Monitoring device and monitoring method for deformation quantity of pressure gas cylinder of fire extinguishing system

Abstract

The invention discloses a device and a method for monitoring the deformation quantity of a pressure gas cylinder of a fire fighting system, and relates to the field of pressure gas cylinder shape monitoring, the device comprises a pressure gas cylinder body used for collecting a stress change signal corresponding to the pressure gas cylinder through a sensor; the deformation monitoring processing end is used for predicting the deformation quantity of the pressure gas cylinder according to the stress change signal and a characteristic relation model, and evaluating the damage degree of the pressure gas cylinder based on the deformation quantity and the installation environment; and the gas cylinder monitoring control end is used for generating a terminal alarm prompt rule according to the damage degree result, and the management terminal adjusts the use state of the pressure gas cylinder according to the prompt rule. By combining the stress change signal with the characteristic relation model, the deformation quantity of the pressure gas cylinder can be predicted in real time, potential dangers can be recognized in advance, meanwhile, the fatigue life of the pressure gas cylinder is analyzed based on the deformation quantity, and the durability of the gas cylinder in long-term use can be effectively predicted.

Description

Technical Field

[0001] The present invention relates to the field of pressure gas cylinder shape monitoring, and in particular to a device and method for monitoring the shape of a pressure gas cylinder in a fire protection system. Background Art

[0002] Pressure gas cylinders in fire protection systems are usually devices used to store and supply fire extinguishing media (such as carbon dioxide, foam, gas fire extinguishing agents, etc.). Pressure gas cylinders are an important part of the fire protection system and are responsible for providing rapid and efficient fire extinguishing support when a fire occurs. Pressure gas cylinders store fire extinguishing agents at high pressure and can release them quickly when needed, ensuring that the fire extinguishing agents can be sprayed out with sufficient pressure in a short time to ensure the fire extinguishing effect.

[0003] Therefore, pressure gas cylinders in firefighting systems are critical components for ensuring the smooth operation of firefighting operations. Their stored high-pressure gas is rapidly released upon a fire, exerting its extinguishing effect. However, over long-term use, pressure gas cylinders can experience deformation changes due to factors such as internal gas pressure, ambient temperature fluctuations, and the aging of the cylinder's own materials. When deformation exceeds a safe range, the cylinder's seal and structural stability are compromised, impairing not only the firefighting effectiveness but also potentially causing serious safety incidents such as explosions, posing a significant threat to human life and property.

[0004] In the existing technology, the inspection of pressure cylinders in fire protection systems mainly relies on regular manual inspections, and the status of the cylinders is judged through visual inspections, pressure tests, etc. However, this method has obvious defects. For example, the intervals between manual inspections are long, making it difficult to capture real-time changes in the deformation of the cylinders and to detect potential safety hazards in the early stages. At the same time, traditional non-destructive testing technology, when applied to pressure cylinders, has detection limitations due to the particularity of the cylinder structure and usage environment, making it impossible to achieve continuous and effective monitoring of the deformation.

[0005] Currently, no effective solutions have been proposed for the problems in related technologies. Summary of the Invention

[0006] In response to the problems in the related art, the present invention proposes a deformation monitoring device and a monitoring method for a pressure gas cylinder in a fire protection system to overcome the above-mentioned technical problems existing in the existing related art.

[0007] To this end, the specific technical solutions adopted in the present invention are as follows: In a first aspect, the present invention provides a device for monitoring the deformation of a pressure gas cylinder in a fire protection system, the device comprising: The pressure gas cylinder body is used to collect stress change signals corresponding to the pressure gas cylinder through sensors; The deformation monitoring and processing end is used to predict the deformation of the pressure gas cylinder based on the stress change signal and the characteristic relationship model, and to assess the degree of damage to the pressure gas cylinder based on the deformation and installation environment; The gas cylinder monitoring and control terminal is used to generate terminal alarm prompt rules based on the damage degree results, and the management terminal adjusts the use status of the pressure gas cylinder according to the prompt rules.

[0008] Preferably, collecting the stress change signal corresponding to the pressure gas cylinder through the sensor includes: Using the strain sensor at the threaded connection between the middle of the fire-fighting system pressure gas cylinder and the bottle mouth, the strain change signal data caused by the internal pressure, external environment and loose threads of the fire-fighting system pressure gas cylinder is obtained; Signal testing technology is used to analyze the signal transmission stability of the wireless transmission method, and based on the analysis results, it is determined whether there are signal interference and signal loss problems in the transmission process of the strain change signal data; If there is a problem, the transmission parameters of the wireless transmission method are adjusted; if there is no problem, the strain change signal data is uploaded to the characteristic relationship model.

[0009] Preferably, the deformation monitoring processing end includes: A deformation prediction unit is used to combine the stress change signal with the characteristic relationship model to predict the deformation of the pressure gas cylinder; The damage degree prediction unit is used to analyze the fatigue life and burst pressure of the pressure gas cylinder based on the deformation variable, and predict the damage degree of the pressure gas cylinder in combination with the installation environment of the pressure gas cylinder.

[0010] Preferably, combining the stress change signal with the characteristic relationship model to predict the deformation of the pressure gas cylinder includes: Analyze the correlation characteristics between the historical stress data and historical deformation data of pressure cylinders in fire protection systems; Utilize relevant feature information to build a characteristic relationship model with correlation calculation as the core, and output the deformation of the pressure gas cylinder; Preferably, analyzing the relevant characteristic information between the historical stress data and the historical deformation data of the pressure cylinders in the fire protection system includes: Obtain historical stress data and historical deformation data of pressure cylinders in different time periods during the use of the fire protection system, and perform time alignment processing on the historical stress data and historical deformation data; Based on the alignment processing results, a time window is selected to extract the change relationship between stress and deformation. Point pair features are constructed based on the change relationship results, and edge features of the point pair features are analyzed using edge convolution. A full edge connectivity matrix is ​​constructed based on the edge features and change relationships, and the connectivity matrix between stress and deformation is analyzed based on the full edge connectivity matrix. The topological relationship between historical stress data and historical deformation data is analyzed based on the connectivity matrix, and the relevant characteristic information between stress and deformation is analyzed according to the topological relationship.

[0011] Preferably, constructing a full edge connectivity matrix based on edge features and change relationships, and analyzing the connectivity matrix between stress and deformation based on the full edge connectivity matrix includes: Perform node relationship search on edge features and change relationships, generate a node association table based on the search results, use the node association table to establish an adjacency matrix of the connection relationship, and multiply the adjacency relationship by itself; The full edge connectivity matrix is ​​obtained by squaring the self-multiplied result, and the continuous variables whose edge features are positively correlated with the change relationship are identified based on the full edge connectivity matrix; Establish a multiple linear regression equation with continuous variables as dependent variables, and use ridge regression technology to analyze the historical stress data and historical deformation data in different time periods to form the continuous variables of the regression equation; All continuous variables are combined, and the connection matrix between stress and deformation is generated based on the combination results.

[0012] Preferably, a characteristic relationship model with correlation calculation as the core is constructed using relevant feature information, and the deformation variables of the pressure cylinder are output including: The correlation feature information is combined with the self-attention optimization technology to generate feature vectors, and the weight coefficient of the correlation feature information is obtained by calculating the correlation between the feature vectors; The attention feature of the correlation feature coefficient is obtained based on the weight coefficient, and the attention feature is combined with the cross attention technology to generate an optimized feature, and the correlation matrix is ​​obtained according to the optimized feature; The correlation matrix is ​​combined with the correlation model reconstruction technology to perform correlation model reconstruction processing, and a characteristic relationship model is obtained based on the processing results.

[0013] Preferably, analyzing the fatigue life and burst pressure of the pressure gas cylinder based on the deformation variable and predicting the damage degree of the pressure gas cylinder in combination with the installation environment of the pressure gas cylinder includes: Obtain the initial installation environment, structure, and material information of the fire protection system pressure cylinder, and generate a pressure cylinder structural model based on the deformation variables, and obtain the pressure cylinder structural model as the initial image representation diagram; The orthogonal test technology is used to set the test factor table of the pressure cylinder structure model. The fatigue life of the pressure cylinder is obtained based on the test factor table and the life curve analysis method. The damage accumulation analysis of the pressure cylinder is performed based on the damage criterion to obtain the damage accumulation. Predict the burst pressure of pressure gas cylinders under different working conditions based on fatigue life and loss accumulation, and obtain images of the explosion installation environment of pressure gas cylinders when the burst pressure is reached; The similarity between the blasting installation environment image and the initial image representation is compared, and the degree of morphological damage of the pressure gas cylinder under deformation conditions is evaluated based on the similarity results.

[0014] Preferably, comparing the similarity between the blasting installation environment image and the initial image representation image, and evaluating the degree of morphological damage of the pressure gas cylinder under the deformation condition based on the similarity result includes: Select two reference points on the main axis of the initial image representation, analyze the frequency graph between the initial image representation and the reference points, and combine the two sets of frequency graphs for normalization to obtain a direction vector; Extracting the representation feature parameters of the initial image representation map according to the direction vector, and generating a feature set based on the representation feature parameters to construct an image cloud model that can reflect the morphological image; Inputting the representation feature parameters corresponding to the blasting installation environment image into the image cloud model for similarity analysis, and obtaining the similarity results between the blasting installation environment image and the initial image representation map; An evaluation system is constructed based on the similarity results and the set indicators to analyze the degree of morphological damage of the pressure cylinder under deformation conditions.

[0015] Preferably, two reference points are selected on the main axis of the initial image representation, a frequency graph between the initial image representation and the reference points is analyzed, and the two sets of frequency graphs are combined and normalized to obtain a direction vector, which includes: Analyze the centroid position according to the pixel coordinates of the initial image representation map, combine the centroid position with the pixels of the initial image representation map to calculate the covariance matrix, and generate eigenvalues ​​and eigenvectors based on the covariance matrix; Determine the rotation angle between the main axis and the horizontal axis of the initial image representation based on the eigenvalue and the eigenvector, obtain the main axis direction, and obtain the distance from each pixel coordinate point in the initial image representation to the center of mass; Select the maximum value of the distance result to determine the two reference points on the main axis, analyze the coordinate values ​​of the two reference points on the initial image representation map, and use the frequency map between the coordinate values ​​and the initial image representation map; The two sets of frequency maps are merged, and normalized, median filtered and noise removed. According to the processing results, spatial features are extracted from the frequency maps as the direction vectors of the initial image representation map.

[0016] Preferably, extracting characteristic feature parameters of the initial image representation graph according to the direction vector, and generating a feature set based on the characteristic feature parameters to construct an image cloud model that can reflect the morphological image includes: Selecting gray-level co-occurrence matrices in different directions from the direction vector, analyzing the entropy and correlation characteristics of the gray-level co-occurrence matrix, and extracting the representation feature parameters of the initial image representation map; The representation feature parameters are combined with the membership of the central node reflecting the direction vector generated by the reverse cloud generator to the initial image representation map, and a preset number of representation feature parameters are selected based on the membership results to generate a feature set; Determine the mean and variance of the feature set. According to the analysis of the mean and variance results, the expected value and information entropy of the characteristic parameters corresponding to the morphological image can be reflected. Based on the expected value and information entropy results, the characteristic parameters are converted into an image cloud model.

[0017] In a second aspect, the present invention further proposes a method for monitoring the deformation of a pressure gas cylinder in a fire protection system, the method comprising: The stress change signal corresponding to the pressure gas cylinder is collected through the sensor; Predict the deformation of the pressure gas cylinder based on the stress change signal and characteristic relationship model, and evaluate the degree of damage to the pressure gas cylinder based on the deformation and installation environment; The terminal alarm prompt rules are generated according to the damage degree results, and the management terminal adjusts the use status of the pressure gas cylinder according to the prompt rules.

[0018] The beneficial effects of the present invention are: 1. By combining stress change signals with characteristic relationship models, the present invention can predict the deformation of pressure gas cylinders in real time, identifying potential hazards in advance. At the same time, based on the deformation analysis, the fatigue life of the pressure gas cylinders can effectively predict the durability of the gas cylinders in long-term use. By knowing the fatigue life of the pressure gas cylinders in advance and predicting their degree of damage, it helps to assess the safety of the gas cylinders before they are damaged, facilitating the provision of more accurate maintenance and replacement plans.

[0019] 2. The present invention can monitor the deformation of pressure gas cylinders in real time and continuously, capture any subtle changes in time, and discover safety hazards at the first time. It adopts high-precision micro-strain sensors and algorithms, fully considers the special circumstances of pressure gas cylinders, and ensures the high accuracy of deformation monitoring. Once the deformation exceeds the standard, the system responds quickly and issues an alarm, thus buying valuable emergency response time for the fire department and effectively reducing the risk caused by pressure gas cylinder failure when a fire occurs. At the same time, by obtaining historical stress data and deformation data of the fire protection system in different time periods, a characteristic relationship model is constructed to output the deformation of the pressure gas cylinder, which is convenient for later diagnosis of potential problems with the pressure gas cylinder, taking repair measures in advance, reducing equipment failure rate, and improving the safety and stability of the fire protection system.

[0020] 3. By acquiring the initial installation environment, structure, and material information of the pressure gas cylinder and combining it with the deformation variable to generate an accurate pressure gas cylinder structural model, the present invention can evaluate the performance of the pressure gas cylinder under different operating conditions and obtain the bursting pressure of the gas cylinder under different operating conditions. At the same time, by comparing the similarity between the bursting installation environment image and the initial image representation diagram, the degree of morphological damage of the pressure gas cylinder under the deformation variable conditions can be evaluated, thereby helping managers understand the morphological changes that may occur in the pressure gas cylinder when it is operated for a long time or is subjected to abnormal pressure, and infer potential structural damage. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0022] Figure 1 This is a principle block diagram of a device for monitoring deformation of a pressure gas cylinder in a fire protection system according to an embodiment of the present invention; Figure 2 The present invention is a flowchart of a method for monitoring the deformation of a pressure gas cylinder in a fire protection system according to an embodiment of the present invention.

[0023] In the picture: 1. Pressure gas cylinder body; 2. Deformation monitoring and processing terminal; 201. Deformation prediction unit; 202. Damage degree prediction unit; 3. Gas cylinder monitoring and control terminal. DETAILED DESCRIPTION

[0024] To further illustrate each embodiment, the present invention provides drawings, which are part of the disclosure of the present invention. They are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. By referring to these contents, ordinary technicians in this field should be able to understand other possible implementation methods and the advantages of the present invention.

[0025] According to an embodiment of the present invention, a device and method for monitoring the deformation of a pressure gas cylinder in a fire protection system are provided.

[0026] The present invention will now be further described with reference to the accompanying drawings and specific embodiments. Figure 1 As shown, the deformation monitoring device of the pressure gas cylinder of the fire protection system according to the embodiment of the present invention includes: The pressure gas cylinder body 1 is used to collect stress change signals corresponding to the pressure gas cylinder through sensors.

[0027] It should be explained that the body of the pressure gas cylinder 1 is usually made of high-strength steel or composite materials and must pass pressure resistance and sealing tests. The valve device mainly includes a start valve (electric / manual) and a release valve to control the release of the fire extinguishing agent. The sensor monitors the stress changes in the cylinder in real time to ensure that the changes are within the normal range. The safety pressure relief device prevents the cylinder from exploding due to overpressure or high temperature. The siphon / release pipe ensures that the fire extinguishing agent is efficiently sprayed in liquid or gaseous form. Specifically, when collecting the stress change signal corresponding to the pressure gas cylinder through the sensor, the strain sensor at the threaded connection between the middle part of the bottle body and the bottle mouth of the fire protection system pressure gas cylinder is used to obtain the strain change signal data caused by the internal pressure, external environment and loose threads of the fire protection system pressure gas cylinder; the signal testing technology is used to analyze the signal transmission stability of the wireless transmission method, and according to the analysis results, it is judged whether there is signal interference and signal loss problem in the transmission process of the strain change signal data; if there is a problem, the transmission parameters of the wireless transmission method are adjusted; if there is no problem, the strain change signal data is uploaded to the characteristic relationship model During the specific implementation, multiple miniature high-precision strain sensors are reasonably and evenly arranged at key stress concentration locations such as the bottle body and bottle mouth threaded connections of the fire system pressure cylinder. The strain sensors can sensitively sense extremely subtle strain changes on the surface of the fire system pressure cylinder and quickly convert them into electrical signal output. Due to the special structure and usage scenarios of the fire system pressure cylinder, the selected miniature strain sensors have the characteristics of small size and strong anti-interference ability, and will not affect the normal use of the fire system pressure cylinder.

[0028] At the same time, the electrical signal output by the strain sensor is transmitted to a dedicated data processing center through a highly flexible, wear-resistant cable (wired method) or a low-power, highly stable wireless module (wireless method). The data processing center has a built-in algorithm (characteristic relationship model) optimized for the characteristics of fire protection system pressure cylinders. It conducts in-depth analysis and processing of the collected strain signals and accurately calculates the deformation of the fire protection system pressure cylinders based on the pre-established characteristic relationship model applicable to the fire protection system pressure cylinders.

[0029] The data processing center will compare the calculated deformation variable with the safety threshold set according to the safety standard of fire protection system pressure cylinders in real time. Once the deformation variable exceeds the safety threshold, the high-decibel sound and light alarm device will be immediately activated, emitting a strong alarm sound and flashing lights to attract the attention of on-site personnel; at the same time, detailed abnormal information, such as the specific deformation variable value, the time of exceeding the standard, etc., will be sent to the management terminal of the fire control center and the mobile devices such as mobile phones of the relevant responsible persons through the Internet of Things communication module, so that timely measures such as replacement of fire protection system pressure cylinders and maintenance can be taken.

[0030] Specifically, after the fire protection system pressure cylinder is manufactured or during regular maintenance, in accordance with strict design requirements, the micro strain sensor is firmly attached to the predetermined position of the fire protection system pressure cylinder using special high-temperature resistant and corrosion-resistant glue. During the pasting process, ensure that the strain sensor is completely and tightly fitted to the surface of the fire protection system pressure cylinder to avoid bubbles, looseness, etc., so as to ensure the accuracy and stability of the measurement data.

[0031] Specifically, three miniature high-precision strain sensors are evenly arranged along the circumferential direction in the middle of the fire-fighting system pressure cylinder, marked as E1, E2, and E3. These three strain sensors 2 can effectively monitor the circumferential strain changes of the bottle body caused by internal pressure and external environmental factors. Two miniature strain sensors E4 and E5 are symmetrically installed at the threaded connection of the bottle mouth. This is the place where stress concentration is prone to occur in the fire-fighting system pressure cylinder. These two strain sensors can accurately capture the strain conditions at the threaded connection of the bottle mouth, preventing safety hazards caused by problems such as loose threads.

[0032] After the strain sensor is installed, first calibrate the strain sensor to ensure that its measurement accuracy meets the requirements. Then test the signal transmission stability of the signal transmission cable or wireless transmitter to check whether there are problems such as signal interference and loss.

[0033] Multiple miniature high-precision strain sensors (E1-E5) serve as data acquisition terminals, transmitting the collected strain electrical signals through signal transmission technology. If a wired method is used, the signal will be connected to the data transmission module through a highly flexible, wear-resistant cable; if a wireless method is used, a low-power, highly stable wireless transmitter will be responsible for signal transmission.

[0034] The deformation monitoring processing terminal 2 is used to predict the deformation of the pressure gas cylinder based on the stress change signal and the characteristic relationship model, and to evaluate the degree of damage to the pressure gas cylinder based on the deformation and the installation environment.

[0035] In one embodiment, the deformation monitoring processing end 2 includes: The deformation prediction unit 201 is used to combine the stress change signal with the characteristic relationship model to predict the deformation of the pressure gas cylinder.

[0036] Specifically, when combining the stress change signal with the characteristic relationship model to predict the deformation of the pressure gas cylinder, the relevant characteristic information between the historical stress data and the historical deformation data of the pressure gas cylinder in the fire protection system can be analyzed; the relevant characteristic information is used to construct a characteristic relationship model with correlation calculation as the core to output the deformation of the pressure gas cylinder.

[0037] Among them, when analyzing the relevant characteristic information between the historical stress data and the historical deformation data of the pressure cylinders in the fire protection system, the historical stress data and the historical deformation data of the pressure cylinders in different time periods during the use of the fire protection system can be obtained, and the historical stress data and the historical deformation data can be time-aligned; based on the alignment processing result, a time window is selected to extract the change relationship between stress and deformation, and a point pair feature is constructed according to the change relationship result, and the edge feature of the point pair feature is analyzed using edge convolution; a full edge connectivity matrix is ​​constructed according to the edge feature and the change relationship, and the connectivity matrix between stress and deformation is analyzed based on the full edge connectivity matrix; the topological relationship between the historical stress data and the historical deformation data is analyzed based on the connectivity matrix, and the relevant characteristic information between stress and deformation is analyzed based on the topological relationship.

[0038] At the same time, when constructing a full edge connectivity matrix based on edge features and change relationships, and analyzing the connectivity matrix between stress and deformation based on the full edge connectivity matrix, a node relationship search can be performed on the edge features and change relationships, and a node association table can be generated based on the search results. The node association table is used to establish an adjacency matrix of the connection relationship, and the adjacency relationship is multiplied by itself; the self-multiplication result is squared to obtain the full edge connectivity matrix, and the continuous variables that are positively correlated with the edge features and change relationships are identified based on the full edge connectivity matrix; a multiple linear regression equation with continuous variables as dependent variables is established, and the ridge regression technique is combined to analyze the continuous variables of the regression equation composed of historical stress data and historical deformation data in different time periods; all continuous variables are combined, and the connectivity matrix between stress and deformation is generated based on the combination result.

[0039] Among them, when using relevant feature information to construct a feature relationship model with correlation calculation as the core, the correlation feature information can be combined with the self-attention optimization technology to generate a feature vector, and by calculating the correlation between the feature vectors, the weight coefficient of the correlation feature information can be obtained; the attention feature of the correlation feature coefficient is obtained based on the weight coefficient, and the attention feature is combined with the cross-attention technology to generate an optimized feature, and the correlation matrix is ​​obtained according to the optimized feature; the correlation matrix is ​​combined with the correlation model reconstruction technology to perform correlation model reconstruction processing, and the feature relationship model is obtained based on the processing results.

[0040] In order to facilitate understanding of the above technical solutions of the present invention, the working principle of the present invention for predicting the deformation amount of a pressure gas cylinder in an actual process is described in detail below.

[0041] Step 1: Data collection; Five miniature high-precision strain sensors (E1-E5) are arranged on the fire protection system pressure cylinder 1. The specific arrangement is as follows: E1, E2, and E3: Installed in the middle of the bottle, evenly distributed around the circumference of the bottle at intervals of 120°, they are used to monitor the circumferential strain of the bottle caused by internal pressure and external environmental factors. The specific coordinates of the installation locations are: E1: 50mm from the bottom of the bottle; E2: 50mm from the bottom of the bottle; E3: 50mm from the bottom of the bottle; E4 and E5: Installed symmetrically at the threaded connection of the bottle mouth to monitor stress concentration at the threaded connection. The installation position of the threaded connection is: E4 and E5 are both installed at the bottle mouth thread, 15mm from the outer edge of the bottle mouth; The sensor size is 6mm in diameter and 1mm thick for each sensor, which is suitable for close installation. The measurement accuracy of the strain sensor is ±0.1%. The sensor has strong anti-electromagnetic interference ability and is suitable for high noise environment. At the same time, data transmission is carried out after signal transmission test.

[0042] Step 2: Extract change relations; Historical stress and deformation data are collected every hour. The stress data range is 0-100 MPa (based on the internal pressure of the gas cylinder). The collection interval of each data point is 1 hour. The deformation data range is 0-0.5 mm (fluctuates with pressure and environmental changes). The historical stress data and historical deformation data are aligned so that each stress data point corresponds to a deformation data point. A 24-hour time window is selected to extract the relationship between stress and deformation changes during this period: The change relationship analysis is performed every 24 hours in the time window size; when the stress change is greater than 5MPa per hour in the change rate, the strain change is greater than 0.05mm and the edge convolution analysis is performed based on the change relationship between stress and deformation. The convolution kernel size uses a 3×3 convolution kernel for edge feature extraction, and the edge feature extraction accuracy is ±0.02mm.

[0043] Step 3: Connection matrix construction; A node relationship search algorithm based on graph theory is used to construct the topological relationship between stress and deformation. A time period of 1 hour is selected for each search to ensure that high dynamic changes can be captured. A maximum of 100 data points are captured as nodes in each time window. Based on the node relationship search results, an adjacency matrix is ​​generated. Assuming that the similarity between nodes is greater than 0.8, it is considered that there is a connection relationship. The size of the adjacency matrix is ​​a 5×5 matrix (i.e., 5 nodes). The connection matrix uses 1 to represent connection and 0 to represent disconnection. The adjacency matrix is ​​multiplied by itself to obtain the full edge connectivity matrix. After calculating the full edge connectivity matrix, a 6×6 matrix is ​​obtained, which represents the connectivity relationship between each node; The self-attention mechanism is used to optimize the feature vector to calculate the correlation between each stress and deformation data point. The length of each feature vector is 64 dimensions to ensure that the model can cover enough information. The feature vector is optimized through the cross-attention technology to calculate the weight coefficient of each feature. Based on the correlation between the features, the attention coefficient of each feature is extracted, which ranges from 0 to 1.

[0044] Step 4: Constructing the characteristic relationship model; Based on the attention coefficient, a feature correlation matrix is ​​generated for model optimization. The matrix size is 8×8, which represents the relationship between features. The ridge regression technique is used to generate the regression equation: k = z 0+ z 1 e 1+ z 2 e 2+…+ z n e n Where, Y represents the shape variable, z n represents the regression coefficient, e n Represents stress data; The current deformation is calculated based on the output. Assume that the safety deformation threshold of the pressure cylinder is set to 0.1mm. If the deformation exceeds 0.1mm, an alarm is triggered.

[0045] Therefore, by obtaining the historical stress data and deformation data of the fire protection system in different time periods, a characteristic relationship model is constructed to output the deformation of the pressure gas cylinder, which is convenient for later diagnosis of potential problems in the pressure gas cylinder, taking repair measures in advance, reducing equipment failure rate, and improving the safety and stability of the fire protection system.

[0046] The damage degree prediction unit 202 is used to analyze the fatigue life and burst pressure of the pressure gas cylinder based on the deformation variable, and predict the damage degree of the pressure gas cylinder in combination with the installation environment of the pressure gas cylinder.

[0047] In one embodiment, when analyzing the fatigue life and bursting pressure of a pressure cylinder based on deformation variables and predicting the degree of damage to the pressure cylinder in combination with the installation environment of the pressure cylinder, the initial installation environment, structure, and material information of the fire protection system pressure cylinder can be obtained, and a pressure cylinder structural model can be generated in combination with the deformation variables, and the pressure cylinder structural model can be obtained as an initial image representation diagram; an experimental factor table of the pressure cylinder structural model is set using orthogonal experimental technology, and the fatigue life of the pressure cylinder is obtained based on the experimental factor table and the life curve analysis method, and a damage accumulation analysis is performed on the pressure cylinder based on the damage criterion to obtain damage accumulation; based on the fatigue life and loss accumulation, the bursting pressure of the pressure cylinder under different working conditions is predicted, and an image of the bursting installation environment of the pressure cylinder when the bursting pressure is reached is obtained; the similarity between the bursting installation environment image and the initial image representation diagram is compared, and the degree of morphological damage of the pressure cylinder under deformation variable conditions is evaluated based on the similarity results.

[0048] It should be explained that when generating the pressure cylinder structural model, it is necessary to obtain the initial environment, structure, and material information of the fire protection system pressure cylinder. This information will help understand the behavioral characteristics of the cylinder under normal working conditions: Record the initial operating environment temperature of the pressure gas cylinder, which is usually required to be at a standard ambient temperature (such as 20°C to 25°C). At the same time, use a temperature sensor to monitor the installation environment of the gas cylinder in real time and record the initial temperature. The installation location of the pressure gas cylinder may affect the external load and temperature changes it withstands, so it is necessary to record its installation location (such as indoors, basement, rooftop, etc.) and take into account the differences in the external environment at different locations.

[0049] Record the key structural dimensions of the gas cylinder, including the diameter, length, mouth size, and wall thickness of the cylinder. These dimensional data are used to build the structural model. Specifically: Assume that the cylinder body diameter is 300mm, the cylinder body length is 800mm, the bottle mouth diameter is 50mm, and the bottle body wall thickness is 5mm. Based on the structural type of the gas cylinder, determine whether it is cylindrical, conical, or other shapes. Record the thread type, pitch, and other information of the bottle mouth for calculating the stress concentration area. Assume that the thread is M50×2mm.

[0050] Record the material of the gas cylinder, such as steel, composite material, aluminum alloy, etc. Gas cylinders of different materials have different stress, deformation, and life characteristics. Assume that the gas cylinder is made of high-strength steel with a tensile strength of 500 MPa, a compressive strength of 600 MPa, and an elastic modulus of 200 GPa. Record the material thickness of the gas cylinder wall to facilitate subsequent calculation of stress and deformation in finite element analysis.

[0051] By combining the acquired installation environment, structure, and material information with deformation data, finite element analysis (FEA) can be used to generate a structural model of the pressure cylinder, thereby predicting its deformation and stress distribution. Specifically, computer-aided design software can be used to create a geometric model of the cylinder. Based on the acquired structural dimensions and shape parameters, a high-precision geometric model can be established.

[0052] During the modeling process, the influence of the initial ambient temperature on the cylinder structure is taken into account. Combined with the deformation data of the pressure cylinder (for example, the deformation measured by the strain sensor), these deformation data are combined with finite element analysis as one of the boundary conditions of the deformation, making it easier to predict the stress and deformation of the pressure cylinder under different working conditions through the deformation.

[0053] Specifically, the orthogonal test technology is used to set the test factor table of the pressure gas cylinder structural model, the fatigue life of the pressure gas cylinder is obtained according to the test factor table and the life curve analysis method, and the damage accumulation analysis of the pressure gas cylinder is performed based on the damage criterion. In the process of obtaining the damage accumulation, some factors that have a greater impact on the fatigue life can be selected as test factors according to the design structure and working environment of the pressure gas cylinder. The factors usually include: pressure (the pressure inside the gas cylinder), temperature (the temperature of the working environment), vibration frequency (vibration conditions in the working environment of the gas cylinder), material properties (the type of material used in the gas cylinder) (such as high-strength steel, aluminum alloy, etc.), wall thickness (the wall thickness of the gas cylinder directly affects its load-bearing capacity) and stress concentration at the threaded connection (the structure and stress state of the threaded connection at the bottle mouth).

[0054] For each test factor, select several level values ​​that should cover the possible working range of the cylinder. For example: Pressure: select 3 levels, such as 0.5MPa, 1.0MPa, 1.5Mpa; Temperature: select 3 levels, such as 20°C, 40°C, 60°C; Vibration frequency: select 3 levels, such as 10Hz, 20Hz, 30Hz; Material properties: Choose different types of materials, taking into account different tensile strengths (e.g. steel, aluminum alloy).

[0055] Using orthogonal experimental design (such as L9 orthogonal table), all experimental factors and their levels are combined into an experimental plan. Taking L9 orthogonal table as an example, the experimental design table is shown in Table 1: Table 1: Experimental design table According to the test combination of the orthogonal test table, each test is performed one by one to obtain the working status data of the gas cylinder under different factors, including the stress, deformation, life, etc. of each test.

[0056] After the orthogonal test is completed, the fatigue life curve of the pressure gas cylinder is fitted by fitting the experimental data. The life curve reflects the relationship between the fatigue life of the gas cylinder under specific working conditions and factors such as stress and temperature. Commonly used life curves include: SN curve: represents the fatigue life of the gas cylinder after n cycles under different stress amplitudes (S); stress-strain life curve: combines the stress-strain characteristics of the material to calculate the life. By analyzing the results of the orthogonal test, the relationship between the changes in factors such as stress and temperature and the life is fitted to obtain the life curve.

[0057] Based on the life curve, the fatigue life of the cylinder is calculated under different operating conditions (such as different pressures, temperatures, vibration frequencies, etc.). Based on experimental data, the SN curve or stress-strain life curve is used to obtain the expected service life of the cylinder under different conditions.

[0058] Damage accumulation analysis primarily assesses fatigue damage in gas cylinders based on damage criteria. Common damage criteria include: Miner's Rule: The Miner's Rule is used for fatigue damage accumulation under multiple loads. This rule assumes that damage accumulates linearly, meaning that the damage values ​​at each loading stage are summed.

[0059] Based on orthogonal experiments and life curves, information such as stress amplitude and number of load cycles under different loading conditions is obtained. For each loading condition, the corresponding damage value is calculated. The damage values ​​of each condition are accumulated to obtain the total damage value of the gas cylinder during the entire service life.

[0060] Based on the cumulative damage value, the fatigue failure time of the gas cylinder is predicted. If the total damage value reaches 1 or higher, it means that the life of the gas cylinder has been exhausted and maintenance, replacement or repair is required. The final results of the damage accumulation analysis include the following aspects: Total damage value: The cumulative damage value calculated by the mineral criterion is used to determine whether the cylinder has approached or reached the fatigue limit; Fatigue failure time: The fatigue failure time of the cylinder is predicted by the damage accumulation value, which usually results in a specific time range or cumulative load cycle; Safety measures that need to be taken: If the damage value of the gas cylinder is close to or exceeds 1, relevant safety measures need to be taken, such as replacing the gas cylinder, regular maintenance, etc.

[0061] Orthogonal test technology is used to set the test factor table for pressure gas cylinders. The fatigue life of the cylinders can be calculated through life curve analysis, and damage accumulation analysis is performed based on the mineral criterion. This analysis method can help predict the fatigue life of pressure gas cylinders under different working conditions, and promptly identify potential fatigue failure risks, ensuring the safe use and timely maintenance of pressure gas cylinders.

[0062] It should be explained that in the process of predicting the burst pressure of pressure gas cylinders under different operating conditions based on fatigue life and loss accumulation, and obtaining an image of the burst installation environment of the pressure gas cylinder when the burst pressure is reached, factors affecting the burst pressure (the burst pressure is the pressure value when the pressure gas cylinder fails due to fatigue or overload) can be determined. Factors affecting the burst pressure include: The material, wall thickness, operating environment temperature and temperature fluctuations, aging degree and usage time of the gas cylinder, internal and external pressure fluctuations, and the bottle mouth structure and connection of the gas cylinder; based on the previous steps, the fatigue life of the gas cylinder is predicted through the life curve analysis method, and the fatigue life data of the gas cylinder under different operating conditions is obtained. The fatigue life curve and damage accumulation criterion will help evaluate the behavior of the gas cylinder under multiple load cycles and the formation of fatigue cracks.

[0063] Based on the design parameters of the gas cylinder (such as size, material, wall thickness, etc.), stress analysis of the pressure gas cylinder is carried out. The finite element analysis (FEA) tool can be used to simulate the gas cylinder structure and evaluate the stress distribution of the gas cylinder under different load and environmental conditions. Through the stress analysis of the gas cylinder, the maximum safe pressure that the gas cylinder can withstand is calculated. The bursting pressure of the gas cylinder is close to its yield strength or ultimate failure stress, and the gas cylinder is located at the bottle mouth, threaded connection, weak area of ​​the bottle body, etc.

[0064] The bursting pressure of the gas cylinder is estimated by combining the tensile strength, yield strength, fatigue strength and other data of the gas cylinder material. For example: Tensile strength: Assuming it is 600MPa, the gas cylinder explosion usually occurs when the tensile strength is close to it; Yield strength: Assuming it is 500MPa, permanent deformation will occur near the yield strength, and further increasing the load may cause explosion.

[0065] Temperature and aging can affect the strength performance of gas cylinders. The fatigue life curve and burst pressure prediction of the material can be adjusted according to the ambient temperature of the gas cylinder (such as high temperature, low temperature, etc.): Temperature correction factor: Temperature affects the material strength of the gas cylinder, and the model needs to be adjusted according to changes in ambient temperature; Aging effect: Over time, the material of the gas cylinder may lose strength due to oxidation, corrosion or fatigue accumulation, and the degree of material aging needs to be considered.

[0066] Based on the mineral criterion or other damage criteria, a damage accumulation analysis is performed on the gas cylinder during use: considering the load fluctuations experienced by the gas cylinder under different working conditions, the damage values ​​at different working stages are calculated.

[0067] Fatigue failure prediction: By calculating the cumulative damage value, the fatigue damage degree of the gas cylinder when it reaches the bursting pressure is predicted. If the total damage value reaches or exceeds 1, it indicates that the gas cylinder may be about to explode. Based on the cumulative fatigue damage and crack growth model, the growth of fatigue cracks is predicted, and the bursting pressure is estimated using crack growth theory and critical crack size.

[0068] Once the bursting pressure of the gas cylinder is determined, the following steps can be used to generate an image of the pressure gas cylinder explosion installation environment: simulate the temperature conditions when the explosion occurs, assuming a high temperature environment (for example, 60°C); the surrounding environment of the gas cylinder: vibration or other dynamic loads can be set in the surrounding area; the working pressure and bursting pressure of the gas cylinder: determine the working pressure and bursting pressure of the gas cylinder, assuming the working pressure is 50MPa and the bursting pressure is 70MPa.

[0069] Finite element analysis or fluid dynamics simulation (e.g., CFD analysis) is used to evaluate the stress distribution and deformation of the cylinder during a rupture. The crack propagation and local deformation areas of the cylinder are analyzed. Combined with the above analysis, an image of the environment during a pressure cylinder rupture is generated: Time series of the explosion process: showing the deformation of the gas cylinder before and after the explosion; Stress concentration area: shows the area where stress is most concentrated before bursting (such as the weak area of ​​the bottle mouth or bottle body); Temperature Effect: Corrected by ambient temperature, presents the effect of temperature distribution on the burst pressure of gas cylinders.

[0070] Images can be generated using computer-aided design (CAD) or finite element analysis (FEA) software, or they can be simulated using physical experiments to generate more realistic images of the blasting environment.

[0071] Through fatigue life prediction and damage accumulation analysis, combined with information such as the cylinder's design, material, and working environment, the burst pressure of a pressure cylinder can be accurately predicted and verified through stress analysis, fatigue damage analysis, and temperature impact analysis. Ultimately, based on these analysis results, an image of the installation environment at the time of cylinder explosion can be generated, helping to assess the safety of the cylinder and providing a basis for preventing explosion accidents.

[0072] Specifically, when comparing the similarity between the blasting installation environment image and the initial image representation map, and evaluating the degree of morphological damage of the pressure gas cylinder under deformation variable conditions based on the similarity results, two reference points can be selected on the main axis of the initial image representation map, the frequency map between the initial image representation map and the reference points can be analyzed, and the two sets of frequency maps can be merged and normalized to obtain a direction vector; the representation feature parameters of the initial image representation map are extracted according to the direction vector, and a feature set is generated based on the representation feature parameters to construct an image cloud model that can reflect the morphological image; the representation feature parameters corresponding to the blasting installation environment image are input into the image cloud model for similarity analysis to obtain the similarity results between the blasting installation environment image and the initial image representation map; an evaluation system is constructed based on the similarity results and the set indicators to analyze the degree of morphological damage of the pressure gas cylinder under deformation variable conditions.

[0073] Among them, two reference points are selected on the main axis of the initial image representation map, the frequency map between the initial image representation map and the reference points is analyzed, and the two sets of frequency maps are merged for normalization processing to obtain the direction vector. The center of mass position can be analyzed according to the pixel coordinates of the initial image representation map, and the center of mass position and the pixels of the initial image representation map are combined to calculate the covariance matrix, and the eigenvalues ​​and eigenvectors are generated based on the covariance matrix; the rotation angle between the main axis of the initial image representation map and the horizontal axis is determined based on the eigenvalues ​​and eigenvectors to obtain the main axis direction and obtain the distance from each pixel coordinate point in the initial image representation map to the center of mass; the maximum value of the distance result is selected to determine the two reference points on the main axis, and the coordinate values ​​of the two reference points on the initial image representation map are analyzed, and according to the frequency map between the coordinate values ​​and the initial image representation map; the two sets of frequency maps are merged, and normalization, median filtering and noise removal processing are performed, and according to the processing results, spatial features are extracted from the frequency map as the direction vector of the initial image representation map.

[0074] At the same time, when extracting the representation feature parameters of the initial image representation map according to the direction vector, and generating a feature set based on the representation feature parameters to construct an image cloud model that can reflect the morphological image, the grayscale co-occurrence matrix in different directions can be selected from the direction vector, and the entropy and correlation characteristics of the grayscale co-occurrence matrix can be analyzed to extract the representation feature parameters of the initial image representation map; the representation feature parameters and the central node generated by the inverse cloud generator reflect the membership of the direction vector to the initial image representation map, and a preset number of representation feature parameters are screened out according to the membership results to generate a feature set; the mean and variance of the feature set are judged, and the expected value and information entropy of the representation feature parameters corresponding to the morphological image can be reflected according to the analysis of the mean difference and variance results, and the representation feature parameters are converted into an image cloud model based on the expected value and information entropy results.

[0075] The calculation formula of the correlation feature is: ; Where, L Represents the correlation characteristics, T represents the gray level of the gray level co-occurrence matrix, P ( x , y ) indicates that the gray levels of the two pixels are x and y The probability of x and y Both represent the gray levels in the gray-level co-occurrence matrix, f 1 represents the standard deviation of the gray-level co-occurrence matrix along the horizontal direction, f 2 represents the standard deviation of the gray-level co-occurrence matrix along the vertical direction, a Represents the statistical step size of the gray-level co-occurrence matrix.

[0076] The calculation formula for characterizing the information entropy of characteristic parameters is: ; Where, W represents the information entropy of the characteristic parameters, N represents the total number of feature sets, r m Indicates the m The characteristic parameter corresponding to the feature set is Indicates the m The mean of the characteristic parameter corresponding to the feature set, Represents the mean of the feature set.

[0077] The gas cylinder monitoring and control terminal 3 is used to generate terminal alarm prompt rules according to the damage degree results, and the management terminal adjusts the use status of the pressure gas cylinder according to the prompt rules.

[0078] In one embodiment, when a terminal alarm prompt rule is generated based on the damage degree result and the management terminal adjusts the use status of the pressure cylinder according to the prompt rule, the deformation variable monitored in real time (or the maximum deformation variable calculated by simulation) can be compared with the safety threshold to obtain a difference. If the difference is ≥0, it means that the deformation variable of the cylinder does not exceed the safety threshold and the cylinder is still within the safe working range; if the difference is <0, it means that the deformation variable of the cylinder exceeds the safety threshold and the cylinder may have entered a dangerous area and is at risk of damage or rupture.

[0079] Based on the difference in deformation, the degree of damage to the gas cylinder can be further estimated. For example, if the difference is close to zero, it means that the gas cylinder is close to the safety limit and needs further monitoring. If the difference is negative and large, it means that the gas cylinder may have undergone significant deformation or damage and is close to explosion or failure.

[0080] The degree of damage can be classified into the following levels: Slight damage (difference close to 0): The cylinder deformation is slightly close to the threshold, and the operator needs to be reminded to check and monitor; Moderate damage (the difference is negative but within a certain range): the cylinder deformation has exceeded the safety threshold and measures must be taken as soon as possible; Severe damage (difference greater than the negative threshold): The cylinder deformation has reached or exceeded the critical value. Stop using it immediately and perform maintenance or replacement.

[0081] Based on the difference between the deformation and the safety threshold and the degree of damage, the terminal alarm prompt rules can be set. The specific rules include the following levels: Mild alarm (reminder to check): When the difference is close to 0 or slightly negative (for example, the difference is between 0 and -0.5), the operator is reminded that the cylinder deformation is close to the threshold and regular inspections are recommended. The prompt content is "The cylinder deformation is close to the safety threshold, please perform routine inspections" or "Please check the operating status of the cylinder and monitor the deformation changes."

[0082] Moderate alarm (check immediately): When the difference is negative and in the medium range (for example, the difference is between -0.5 and -1.0).

[0083] Alarm rules: When the deformation of the gas cylinder exceeds the safety threshold, a moderate alarm should be issued, requiring immediate inspection of the gas cylinder status and repair or decompression.

[0084] Prompt content: "The deformation of the gas cylinder exceeds the safety threshold. Please check immediately and reduce the operating pressure" or "Please check the gas cylinder as soon as possible to ensure that there is no damage or deformation to avoid further damage."

[0085] Serious alarm (stop using): When the difference is significantly negative (for example, the difference is less than -1.0); Alarm rules: When the deformation of the gas cylinder is close to the critical value or exceeds the safe working range, the gas cylinder should be stopped from use immediately and arranged for repair or replacement. The prompt content is: "The deformation of the gas cylinder exceeds the serious threshold and is close to failure. Please stop using it immediately" or "The gas cylinder is seriously damaged. It is recommended to replace it immediately and contact maintenance personnel."

[0086] According to the terminal alarm prompt rules, the management terminal automatically adjusts the use status of the pressure gas cylinder according to different alarm levels. The specific steps are as follows: Mild alarm (continue to use, but monitor): continue to use the gas cylinder, but strengthen the monitoring of the gas cylinder and regularly check the deformation, pressure and ambient temperature of the gas cylinder.

[0087] Moderate alarm (reduce usage or reduce pressure), reduce the cylinder's operating pressure, reduce its workload, or perform depressurization operations where appropriate, increase the frequency of cylinder inspections, especially in areas with large deformation, reduce usage time, and consider repairing or replacing.

[0088] Serious alarm (stop using, repair or replace), stop using the gas cylinder immediately. The gas cylinder may be in a dangerous state and must be inspected or replaced. Carry out gas cylinder inspection, testing and safety assessment, and prepare to replace the gas cylinder. The operator can directly receive an emergency stop command.

[0089] Therefore, by comparing the difference between the deformation variable and the safety threshold, and combining the size of the difference with the degree of damage, an alarm prompt rule is generated. The management terminal adjusts the use status of the gas cylinder in time according to the alarm rule, which can effectively prevent the pressure gas cylinder from exploding or failing and ensure safe operation.

[0090] like Figure 2 As shown, according to another embodiment of the present invention, a method for monitoring the deformation of a pressure gas cylinder in a fire protection system is also provided, the method comprising: S1, collects the stress change signal corresponding to the pressure cylinder through the sensor; S2, predicting the deformation of the pressure gas cylinder based on the stress change signal and the characteristic relationship model, and evaluating the damage degree of the pressure gas cylinder based on the deformation and installation environment; S3, generating terminal alarm prompt rules according to the damage degree result, and the management terminal adjusts the use status of the pressure gas cylinder according to the prompt rules.

[0091] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A deformation monitoring device for pressure gas cylinders in a fire protection system, characterized in that: The device includes: The pressure gas cylinder body is used to collect stress change signals corresponding to the pressure gas cylinder through sensors; The deformation monitoring and processing end is used to predict the deformation of the pressure gas cylinder based on the stress change signal and the characteristic relationship model, and to assess the degree of damage to the pressure gas cylinder based on the deformation and installation environment; Wherein, the deformation monitoring processing end includes: The deformation prediction unit is used to combine the stress change signal with the characteristic relationship model to predict the deformation of the pressure gas cylinder; it includes: Analyze the correlation characteristics between the historical stress data and historical deformation data of pressure cylinders in fire protection systems; Utilize relevant feature information to build a characteristic relationship model with correlation calculation as the core, and output the deformation of the pressure gas cylinder; A damage degree prediction unit is used to analyze the fatigue life and burst pressure of the pressure gas cylinder based on the deformation variable, and to predict the damage degree of the pressure gas cylinder in combination with the installation environment of the pressure gas cylinder; The gas cylinder monitoring and control terminal is used to generate terminal alarm prompt rules based on the damage degree results, and the management terminal adjusts the use status of the pressure gas cylinder according to the prompt rules.

2. A fire protection system pressure gas cylinder deformation monitoring device according to claim 1, characterized in that: The collecting of stress change signals corresponding to the pressure cylinder by the sensor includes: Using the strain sensor at the threaded connection between the middle of the fire-fighting system pressure gas cylinder and the bottle mouth, the strain change signal data caused by the internal pressure, external environment and loose threads of the fire-fighting system pressure gas cylinder is obtained; Signal testing technology is used to analyze the signal transmission stability of the wireless transmission method, and based on the analysis results, it is determined whether there are signal interference and signal loss problems in the transmission process of the strain change signal data; If there is a problem, the transmission parameters of the wireless transmission method are adjusted; if there is no problem, the strain change signal data is uploaded to the characteristic relationship model.

3. A fire protection system pressure gas cylinder deformation monitoring device according to claim 1, characterized in that: The analysis of the relevant characteristic information between the historical stress data and the historical deformation data of the pressure cylinders in the fire protection system includes: Obtain historical stress data and historical deformation data of pressure cylinders in different time periods during the use of the fire protection system, and perform time alignment processing on the historical stress data and historical deformation data; Based on the alignment processing results, a time window is selected to extract the change relationship between stress and deformation. Point pair features are constructed based on the change relationship results, and edge features of the point pair features are analyzed using edge convolution. According to the edge characteristics and change relationship, a full edge connectivity matrix is ​​constructed, and the connectivity matrix between stress and deformation is analyzed based on the full edge connectivity matrix; The topological relationship between historical stress data and historical deformation data is analyzed based on the connectivity matrix, and the relevant characteristic information between stress and deformation is analyzed according to the topological relationship.

4. A fire protection system pressure cylinder deformation monitoring device according to claim 1, characterized in that: According to the edge characteristics and change relationship, a full edge connectivity matrix is ​​constructed, and the connectivity between stress and deformation is analyzed based on the full edge connectivity matrix. The connection matrix includes: Perform node relationship search on edge features and change relationships, generate a node association table based on the search results, use the node association table to establish an adjacency matrix of the connection relationship, and multiply the adjacency relationship by itself; The full edge connectivity matrix is ​​obtained by squaring the self-multiplied result, and the continuous variables whose edge features are positively correlated with the change relationship are identified based on the full edge connectivity matrix; Establish a multiple linear regression equation with continuous variables as dependent variables, and use ridge regression technology to analyze the historical stress data and historical deformation data in different time periods to form the continuous variables of the regression equation; All continuous variables are combined, and the connection matrix between stress and deformation is generated based on the combination results.

5. The deformation monitoring device for pressure gas cylinders in fire protection systems according to claim 1, characterized in that: The characteristic relationship model based on correlation calculation is constructed by using relevant feature information to output the deformation variables of the pressure gas cylinder, including: The correlation feature information is combined with the self-attention optimization technology to generate feature vectors, and the weight coefficient of the correlation feature information is obtained by calculating the correlation between the feature vectors; The attention feature of the correlation feature coefficient is obtained based on the weight coefficient, and the attention feature is combined with the cross attention technology to generate an optimized feature, and the correlation matrix is ​​obtained according to the optimized feature; The correlation matrix is ​​combined with the correlation model reconstruction technology to perform correlation model reconstruction processing, and a characteristic relationship model is obtained based on the processing results.

6. The deformation monitoring device for pressure gas cylinders in fire protection systems according to claim 1, characterized in that: The analysis of the fatigue life and burst pressure of the pressure gas cylinder based on the deformation variable and the prediction of the damage degree of the pressure gas cylinder in combination with the installation environment of the pressure gas cylinder include: Obtain the initial installation environment, structure, and material information of the fire protection system pressure cylinder, and generate a pressure cylinder structural model based on the deformation variables, and obtain the pressure cylinder structural model as the initial image representation diagram; The orthogonal test technology is used to set the test factor table of the pressure cylinder structure model. The fatigue life of the pressure cylinder is obtained based on the test factor table and the life curve analysis method. The damage accumulation analysis of the pressure cylinder is performed based on the damage criterion to obtain the damage accumulation. Predict the burst pressure of pressure gas cylinders under different working conditions based on fatigue life and loss accumulation, and obtain images of the explosion installation environment of pressure gas cylinders when the burst pressure is reached; The similarity between the blasting installation environment image and the initial image representation is compared, and the degree of morphological damage of the pressure gas cylinder under deformation conditions is evaluated based on the similarity results.

7. A fire protection system pressure cylinder deformation monitoring device according to claim 6, characterized in that: The similarity between the blasting installation environment image and the initial image representation image and evaluating the degree of morphological damage of the pressure gas cylinder under the deformation condition based on the similarity result include: Select two reference points on the main axis of the initial image representation, analyze the frequency graph between the initial image representation and the reference points, and combine the two sets of frequency graphs for normalization to obtain a direction vector; Extracting the representation feature parameters of the initial image representation map according to the direction vector, and generating a feature set based on the representation feature parameters to construct an image cloud model that can reflect the morphological image; Inputting the representation feature parameters corresponding to the blasting installation environment image into the image cloud model for similarity analysis, and obtaining the similarity results between the blasting installation environment image and the initial image representation map; An evaluation system is constructed based on the similarity results and the set indicators to analyze the degree of morphological damage of the pressure cylinder under deformation conditions.

8. The deformation monitoring device for pressure gas cylinders in fire protection systems according to claim 7, characterized in that: The step of selecting two reference points on the main axis of the initial image representation graph, analyzing the frequency graph between the initial image representation graph and the reference points, and merging the two sets of frequency graphs for normalization to obtain a direction vector includes: Analyze the centroid position according to the pixel coordinates of the initial image representation map, combine the centroid position with the pixels of the initial image representation map to calculate the covariance matrix, and generate eigenvalues ​​and eigenvectors based on the covariance matrix; Determine the rotation angle between the main axis and the horizontal axis of the initial image representation based on the eigenvalue and the eigenvector, obtain the main axis direction, and obtain the distance from each pixel coordinate point in the initial image representation to the center of mass; Select the maximum value of the distance result to determine the two reference points on the main axis, analyze the coordinate values ​​of the two reference points on the initial image representation map, and use the frequency map between the coordinate values ​​and the initial image representation map; The two sets of frequency maps are merged, and normalized, median filtered and noise removed. According to the processing results, spatial features are extracted from the frequency maps as the direction vectors of the initial image representation map.

9. The deformation monitoring device for pressure gas cylinders in fire protection systems according to claim 8, characterized in that: The step of extracting the representation feature parameters of the initial image representation graph according to the direction vector and generating a feature set based on the representation feature parameters to construct an image cloud model that can reflect the morphological image includes: Selecting gray-level co-occurrence matrices in different directions from the direction vector, analyzing the entropy and correlation characteristics of the gray-level co-occurrence matrix, and extracting the representation feature parameters of the initial image representation map; The representation feature parameters are combined with the membership of the central node reflecting the direction vector generated by the reverse cloud generator to the initial image representation map, and a preset number of representation feature parameters are selected based on the membership results to generate a feature set; Determine the mean and variance of the feature set. According to the analysis of the mean and variance results, the expected value and information entropy of the characteristic parameters corresponding to the morphological image can be reflected. Based on the expected value and information entropy results, the characteristic parameters are converted into an image cloud model.

10. A method for monitoring the deformation of a pressure gas cylinder in a fire protection system, implemented by using a device for monitoring the deformation of a pressure gas cylinder in a fire protection system according to any one of claims 1 to 9, characterized in that: The method includes: The stress change signal corresponding to the pressure gas cylinder is collected through the sensor; Predict the deformation of the pressure gas cylinder based on the stress change signal and characteristic relationship model, and evaluate the degree of damage to the pressure gas cylinder based on the deformation and installation environment; The terminal alarm prompt rules are generated according to the damage degree results, and the management terminal adjusts the use status of the pressure gas cylinder according to the prompt rules.

Images