simulated fire extinguishing analysis module, which is used to input the fire condition feature set into a multi-party prediction channel for comprehensive analysis, output fire condition prediction information, and use the fire extinguishing analysis twin model to simulate the fire condition prediction. The information is used to simulate fire extinguishing analysis to determine the simulated flow parameters of the working deluge valve; a simulated flow compensation module is used to obtain the associated pipeline characteristics of the working deluge valve, determine the flow loss coefficient based on the associated pipeline characteristics analysis, optimize and compensate the simulated flow parameters according to the flow loss coefficient, and obtain the optimal flow parameters; a fire trend prediction module is used to control the working deluge valve to perform fire extinguishing operations according to the optimal flow parameters, and to send back the real-time fire feature set to the trend analysis channel at a fixed point to obtain fire trend prediction information, wherein the trend analysis channel is constructed based on the long short-term memory network; a flow optimization and control module is used to optimize and control the flow parameters of the previous node in real time according to the fire trend prediction information.

[0007] On the second aspect, the present application provides a method for intelligent control of a deluge valve based on fire prediction, which is implemented through an intelligent control system of a deluge valve based on fire prediction, including: based on digital twins, simulation is performed according to substation structure information, deluge valve attribute characteristics and fire extinguishing equipment information to construct a fire extinguishing analysis twin model; based on a multi-source sensor array, fire monitoring is performed on the target substation, and multi-dimensional monitoring data is extracted according to predetermined correlation characteristics to obtain a fire feature set; the fire feature set is input into a multi-party prediction channel for comprehensive analysis, and fire prediction information is output; the fire prediction information is analyzed by the fire extinguishing analysis twin model Conduct simulated fire extinguishing analysis to determine the simulated flow parameters of the working deluge valve; obtain the associated pipeline characteristics of the working deluge valve, determine the flow loss coefficient based on the associated pipeline characteristics analysis, optimize and compensate the simulated flow parameters according to the flow loss coefficient, and obtain the optimal flow parameters; control the working deluge valve to perform fire extinguishing operations according to the optimal flow parameters, and send back the real-time fire feature set to the trend analysis channel at a fixed point to obtain fire trend prediction information, wherein the trend analysis channel is constructed based on a long short-term memory network; and optimize and control the flow parameters of the previous node in real time according to the fire trend prediction information.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0009] The fire condition feature set is acquired through monitoring, and the fire condition feature set is predicted and analyzed through the multi-party prediction channel constructed based on the multi-party prediction operator to obtain the fire condition prediction information, which can improve the accuracy and reliability of the fire condition prediction; then the fire condition prediction information is simulated and analyzed through the fire extinguishing analysis twin model to determine the working deluge valve and the simulated flow parameters of the working deluge valve, which can improve the adaptability of the simulated flow parameters to the fire condition prediction information; further based on the associated pipeline characteristic analysis of the working deluge valve, the pipeline flow loss coefficient is determined, and the simulated flow parameters are optimized and compensated according to the pipeline flow loss coefficient to obtain the optimal flow parameters; then according to the optimal flow The parameter-controlled working deluge valve performs fire extinguishing operations; on the other hand, during the fire extinguishing operation, the real-time fire feature set transmitted back from the fixed point is predicted through the trend analysis channel constructed based on the long short-term memory network to obtain the fire trend prediction information; finally, the flow parameters of the previous node are optimized and regulated in real time according to the fire trend prediction information; the above steps can improve the accuracy of substation fire analysis and prediction, and then the fire extinguishing control parameters can be optimized and adjusted in time according to the fire prediction results, so as to improve the flexibility, accuracy and reliability of fire extinguishing control, thereby achieving the technical effect of reducing water waste and improving the efficiency and quality of substation fire handling.

[0010] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented according to the contents of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically cited below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0012] Figure 1 This is a flow chart of a method for intelligently controlling a deluge valve based on fire prediction in this application;

[0013] Figure 2 A schematic diagram of a flow chart for constructing a multi-party prediction channel in a method for intelligently controlling a deluge valve based on fire prediction in this application;

[0014] Figure 3This is a structural schematic diagram of a deluge valve intelligent control system based on fire prediction in this application.

[0015] Description of reference numerals:

[0016] 11 Fire extinguishing analysis twin model construction module, 12 Correlation feature extraction module, 13 Simulation fire extinguishing analysis module, 14 Simulation flow compensation module, 15 Fire trend prediction module, 16 Flow optimization and control module. DETAILED DESCRIPTION

[0017] This application provides a deluge valve intelligent control system and method based on fire prediction, which solves the existing deluge valve control method for substations. Due to the poor accuracy of substation fire analysis and prediction, it is impossible to set the corresponding fire extinguishing control parameters according to the fire prediction results, which makes the accuracy and reliability of fire extinguishing control poor, resulting in low efficiency and quality of substation fire handling, and there is also a technical problem of water resource waste. The accuracy of substation fire analysis and prediction can be improved, and then the fire extinguishing control parameters can be optimized and adjusted in time according to the fire prediction results, and the flexibility, accuracy and reliability of fire extinguishing control can be improved, so as to achieve the technical effect of reducing water resource waste and improving the efficiency and quality of substation fire handling.

[0018] Below, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited to the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application. It should also be noted that, for the convenience of description, only the parts related to the present application are shown in the accompanying drawings, rather than all of them.

[0019] For example, please refer to the attached Figure 1 The present application provides a method for intelligently controlling a deluge valve based on fire prediction, which is applied to an intelligent control system of a deluge valve based on fire prediction, and specifically comprises the following steps:

[0020] Step 1: Based on the digital twin, simulation is performed according to the substation structure information, deluge valve property characteristics and fire extinguishing equipment information to build a fire extinguishing analysis twin model.

[0021] Specifically, first, obtain the substation structure information, deluge valve attribute characteristics and fire extinguishing equipment information, among which the substation structure information includes building structure, equipment type, equipment layout and electrical lines, etc. The equipment types include transformers, circuit breakers, switch cabinets, busbars, cables, etc.; the deluge valve attribute characteristics include deluge valve type, working principle, specifications, location coordinates and other parameters; the fire extinguishing equipment information includes data such as fire extinguishing equipment type and equipment location, such as water sprinkler fire extinguishing system, spray fire extinguishing system, etc.

[0022] Digital twin technology is a virtual, digital entity that models and simulates physical entities in the real world through sensors, data analysis, and simulation technology to achieve simulated operation of physical entities. Then, based on digital twin technology, simulation modeling is performed according to the substation structure information, deluge valve attribute characteristics, and fire extinguishing equipment information. For example, professional modeling software is used to create a 3D model of the substation based on the substation structure information, and then the information of the deluge valve and fire extinguishing equipment is integrated into the model to ensure the accuracy and completeness of the model. The fire extinguishing analysis twin model is obtained. By constructing a fire extinguishing analysis twin model, the accuracy and reliability of subsequent simulated fire extinguishing analysis can be improved.

[0023] Step 2: Fire monitoring of the target substation is performed based on a multi-source sensor array, and multi-dimensional monitoring data is extracted according to predetermined correlation features to obtain a fire feature set.

[0024] Specifically, first, a multi-source sensor array is constructed, where the multi-source sensor array includes multiple sensor groups, each of which includes a temperature sensor, a smoke detector, and a flame detector; then the sensor groups are evenly deployed in the target substation to ensure coverage of key areas and equipment, and the building structure, equipment layout, and environmental conditions of the substation should be considered during deployment.

[0025] Then, the target substation is monitored for fire based on the multi-source sensor array, that is, the sensors in the multi-source sensor array (such as temperature sensors, smoke detectors and flame detectors) are used to collect environmental data in the substation in real time to obtain multi-dimensional monitoring data; the multi-dimensional monitoring data is further extracted according to the predetermined associated features to obtain a fire feature set, which includes the type of fire equipment. The predetermined associated features refer to key features related to fire monitoring. These features can reflect the state and trend of the fire and can be set according to actual conditions, such as temperature change rate, smoke concentration, flame brightness, flame color and other features. By obtaining the fire feature set, data support is provided for the next step of fire prediction analysis.

[0026] Step three: input the fire feature set into the multi-party prediction channel for comprehensive analysis, output fire prediction information, perform simulated fire extinguishing analysis on the fire prediction information through the fire extinguishing analysis twin model, and determine the simulated flow parameters of the working deluge valve.

[0027] Specifically, first, configure the multi-party prediction operator, which refers to the algorithm or model used for fire prediction, wherein the multi-party prediction operator includes at least BP neural network, support vector machine and random decision forest, and additional adaptive prediction operators can also be added according to actual needs; then, build a prediction channel based on the multi-party prediction operator, and train the prediction channel to a convergence state through sample data; further establish a mapping association between equipment type and prediction channel, and build a multi-party prediction channel based on the mapping association, wherein the multi-party prediction channel includes multiple equipment types and multiple prediction channels, and the equipment type and the prediction channel correspond one to one.

[0028] Then, the fire condition feature set is input into the multi-party prediction channel. First, the corresponding prediction channel is determined according to the type of fire equipment. Then, the fire condition feature set is intelligently predicted and analyzed according to the matching prediction channel to obtain fire condition prediction information, wherein the fire condition prediction information includes the fire type, spread speed, fire size, affected range and severity. Then, the fire condition prediction information is input into the fire extinguishing analysis twin model to determine the working deluge valve and related fire extinguishing equipment corresponding to the fire condition prediction information. Further, based on the constraint condition of meeting the expected fire extinguishing index, a simulated fire extinguishing analysis is performed based on the working deluge valve and the related fire extinguishing equipment to obtain the simulated flow parameter of the working deluge valve, wherein the simulated flow parameter refers to the minimum flow parameter under the condition of meeting the expected fire extinguishing index. By constructing a fire extinguishing analysis twin model to simulate fire extinguishing analysis on the fire condition prediction information, the adaptability of the simulated flow parameter and the fire condition prediction information can be improved, thereby improving the accuracy and reliability of fire extinguishing control, and at the same time, it can avoid the waste of water resources caused by excessive flow parameters.

[0029] Step 4: Obtain the associated pipeline characteristics of the working deluge valve, determine the flow loss coefficient based on the associated pipeline characteristics analysis, optimize and compensate the simulated flow parameters according to the flow loss coefficient, and obtain the optimal flow parameters.

[0030] Specifically, the associated pipeline characteristics of the working deluge valve are obtained, wherein the associated pipeline characteristics include the inner diameter of the pipeline, the material of the pipeline, the length of the pipeline, the structural characteristics of the pipeline and the degree of pipeline aging; then, based on the associated pipeline characteristics, the pipeline flow loss analysis is performed to determine the pipeline flow loss coefficient, wherein the larger the pipeline flow loss coefficient, the larger the pipeline flow loss. Further, the simulated flow parameters are optimized and compensated according to the flow loss coefficient, that is, the simulated flow parameters are adjusted, and the part of the pipeline loss is added to the simulated flow parameters to ensure the match between the actual flow and the expected flow, wherein the larger the pipeline flow loss coefficient, the greater the compensation; and the optimal flow parameters are obtained, wherein the optimal flow parameters are the optimal flow settings under the influence of the flow loss coefficient, which can achieve the expected fire extinguishing effect.

[0031] By optimizing and compensating the simulated flow parameters based on the flow loss coefficient, the impact of flow loss can be reduced, ensuring that the actual flow of the deluge valve can meet the expected fire extinguishing needs, thereby improving the accuracy and reliability of fire extinguishing control.

[0032] Step 5: Control the working deluge valve to perform fire extinguishing operations according to the optimal flow parameters, and transmit the real-time fire feature set to the trend analysis channel at a fixed point to obtain fire trend prediction information, wherein the trend analysis channel is constructed based on a long short-term memory network.

[0033] Specifically, the working deluge valve is then controlled according to the optimal flow parameters to perform fire extinguishing operations. At the same time, during the fire extinguishing operation, at a predetermined time node, the real-time fire feature set transmitted back by the multi-source sensor array is monitored and acquired; on the other hand, a trend analysis channel is constructed based on the long short-term memory network. First, the structure of the long short-term memory network is designed, including an input layer, a hidden layer, and an output layer. The hidden layer contains multiple LSTM units, each of which can capture long-term dependencies in time series data; then the long short-term memory network is trained using historical data so that the network can predict the development trend of the fire and obtain a trend analysis channel that meets the convergence state. The real-time fire feature set is further input into the trend analysis channel to predict the fire trend and obtain fire trend prediction information.

[0034] Step 6: Optimize and control the flow parameters of the previous node in real time according to the fire trend prediction information.

[0035] Specifically, the historical fire trend prediction information of the adjacent previous node is obtained, and the fire trend prediction information is compared and analyzed according to the historical fire trend prediction information to determine the fire trend deviation; finally, the flow parameters of the previous node are optimized and regulated in real time according to the fire trend deviation, that is, the flow parameters are adjusted to adapt to the changes in the fire situation to ensure the optimization of the fire extinguishing effect. By optimizing and regulating the flow parameters in real time according to the fire trend prediction information, the flexibility, accuracy and reliability of fire extinguishing control can be improved, which helps to improve the fire safety level of the substation and save the consumption of fire water resources.

[0036] The method for intelligent control of a deluge valve based on fire prediction is applied to an intelligent control system for a deluge valve based on fire prediction, which can solve the existing deluge valve control method for substations. Due to the poor accuracy of fire analysis and prediction of substations, it is impossible to set the corresponding fire extinguishing control parameters according to the fire prediction results, resulting in poor accuracy and reliability of fire extinguishing control, resulting in low efficiency and quality of fire handling in substations, and there is also a technical problem of water resource waste. The fire feature set is acquired through monitoring, and the fire feature set is predicted and analyzed through a multi-party prediction channel constructed based on a multi-party prediction operator to obtain fire prediction information, which can improve the accuracy and reliability of fire prediction; then, the fire prediction information is simulated and extinguished through a fire extinguishing analysis twin model to determine the working deluge valve and the simulated flow parameters of the working deluge valve, which can improve the adaptability of the simulated flow parameters to the fire prediction information; further, the pipeline flow loss coefficient is determined based on the associated pipeline characteristic analysis of the working deluge valve, and the simulated flow parameters are optimized and compensated according to the pipeline flow loss coefficient to obtain the optimal flow parameters; then, according to the optimal flow The parameter-controlled working deluge valve performs fire extinguishing operations; on the other hand, during the fire extinguishing operation, the real-time fire feature set transmitted back from the fixed point is predicted through the trend analysis channel constructed based on the long short-term memory network to obtain the fire trend prediction information; finally, the flow parameters of the previous node are optimized and regulated in real time according to the fire trend prediction information; the above steps can improve the accuracy of substation fire analysis and prediction, and then the fire extinguishing control parameters can be optimized and adjusted in time according to the fire prediction results, so as to improve the flexibility, accuracy and reliability of fire extinguishing control, thereby achieving the technical effect of reducing water waste and improving the efficiency and quality of substation fire handling.

[0037] Further, a fire feature set is obtained. Step 2 of this application includes:

[0038] The multi-source sensor array includes multiple sensor groups, which are evenly distributed in the target substation, wherein each sensor group includes a temperature sensor, a smoke detector and a flame detector, and has a positioning mark; based on the multi-source sensor array, the target substation is monitored for fire, and feature extraction is performed on the multidimensional monitoring data set according to predetermined associated features to obtain a fire feature set, wherein the fire features at least include temperature distribution, smoke shape, smoke color, smoke concentration, flame shape, flame size and flame brightness, and are marked with the type of ignition equipment and the fire location coordinates.

[0039] Specifically, a multi-source sensor array is constructed, wherein the multi-source sensor array includes multiple sensor groups, and each sensor group includes a temperature sensor, a smoke detector and a flame detector, and then the multiple sensor groups are evenly arranged in the target substation to ensure coverage of key areas and equipment.

[0040] The target substation is monitored for fire through a multi-source sensor array to obtain a multidimensional monitoring data set; predetermined associated features are obtained, which refer to key features related to fire monitoring, and these features can reflect the status and trend of the fire; then, the multidimensional monitoring data set is extracted according to the predetermined associated features to obtain a fire feature set, wherein the fire characteristics at least include temperature distribution, smoke shape, smoke color, smoke concentration, flame shape, flame size and flame brightness, and are marked with the type of ignition equipment and the fire location coordinates.

[0041] Further, construct a multi-party prediction channel, as shown in the attached Figure 2 As shown, step three of this application includes:

[0042] Based on the substation structure information, multiple equipment types are obtained, and a first equipment type is randomly selected. With the first equipment type, deluge valve attribute characteristics and fire extinguishing equipment information as constraints, a sample data set is obtained based on power big data retrieval, wherein the sample data includes sample fire characteristics and sample fire information; a multi-party prediction operator is configured, and multiple prediction units are constructed based on the multi-party prediction operator, wherein the multi-party prediction operator at least includes a BP neural network, a support vector machine and a random decision forest; the sample data set is divided into Q parts, and Q parts are selected with replacement Q times to construct a first sample set, and Q sample sets are selected iteratively, wherein Q is the number of prediction operators; with the sample fire characteristics as input and the sample fire information as output, the Q sample sets are used to supervise the training of multiple prediction units to obtain multiple converged prediction units, wherein the converged prediction units have convergence degree identifiers; a fusion output layer is constructed according to the multiple convergence degrees, and a first prediction channel is constructed based on the multiple convergence prediction units and the fusion output layer; based on the binary tree principle, the multi-party prediction channel is constructed according to the mapping association between the first equipment type and the first prediction channel.

[0043] Specifically, first, multiple equipment types are obtained based on the substation structure information, and a first equipment type is randomly selected from the multiple equipment types, and the first equipment type is any one of the multiple equipment types; then, the first equipment type, the deluge valve attribute characteristics and the fire extinguishing equipment information are used as retrieval constraints, and information retrieval is performed based on the power big data, that is, homologous sample data with a high degree of correlation with the constraint conditions is retrieved to obtain a sample data set, wherein the sample data includes sample fire characteristics and sample fire information, and the sample fire information includes fire type, spread speed, fire size, affected range and severity.

[0044] Then configure the multi-party prediction operator, wherein the multi-party prediction operator includes at least BP neural network, support vector machine and random decision forest. These operators have different characteristics and advantages, and their combined use can improve the accuracy and robustness of the prediction. BP neural network is a feedforward artificial neural network, which is usually used for classification and regression tasks and can be iteratively optimized and trained; support vector machine is a supervised learning algorithm based on statistical learning theory, mainly used for classification and regression; random decision forest is an integrated learning method that improves prediction performance by constructing multiple decision trees and voting or averaging their results. Then, multiple prediction units are constructed based on the multi-party prediction operator, that is, each prediction unit adopts a specific prediction algorithm.

[0045] On the other hand, the sample data set is divided into Q parts to obtain Q data sets, ensuring that the number of samples in each part is roughly equal; then, Q samples are selected with replacement from the Q data sets to construct a first sample set, and the same method is further used to iteratively select Q samples to obtain Q sample sets, where Q is the number of prediction operators. Through this random sampling method with replacement, multiple sample sets can be constructed for training different prediction operators. This method helps to improve the prediction accuracy and generalization ability of the model, thereby improving the accuracy and reliability of fire prediction.

[0046] Then, the sample fire characteristics are used as input and the sample fire information is used as output. The Q sample sets are used to perform supervised training on multiple prediction units, wherein the supervised training can be performed using a back propagation algorithm, a gradient descent algorithm, etc., to obtain multiple converged prediction units that meet the expected output accuracy. The convergent prediction unit has a convergence degree identifier, and the convergence degree refers to the actual output accuracy of the convergent prediction unit, and the convergence degree of each convergent prediction unit is different. For example, if the expected output accuracy is 90%, the convergence degree of the prediction unit constructed based on the BP neural network is 95%.

[0047] Further, a fusion output layer is constructed according to multiple convergence degrees, wherein the fusion output layer is a multi-input single-output layer for integrating information from multiple prediction units, that is, weights are set according to the convergence degree, wherein the greater the convergence degree, the greater the weight, and then the output results of the multiple prediction units are weighted according to the weights, and the weighted calculation results are output; then a first prediction channel is constructed based on the multiple convergence prediction units and the fusion output layer. By constructing a prediction channel based on multiple prediction operators, since these operators each have different characteristics and advantages, the prediction error of a single operator can be reduced, and the accuracy and reliability of fire prediction can be improved.

[0048] Using the same method as that of obtaining the first prediction channel, multiple device types and corresponding prediction channels are obtained in turn, and further based on the binary tree principle, the multi-party prediction channel is constructed with the device type as the child node and the prediction channel as the leaf node of the child node.

[0049] Further, after outputting the fire prediction information, step three of this application includes:

[0050] Retrieve and obtain a sample verification set, perform verification tests on multiple prediction channels within the multi-party prediction channel respectively, and determine multiple comprehensive prediction errors, wherein the comprehensive prediction errors correspond to the prediction channels one by one; match the multiple comprehensive prediction errors based on the type of fire equipment to determine the actual prediction error, and compensate the fire prediction information according to the actual prediction error.

[0051] Specifically, a sample verification set is obtained based on the retrieval of electric power big data, and multiple prediction channels in the multi-party prediction channel are respectively verified and tested through the sample verification set to determine multiple comprehensive prediction errors, wherein the comprehensive prediction errors and prediction channels correspond one to one; then, the prediction channels corresponding to the ignition equipment types are matched in the multiple comprehensive prediction errors to obtain the actual prediction errors of the ignition equipment types; finally, the fire prediction information is compensated according to the actual prediction error, that is, the fire prediction information is expanded according to the actual prediction error, for example: if the actual prediction error is 5%, the fire prediction information is additionally expanded by 5%. By compensating the fire prediction information according to the actual prediction error, the influence of the prediction error can be reduced, and the accuracy and reliability of the fire prediction can be improved.

[0052] Further, to determine the simulated flow parameters of the working deluge valve, step three of this application includes:

[0053] The fire prediction information is input into the fire extinguishing analysis twin model, and the working deluge valve is determined based on the fire location coordinate mapping, wherein there is at least one working deluge valve; based on the fire extinguishing analysis twin model, with the expected fire extinguishing index being met as a constraint, a simulated fire extinguishing test is performed on the fire prediction information, and the expected flow parameters of the associated fire extinguishing equipment are output, wherein the expected flow parameters are the minimum flow parameters that meet the expected fire extinguishing index; according to the association characteristics between the deluge valve and the fire extinguishing equipment, the simulated flow parameters of the working deluge valve are determined based on the expected flow parameter mapping.

[0054] Specifically, first, the fire prediction information is input into the fire extinguishing analysis twin model, and the working deluge valve is determined according to the fire location coordinate mapping, wherein in the twin model, the location coordinates in the fire prediction information are corresponded to the actual layout of the substation, and the coordinate mapping in the twin model ensures the consistency between the fire location and the actual physical location; then, according to the fire location coordinates and the prediction information, the working deluge valve closest to the fire location is determined in the twin model, that is, a suitable deluge valve is selected to achieve effective control and extinguishing of the fire, wherein there is at least one working deluge valve.

[0055] The expected fire extinguishing indicators are obtained, wherein the expected fire extinguishing indicators include the fire control time and the fire control range, which can be set according to the actual situation to ensure that the fire extinguishing effect meets the safety requirements; then, based on the fire extinguishing analysis twin model, with the expected fire extinguishing indicators as a constraint, a simulated fire extinguishing test is performed on the fire prediction information. The model simulates the development of the fire and the fire extinguishing process according to the fire prediction information. At the same time, during the simulation process, with the expected fire extinguishing indicators as a constraint, the effects of different fire extinguishing strategies and parameter settings are evaluated. According to the test results, the expected flow parameters of the associated fire extinguishing equipment are output, wherein the expected flow parameters are the minimum flow parameters that meet the expected fire extinguishing indicators. Finally, according to the association characteristics between the deluge valve and the fire extinguishing equipment, the simulated flow parameters of the working deluge valve are determined based on the expected flow parameter mapping. First, the association characteristics between the deluge valve and the fire extinguishing equipment are analyzed, including the type, capacity, layout, etc. of the equipment, to determine the mapping relationship between these characteristics and the expected flow parameters; then, according to the results of the association characteristic analysis, a mapping relationship between the expected flow parameters and the performance of the fire extinguishing equipment is established, wherein the mapping relationship should consider factors such as the water supply capacity and flow restriction of the fire extinguishing equipment; finally, based on the mapping relationship, the simulated flow parameters of the working deluge valve are determined, and the simulated flow parameters should be able to meet the needs of the fire extinguishing equipment and ensure the effectiveness of the fire extinguishing effect.

[0056] Further, to determine the flow loss coefficient, step 4 of this application includes:

[0057] Obtain associated pipeline characteristics of the working deluge valve, wherein the associated pipeline characteristics include pipeline inner diameter, pipeline material, pipeline length, pipeline structural characteristics and pipeline aging degree; with the associated pipeline characteristics as constraints, obtain a sample flow loss ratio set through online retrieval; perform mean calculation on the sample flow loss ratio set, perform deviation calculation on the sample flow loss ratio set based on the loss ratio mean, and determine a loss ratio deviation set; eliminate sample flow loss ratios whose loss ratio deviations are greater than a deviation threshold to obtain a standard loss ratio set, and select the mode of the standard loss ratio set as the flow loss coefficient.

[0058] Specifically, first, the associated pipeline characteristics of the working deluge valve are obtained, wherein the associated pipeline characteristics include pipeline inner diameter, pipeline material, pipeline length, pipeline structural characteristics and pipeline aging degree; then, using the associated pipeline characteristics as constraints, an online search is performed to obtain a sample flow loss ratio set, wherein the flow loss ratio refers to the ratio of the loss flow to the pipeline transport flow; then, the sample flow loss ratio set is mean calculated to obtain a loss ratio mean; further, based on the loss ratio mean, deviations of multiple sample flow loss ratios in the sample flow loss ratio set are calculated to obtain a loss ratio deviation set.

[0059] Obtain a deviation threshold, which can be set according to data accuracy requirements, and then eliminate sample flow loss ratios whose loss ratio deviation is greater than the deviation threshold, that is, eliminate unreliable sample data, to obtain a standard loss ratio set, and finally select the mode of the standard loss ratio set as the flow loss coefficient.

[0060] Further, to construct a trend analysis channel, step five of this application includes:

[0061] Taking the first equipment type, deluge valve attribute characteristics and fire extinguishing equipment information as constraints, a sample fire feature time series set is retrieved and obtained, wherein the sample fire feature time series contains historical fire characteristics of multiple adjacent nodes; a first trend analysis unit is constructed based on a long short-term memory network, and the first trend analysis unit is trained to convergence using the sample fire feature time series set through a back propagation algorithm; the trend analysis channel is generated according to the mapping relationship between the first equipment type and the first trend analysis unit.

[0062] Specifically, first, obtain the first equipment type, where the first equipment type is any one of multiple equipment types; then use the first equipment type, deluge valve attribute characteristics and fire extinguishing equipment information as constraints to retrieve and obtain a sample fire feature time series set, where the sample fire feature time series contains historical fire features of multiple adjacent nodes.

[0063] The long short-term memory network is a special recurrent neural network suitable for processing sequential sequence data. First, a first trend analysis unit is constructed based on the long short-term memory network. For example, the structure of the long short-term memory network is designed, including an input layer, a hidden layer and an output layer. The hidden layer contains multiple LSTM units, and each unit can capture the long-term dependency in the time series data. The input data of the long short-term memory network is the fire characteristics of the previous node, and the output data is the fire trend prediction result of the next adjacent node. Further, through the back propagation algorithm, the first trend analysis unit is trained to convergence using the sample fire characteristic time series set, that is, in each training step, the weight and bias of the model are updated by calculating the gradient of the loss function, and iterative training is continuously performed to adjust the parameters of the model to minimize the loss function. At the same time, a convergence condition is set, such as the value of the loss function is less than a certain threshold, or the change of the loss function during training is less than a certain threshold. When the convergence condition is met, the training process is stopped.

[0064] Finally, the same method is used to obtain multiple trend analysis units corresponding to multiple device types, and a trend analysis channel is constructed based on the mapping relationship between the device type and the trend analysis unit.

[0065] Further, the flow parameters of the previous node are optimized and regulated in real time according to the fire trend prediction information. Step 6 of this application includes:

[0066] Based on the historical fire trend prediction information of the previous node, the fire trend prediction information is subjected to deviation calculation to determine the fire trend deviation; the fire trend deviation is input into the flow correction database for matching to obtain the flow correction coefficient, wherein the flow correction database is constructed based on a classification decision tree; and the flow parameters of the previous node are optimized and regulated in real time according to the flow correction coefficient.

[0067] Specifically, first, according to the historical fire trend prediction information of the previous node, the deviation of the fire trend prediction information is calculated to obtain the fire trend deviation, wherein the fire trend deviation refers to the difference between the fire trend prediction information and the historical fire trend prediction information. Then, a flow correction database is constructed based on the classification decision tree, wherein the flow correction database includes multiple sample fire trend deviations and multiple sample flow correction coefficients, and the sample fire trend deviation and the sample flow correction coefficient are in a one-to-one relationship, and then the fire trend deviation is input into the flow correction database for matching to obtain the flow correction coefficient, wherein the flow correction coefficient reflects the adjustment ratio of the flow parameter of the previous node, for example: the fire trend deviation is greater than 0, indicating that the flow parameter needs to be increased, wherein the larger the deviation value, the greater the increase; finally, the flow parameter of the previous node is optimized and regulated in real time according to the flow correction coefficient, so as to achieve the goal of optimizing and adjusting the fire extinguishing control parameters in a timely manner according to the fire prediction results, and achieve the technical effect of improving the flexibility, accuracy and reliability of fire extinguishing control.

[0068] In summary, the intelligent control method of deluge valve based on fire prediction provided by the present application has the following technical effects:

[0069] 1. Obtain the fire condition feature set through monitoring, and perform predictive analysis on the fire condition feature set through a multi-party prediction channel constructed based on a multi-party prediction operator to obtain fire condition prediction information, which can improve the accuracy and reliability of fire condition prediction; then simulate the fire extinguishing analysis of the fire condition prediction information through the fire extinguishing analysis twin model to determine the working deluge valve and the simulated flow parameters of the working deluge valve, which can improve the adaptability of the simulated flow parameters to the fire condition prediction information; further determine the pipeline flow loss coefficient based on the associated pipeline characteristic analysis of the working deluge valve, and optimize and compensate the simulated flow parameters according to the pipeline flow loss coefficient to obtain the optimal flow parameters; then optimize the optimal flow parameters according to the optimal flow On the other hand, during the fire extinguishing process, the trend analysis channel constructed based on the long short-term memory network is used to predict the fire trend of the real-time fire feature set sent back from the fixed point, and the fire trend prediction information is obtained; finally, the flow parameters of the previous node are optimized and regulated in real time according to the fire trend prediction information; the above steps can improve the accuracy of substation fire analysis and prediction, and then the fire extinguishing control parameters can be optimized and adjusted in time according to the fire prediction results, so as to improve the flexibility, accuracy and reliability of fire extinguishing control, thereby achieving the technical effect of reducing water waste and improving the efficiency and quality of substation fire handling.

[0070] 2. By constructing a prediction channel based on multiple prediction operators, since these operators each have different characteristics and advantages, the prediction error of a single operator can be reduced and the accuracy and reliability of fire prediction can be improved.

[0071] 3. By constructing a fire extinguishing analysis twin model to simulate fire extinguishing analysis of fire prediction information, the adaptability of simulated flow parameters and fire prediction information can be improved, thereby improving the accuracy and reliability of fire extinguishing control, and at the same time avoiding the waste of water resources caused by excessive flow parameters.

[0072] Embodiment 2: Based on the same invention concept as the deluge valve intelligent control method based on fire prediction in the above embodiment, this application also provides a deluge valve intelligent control system based on fire prediction, please refer to the attached Figure 3 ,include:

[0073] The fire extinguishing analysis twin model construction module 11 is used to construct a fire extinguishing analysis twin model based on digital twins, by simulating according to substation structure information, deluge valve property characteristics and fire extinguishing equipment information.

[0074] The correlation feature extraction module 12 is used to perform fire monitoring on the target substation based on the multi-source sensor array, extract the multi-dimensional monitoring data according to the predetermined correlation features, and obtain a fire feature set.

[0075] The simulated fire extinguishing analysis module 13 is used to input the fire feature set into the multi-party prediction channel for comprehensive analysis, output fire prediction information, perform simulated fire extinguishing analysis on the fire prediction information through the fire extinguishing analysis twin model, and determine the simulated flow parameters of the working deluge valve.

[0076] The simulated flow compensation module 14 is used to obtain the associated pipeline characteristics of the working deluge valve, determine the flow loss coefficient based on the associated pipeline characteristics analysis, and optimize and compensate the simulated flow parameters according to the flow loss coefficient to obtain the optimal flow parameters.

[0077] The fire trend prediction module 15 is used to control the working deluge valve to perform fire extinguishing operations according to the optimal flow parameters, and to transmit the real-time fire feature set to the trend analysis channel at a fixed point to obtain fire trend prediction information, wherein the trend analysis channel is constructed based on a long short-term memory network.

[0078] The flow optimization and control module 16 is used to optimize and control the flow parameters of the previous node in real time according to the fire trend prediction information.

[0079] Furthermore, the fire prediction-based deluge valve intelligent control system is also used for:

[0080] The multi-source sensor array includes multiple sensor groups, which are evenly distributed in the target substation, wherein each sensor group includes a temperature sensor, a smoke detector and a flame detector, and has a positioning mark; based on the multi-source sensor array, the target substation is monitored for fire, and feature extraction is performed on the multidimensional monitoring data set according to predetermined associated features to obtain a fire feature set, wherein the fire features at least include temperature distribution, smoke shape, smoke color, smoke concentration, flame shape, flame size and flame brightness, and are marked with the type of ignition equipment and the fire location coordinates.

[0081] Furthermore, the fire prediction-based deluge valve intelligent control system is also used for:

[0082] Based on the substation structure information, multiple equipment types are obtained, and a first equipment type is randomly selected. With the first equipment type, deluge valve attribute characteristics and fire extinguishing equipment information as constraints, a sample data set is obtained based on power big data retrieval, wherein the sample data includes sample fire characteristics and sample fire information; a multi-party prediction operator is configured, and multiple prediction units are constructed based on the multi-party prediction operator, wherein the multi-party prediction operator at least includes a BP neural network, a support vector machine and a random decision forest; the sample data set is divided into Q parts, and Q parts are selected with replacement Q times to construct a first sample set, and Q sample sets are selected iteratively, wherein Q is the number of prediction operators; with the sample fire characteristics as input and the sample fire information as output, the Q sample sets are used to supervise the training of multiple prediction units to obtain multiple converged prediction units, wherein the converged prediction units have convergence degree identifiers; a fusion output layer is constructed according to the multiple convergence degrees, and a first prediction channel is constructed based on the multiple convergence prediction units and the fusion output layer; based on the binary tree principle, the multi-party prediction channel is constructed according to the mapping association between the first equipment type and the first prediction channel.

[0083] Furthermore, the fire prediction-based deluge valve intelligent control system is also used for:

[0084] Retrieve and obtain a sample verification set, perform verification tests on multiple prediction channels within the multi-party prediction channel respectively, and determine multiple comprehensive prediction errors, wherein the comprehensive prediction errors correspond to the prediction channels one by one; match the multiple comprehensive prediction errors based on the type of fire equipment to determine the actual prediction error, and compensate the fire prediction information according to the actual prediction error.

[0085] Furthermore, the fire prediction-based deluge valve intelligent control system is also used for:

[0086] The fire prediction information is input into the fire extinguishing analysis twin model, and the working deluge valve is determined based on the fire location coordinate mapping, wherein there is at least one working deluge valve; based on the fire extinguishing analysis twin model, with the expected fire extinguishing index being met as a constraint, a simulated fire extinguishing test is performed on the fire prediction information, and the expected flow parameters of the associated fire extinguishing equipment are output, wherein the expected flow parameters are the minimum flow parameters that meet the expected fire extinguishing index; according to the association characteristics between the deluge valve and the fire extinguishing equipment, the simulated flow parameters of the working deluge valve are determined based on the expected flow parameter mapping.

[0087] Furthermore, the fire prediction-based deluge valve intelligent control system is also used for:

[0088] Obtain associated pipeline characteristics of the working deluge valve, wherein the associated pipeline characteristics include pipeline inner diameter, pipeline material, pipeline length, pipeline structural characteristics and pipeline aging degree; with the associated pipeline characteristics as constraints, obtain a sample flow loss ratio set through online retrieval; perform mean calculation on the sample flow loss ratio set, perform deviation calculation on the sample flow loss ratio set based on the loss ratio mean, and determine a loss ratio deviation set; eliminate sample flow loss ratios whose loss ratio deviations are greater than a deviation threshold to obtain a standard loss ratio set, and select the mode of the standard loss ratio set as the flow loss coefficient.

[0089] Furthermore, the fire prediction-based deluge valve intelligent control system is also used for:

[0090] Taking the first equipment type, deluge valve attribute characteristics and fire extinguishing equipment information as constraints, a sample fire feature time series set is retrieved and obtained, wherein the sample fire feature time series contains historical fire characteristics of multiple adjacent nodes; a first trend analysis unit is constructed based on a long short-term memory network, and the first trend analysis unit is trained to convergence using the sample fire feature time series set through a back propagation algorithm; the trend analysis channel is generated according to the mapping relationship between the first equipment type and the first trend analysis unit.

[0091] Furthermore, the fire prediction-based deluge valve intelligent control system is also used for:

[0092] Based on the historical fire trend prediction information of the previous node, the fire trend prediction information is subjected to deviation calculation to determine the fire trend deviation; the fire trend deviation is input into the flow correction database for matching to obtain the flow correction coefficient, wherein the flow correction database is constructed based on a classification decision tree; and the flow parameters of the previous node are optimized and regulated in real time according to the flow correction coefficient.

[0093] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The intelligent control method for a deluge valve based on fire prediction and the specific examples in the aforementioned embodiment 1 are also applicable to an intelligent control system for a deluge valve based on fire prediction in this embodiment. Through the aforementioned detailed description of an intelligent control method for a deluge valve based on fire prediction, those skilled in the art can clearly understand an intelligent control system for a deluge valve based on fire prediction in this embodiment, so for the sake of brevity of the specification, it will not be described in detail here. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the description of the method part.

[0094] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

[0095] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technology, the present application is also intended to include these modifications and variations.

Claims

1. The intelligent control system of deluge valve based on fire prediction is characterized by: include: The fire extinguishing analysis twin model construction module is used to simulate the fire extinguishing analysis twin model based on the digital twin, according to the substation structure information, deluge valve attribute characteristics and fire extinguishing equipment information; A correlation feature extraction module is used to monitor the fire condition of the target substation based on a multi-source sensor array, extract the multi-dimensional monitoring data according to the predetermined correlation features, and obtain a fire condition feature set; A simulated fire extinguishing analysis module is used to input the fire condition feature set into the multi-party prediction channel for comprehensive analysis, output fire condition prediction information, perform simulated fire extinguishing analysis on the fire condition prediction information through the fire extinguishing analysis twin model, and determine the simulated flow parameters of the working deluge valve; A simulated flow compensation module, used for obtaining the associated pipeline characteristics of the working deluge valve, analyzing and determining the flow loss coefficient based on the associated pipeline characteristics, optimizing and compensating the simulated flow parameters according to the flow loss coefficient, and obtaining the optimal flow parameters; A fire trend prediction module, used to control the working deluge valve to perform fire extinguishing operations according to the optimal flow parameters, and to transmit the real-time fire feature set to the trend analysis channel at a fixed point to obtain fire trend prediction information, wherein the trend analysis channel is constructed based on a long short-term memory network; A flow optimization and control module, used to optimize and control the flow parameters of the previous node in real time according to the fire trend prediction information; The step of inputting the fire condition feature set into a multi-party prediction channel for comprehensive analysis includes: Based on the substation structure information, multiple equipment types are obtained, and a first equipment type is randomly selected. The first equipment type, the attribute characteristics of the deluge valve, and the fire extinguishing equipment information are used as constraints, and a sample data set is obtained based on the power big data retrieval, wherein the sample data includes sample fire characteristics and sample fire information; Configuring a multi-party prediction operator, and constructing a plurality of prediction units based on the multi-party prediction operator, wherein the multi-party prediction operator includes at least a BP neural network, a support vector machine, and a random decision forest; Divide the sample data set into Q equal parts, select Q parts with replacement, construct a first sample set, iterate and select Q parts, and obtain Q sample sets, where Q is the number of prediction operators; Taking the sample fire condition feature as input and the sample fire condition information as output, the Q sample sets are used to perform supervised training on multiple prediction units to obtain multiple converged prediction units, wherein the converged prediction units have convergence degree identifiers; constructing a fusion output layer according to the multiple convergence degrees, and constructing a first prediction channel based on the multiple convergence prediction units and the fusion output layer; Based on the binary tree principle, the multi-party prediction channel is constructed according to the mapping association between the first device type and the first prediction channel.

2. The intelligent control system for deluge valves based on fire prediction according to claim 1 is characterized in that: Get the fire characteristics set, including: The multi-source sensor array includes a plurality of sensor groups, which are evenly arranged in the target substation, wherein each sensor group includes a temperature sensor, a smoke detector and a flame detector, and has a positioning mark; Fire monitoring of the target substation is carried out based on a multi-source sensor array, and feature extraction is performed on the multidimensional monitoring data set according to predetermined correlation features to obtain a fire feature set, wherein the fire characteristics at least include temperature distribution, smoke shape, smoke color, smoke concentration, flame shape, flame size and flame brightness, and are marked with the type of ignition equipment and the fire location coordinates.

3. The intelligent control system for deluge valves based on fire prediction according to claim 1 is characterized in that: Output fire prediction information, followed by: Retrieving and obtaining a sample verification set, performing verification tests on multiple prediction channels in the multi-party prediction channel respectively, and determining multiple comprehensive prediction errors, wherein the comprehensive prediction errors correspond to the prediction channels one by one; The multiple comprehensive prediction errors are matched based on the type of the ignition equipment to determine an actual prediction error, and the fire prediction information is compensated according to the actual prediction error.

4. The intelligent control system for deluge valves based on fire prediction according to claim 1 is characterized in that: Determine the simulated flow parameters of the working deluge valve, including: Inputting the fire prediction information into the fire extinguishing analysis twin model, determining a working deluge valve based on the fire location coordinate mapping, wherein there is at least one working deluge valve; Based on the fire extinguishing analysis twin model, with the expected fire extinguishing index being satisfied as a constraint, a simulated fire extinguishing test is performed on the fire prediction information, and an expected flow parameter of the associated fire extinguishing equipment is output, wherein the expected flow parameter is a minimum flow parameter that satisfies the expected fire extinguishing index; According to the association characteristics between the deluge valve and the fire extinguishing equipment, the simulated flow parameters of the working deluge valve are determined based on the expected flow parameter mapping.

5. The intelligent control system for deluge valves based on fire prediction according to claim 1 is characterized in that: Determine the flow loss factor, including: Obtaining associated pipeline characteristics of the working deluge valve, wherein the associated pipeline characteristics include pipeline inner diameter, pipeline material, pipeline length, pipeline structural characteristics, and pipeline aging degree; Taking the associated pipeline characteristics as constraints, obtaining a sample flow loss ratio set through online retrieval; Performing mean calculation on the sample flow loss ratio set, performing deviation calculation on the sample flow loss ratio set based on the loss ratio mean, and determining a loss ratio deviation set; Eliminate sample flow loss ratios whose loss ratio deviation is greater than a deviation threshold to obtain a standard loss ratio set, and select the mode of the standard loss ratio set to set as the flow loss coefficient.

6. The intelligent control system for deluge valves based on fire prediction according to claim 1 is characterized in that: Construct trend analysis channels, including: Taking the first equipment type, the deluge valve attribute characteristics and the fire extinguishing equipment information as constraints, a sample fire condition feature time series set is retrieved and obtained, wherein the sample fire condition feature time series contains historical fire condition characteristics of multiple adjacent nodes; The first trend analysis unit is constructed based on the long short-term memory network, and the first trend analysis unit is trained to convergence using the sample fire characteristic time series set through the back propagation algorithm; The trend analysis channel is generated according to a mapping relationship between the first device type and the first trend analysis unit.

7. The intelligent control system for deluge valves based on fire prediction according to claim 1 is characterized in that: According to the fire trend prediction information, the flow parameters of the previous node are optimized and regulated in real time, including: Based on the historical fire trend prediction information of the previous node, performing deviation calculation on the fire trend prediction information to determine the fire trend deviation; Inputting the fire trend deviation into a flow correction database for matching to obtain a flow correction coefficient, wherein the flow correction database is constructed based on a classification decision tree; The flow parameter of the previous node is optimized and regulated in real time according to the flow correction coefficient.

8. The intelligent control method of deluge valve based on fire prediction is characterized in that: The method is implemented by the fire prediction-based deluge valve intelligent control system according to any one of claims 1 to 7, comprising: Based on digital twins, simulation is performed according to substation structure information, deluge valve attribute characteristics and fire extinguishing equipment information to build a fire extinguishing analysis twin model; Fire monitoring of the target substation is performed based on a multi-source sensor array, and multi-dimensional monitoring data is extracted according to predetermined correlation features to obtain a fire feature set; Input the fire condition feature set into the multi-party prediction channel for comprehensive analysis, output fire condition prediction information, perform simulated fire extinguishing analysis on the fire condition prediction information through the fire extinguishing analysis twin model, and determine the simulated flow parameters of the working deluge valve; Acquire the associated pipeline characteristics of the working deluge valve, determine the flow loss coefficient based on the associated pipeline characteristics, optimize and compensate the simulated flow parameters according to the flow loss coefficient, and obtain the optimal flow parameters; According to the optimal flow parameter, the working deluge valve is controlled to perform fire extinguishing operation, and a real-time fire feature set is sent back to a trend analysis channel at a fixed point to obtain fire trend prediction information, wherein the trend analysis channel is constructed based on a long short-term memory network; Optimizing and regulating the flow parameters of the previous node in real time according to the fire trend prediction information; The step of inputting the fire condition feature set into a multi-party prediction channel for comprehensive analysis includes: Based on the substation structure information, multiple equipment types are obtained, and a first equipment type is randomly selected. The first equipment type, the attribute characteristics of the deluge valve, and the fire extinguishing equipment information are used as constraints, and a sample data set is obtained based on the power big data retrieval, wherein the sample data includes sample fire characteristics and sample fire information; Configuring a multi-party prediction operator, and constructing a plurality of prediction units based on the multi-party prediction operator, wherein the multi-party prediction operator includes at least a BP neural network, a support vector machine, and a random decision forest; Divide the sample data set into Q equal parts, select Q parts with replacement, construct a first sample set, iterate and select Q parts, and obtain Q sample sets, where Q is the number of prediction operators; Taking the sample fire condition feature as input and the sample fire condition information as output, the Q sample sets are used to perform supervised training on multiple prediction units to obtain multiple converged prediction units, wherein the converged prediction units have convergence degree identifiers; constructing a fusion output layer according to the multiple convergence degrees, and constructing a first prediction channel based on the multiple convergence prediction units and the fusion output layer; Based on the binary tree principle, the multi-party prediction channel is constructed according to the mapping association between the first device type and the first prediction channel.

Citation Information

Patent Citations

  • Intelligent fire control management method and system

    CN115120920A

  • Intelligent fire-fighting video monitoring and alarm linkage control system

    CN116884167A

  • Smoke thermal simulation training system and training effect evaluation method

    CN117357852A

  • KR20220161956A