A method and system for intelligent monitoring and early warning of structural state of a storage tank

Abstract

This invention provides a method and system for intelligent monitoring and early warning of the structural performance of storage tanks. The method includes: deploying a monitoring chain on the tank wall, the monitoring chain including multiple tilt sensors and liquid level sensors, for real-time acquisition of tilt angle data and liquid level data of the tank wall; calculating multiple characteristic data characterizing the structural performance of the storage tank based on the tilt angle data and liquid level data, the characteristic data including structural load-bearing capacity data and fatigue state data; outputting an overall structural performance health index of the storage tank using an intelligent prediction model based on the multiple characteristic data; classifying the structural performance of the storage tank into different levels according to the numerical range of the overall structural performance health index, and triggering corresponding early warning and maintenance decisions according to different levels. By real-time monitoring of the structural performance of the storage tank, multi-dimensional feature fusion, and intelligent early warning decision-making, the limitations of traditional manual inspection are overcome, enabling early detection of hidden dangers and precise intervention, effectively reducing the risk of accidents.

Description

Technical Field

[0001] This invention relates to the field of structural health monitoring and facility integrity management technology, specifically to an intelligent monitoring and early warning method and system for the structural performance of storage tanks. Background Technology

[0002] Large vertical storage tanks are widely used in petroleum, chemical, and energy industries, and their structural safety is directly related to enterprise production and public safety. During long-term operation, these tanks face various structural degradation risks, including: overall tilting due to uneven foundation settlement, localized indentation caused by external impacts, tank wall bulging due to hydrostatic pressure, and material strength reduction caused by corrosion and fatigue. These deformation modes are coupled and dynamically evolve, posing a serious threat to the long-term safety of the storage tanks.

[0003] Currently, tank structure monitoring mainly relies on traditional methods such as manual measurement, hydrostatic levels, and GPS positioning. Manual measurement is time-consuming, has low accuracy, and struggles to capture dynamic deformation processes; hydrostatic levels can only monitor settlement and cannot detect localized deformation of the tank wall; while GPS can achieve continuous monitoring, it cannot detect the internal stress state of the tank. Although tilt sensors have been used to measure structural tilt, current technology only uses them as single-parameter acquisition devices, lacking systematic data fusion and intelligent analysis capabilities. This prevents a comprehensive assessment from discrete measurement points to the overall structural condition, and further hinders the quantitative prediction of the tank's remaining lifespan and health degradation trends. Therefore, a new monitoring technology solution is urgently needed that can integrate multi-source monitoring data to achieve intelligent perception of structural condition and health assessment. Summary of the Invention

[0004] The purpose of this invention is to solve the problems of existing tank monitoring technologies, such as limited monitoring dimensions, isolated data, lack of intelligent analysis capabilities, and inability to quantitatively assess structural health status and predict remaining lifespan.

[0005] A first aspect of the present invention provides a method for intelligent monitoring and early warning of the structural performance of storage tanks, comprising: A monitoring chain is installed on the tank wall, which includes multiple tilt sensors and liquid level sensors to collect tilt angle data and liquid level data of the tank wall in real time. Based on the tilt angle data and the liquid level data, multiple characteristic data characterizing the structural behavior of the storage tank are calculated, including structural load-bearing capacity data and fatigue state data. Based on the aforementioned feature data, an intelligent prediction model is used to output the overall structural health index of the storage tank. Based on the numerical range of the overall structural health index, the structural condition of the storage tank is divided into different levels, and corresponding early warning and maintenance decisions are triggered according to different levels.

[0006] Furthermore, the structural bearing capacity data includes equivalent stiffness and stiffness degradation rate, and the calculation of the structural bearing capacity data includes: Based on the installation height of the tilt sensor and the measured tilt angle value, a deformation curve of the tank wall is constructed. The slope of the deformation curve is integrated, and the maximum radial displacement in the integration result is selected as the deformation amount. Calculate the effective pressure of the liquid on the tank wall based on the liquid level and liquid density; The equivalent stiffness is calculated based on the ratio of the effective pressure to the deformation. The stiffness degradation rate is calculated based on the comparison between the equivalent stiffness and the initial stiffness.

[0007] Furthermore, the stiffness degradation rate includes the relative stiffness change rate and the monthly average stiffness degradation rate; The relative stiffness change rate = 00%; The monthly average stiffness change rate 00%; Among them, K eq，t Let K be the equivalent stiffness of the t-th measurement. eq，0 Let t be the initial equivalent stiffness, and t be time. For time intervals.

[0008] Furthermore, the fatigue state data includes the cumulative amount of fatigue damage, and the calculation process for the cumulative amount of fatigue damage includes: The tank wall curvature is calculated based on the height difference and tilt angle difference between adjacent tilt sensors, and the bending strain is calculated based on the tank wall curvature and tank wall thickness. The stress is calculated based on the bending strain and the material's elastic modulus, and the amplitude and number of stress cycles are statistically analyzed. The equivalent stress amplitude is obtained by correcting the average stress. Calculate the fatigue life corresponding to each stress level based on the stress-fatigue life curve of the material, and calculate the cumulative fatigue damage.

[0009] Furthermore, based on historical monitoring data, the monthly average fatigue damage accumulation rate is calculated, and the remaining fatigue life is also calculated: L 剩余 =(1-D 总 ) / D 月 ; Among them, D 总 For the current cumulative fatigue damage, D 月 This represents the monthly average cumulative rate of fatigue damage.

[0010] Furthermore, the intelligent prediction model is an LSTM neural network model; Input features include the overall tilt angle, maximum relative settlement difference, local maximum curvature change rate, maximum deflection of tank wall, equivalent stiffness, stiffness degradation rate, cumulative fatigue damage, monthly average liquid level, and liquid level fluctuation amplitude of historical data. The output prediction results include the overall settlement, local deformation, equivalent stiffness, fatigue damage, and overall structural health index of the storage tank in the future period.

[0011] Furthermore, the training process of the LSTM neural network model includes: The input features are normalized and transformed to the [0, 1] interval; Historical data is divided into training set, validation set, and test set; Training is performed, and training is stopped when the validation loss does not decrease for several consecutive iterations. The accuracy of predictions is assessed using mean absolute error and root mean square error.

[0012] Furthermore, the overall structural health index HI is divided into the following levels: When HI≥80, the storage tank is in good condition and requires no special maintenance. When 60≤HI<80, the storage tank is in good condition and routine monitoring should be performed. When 40≤HI<60, the storage tank is in an early warning state, the monitoring frequency is increased and a detailed inspection is carried out; When 20≤HI<40, the storage tank is in a dangerous condition, its use should be restricted, and a maintenance plan should be developed. When HI < 20, the storage tank is in a critically dangerous state and should be shut down immediately for emergency treatment.

[0013] A second aspect of the present invention provides an intelligent monitoring and early warning system for the structural performance of storage tanks, comprising: The monitoring module includes a monitoring chain deployed along the tank wall, the monitoring chain including multiple tilt sensors and liquid level sensors, used to collect tilt angle data and liquid level data of the tank wall; The calculation and evaluation module calculates multiple characteristic data representing the structural state of the storage tank based on the tilt angle data and the liquid level data; based on the multiple characteristic data, it outputs the overall structural health index of the storage tank using an intelligent prediction model. The early warning module triggers multi-level early warnings and maintenance decisions based on the overall structural health index.

[0014] Furthermore, it also includes a data transmission module and an application terminal module; The data transmission module is used to transmit the monitoring data to the cloud platform, and the calculation and evaluation module and the early warning module are deployed on the cloud platform; The application terminal module is used to receive and display the early warning signals and maintenance decision suggestions output by the early warning module.

[0015] Compared with existing technologies, this invention has at least the following beneficial effects: By deploying a monitoring chain including tilt sensors and liquid level sensors on the tank wall, dynamic and continuous monitoring of the tank's structural condition is achieved, overcoming the shortcomings of traditional manual inspections, such as long cycles, low accuracy, and inability to capture sudden changes; by transforming the raw monitoring data into four key characteristic parameters—overall settlement, local deformation, load-bearing capacity, and fatigue damage—a mapping relationship from discrete measuring points to the overall structural state is established, solving the problem that traditional methods rely on only a single indicator and cannot comprehensively assess the structural health status; by integrating multi-dimensional characteristic data through an intelligent prediction model to output a quantitative health index, a shift from experience-based judgment to data-driven objective assessment is achieved, avoiding interference from subjective human factors. This enables the timely detection of hidden dangers and intervention measures in the early stages of structural deterioration, effectively reducing the risk of major safety accidents and extending the service life of the tank. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained as provided without creative effort.

[0017] Figure 1 This is a flowchart of an intelligent monitoring and early warning method for the structural performance of storage tanks according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the modules of an intelligent monitoring and early warning system for the structural performance of storage tanks in one embodiment of the present invention; Figure 3 This is a schematic diagram of the intelligent prediction model for tank deformation trends in one embodiment of the present invention.

[0018] Among them, 101-top monitoring ring; 102-bottom monitoring ring; 103-dynamic monitoring chain; 104-locally reinforced monitoring grid; 111-tilt meter; 112-liquid level monitor; 113-tank wall; 114-tank top. Detailed Implementation

[0019] The present invention will now be described in more detail with reference to the accompanying drawings, which illustrate preferred embodiments of the invention. It should be understood that those skilled in the art can modify the invention described herein while still achieving its advantageous effects. Therefore, the following description should be understood as being broadly known to those skilled in the art and is not intended to limit the invention.

[0020] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0021] The invention is described more specifically by way of example in the following paragraphs with reference to the accompanying drawings. The advantages and features of the invention will become clearer as explained below. It should be noted that the drawings are in a very simplified form and use non-precise proportions, and are only used to facilitate and clarify the illustration of the embodiments of the invention.

[0022] Example 1 This embodiment provides an intelligent monitoring and early warning method for the structural performance of storage tanks. Please refer to [link / reference]. Figures 1-3 ,include: A monitoring chain is installed on the tank wall, which includes multiple tilt sensors and liquid level sensors to collect tilt and liquid level data of the tank wall in real time.

[0023] Based on the tilt angle data and the liquid level data, multiple characteristic data that characterize the structural behavior of the storage tank are calculated. The characteristic data includes structural load-bearing capacity data and fatigue state data.

[0024] Based on the aforementioned feature data, an intelligent prediction model is used to output the overall structural health index of the storage tank.

[0025] Based on the numerical range of the overall structural health index, the structural condition of the storage tank is divided into different levels, and corresponding early warning and maintenance decisions are triggered according to different levels.

[0026] Specifically, firstly, a monitoring chain consisting of tilt sensors and liquid level sensors is deployed on the tank wall to collect real-time data on the tank's tilt angle and liquid level. Then, based on this raw data, characteristic parameters reflecting the tank's health status—structural load-bearing capacity data and fatigue state data—are calculated. These parameters characterize the tank's structural behavior from different dimensions. Next, this characteristic data is input into an intelligent prediction model, which comprehensively evaluates and outputs a quantitative "overall structural health index." Finally, based on the value range of this health index, the tank is classified into different safety levels, and corresponding early warning signals and maintenance decision recommendations are automatically triggered for each level. This method achieves an intelligent leap from discrete sensor data to overall structural health assessment: it not only monitors the tank's current deformation state but also predicts future fatigue degradation trends, thus transforming traditional passive periodic inspections into proactive real-time early warnings, providing a scientific and quantitative basis for tank safety management decisions.

[0027] Furthermore, the structural bearing capacity data includes equivalent stiffness and stiffness degradation rate, and the calculation of the structural bearing capacity data includes: Based on the installation height of the tilt sensor and the measured tilt angle value, a deformation curve of the tank wall is constructed. The slope of the deformation curve is integrated, and the maximum radial displacement in the integration result is selected as the deformation amount.

[0028] The effective pressure of the liquid on the tank wall is calculated based on the liquid level and liquid density.

[0029] The equivalent stiffness is calculated based on the ratio of the effective pressure to the characteristic deformation.

[0030] The stiffness degradation rate is calculated based on the comparison between the equivalent stiffness and the initial stiffness.

[0031] Specifically, structural load-bearing capacity data is a key indicator for assessing the ability of a storage tank structure to resist external loads. Equivalent stiffness reflects the overall deformation resistance of the storage tank in its current state, while stiffness degradation rate quantifies the degree to which this deformation resistance changes over time or with usage. By introducing these two specific indicators, we can more precisely characterize the structural health of the storage tank, from macroscopic overall stiffness to microscopic performance degradation trends.

[0032] In this embodiment, the method for calculating the tank deformation displacement by the tank wall inclination angle can be the Lagrange interpolation method. First, the deformation is characterized by the amount of bulging; in this embodiment, the deformation is characterized by the amount of bulging. Then, the slope function of the vertical profile curve of the tank wall is established. Based on the inclination angle measured at height hi Then we have: Assuming n sensors are placed at different heights on the tank wall, a polynomial of degree n-1 can be determined using the Lagrange difference method: .

[0033] in: .

[0034] The equation for the deformation curve of the tank wall profile can be obtained by approximating the integral using the slope function, as follows: .

[0035] The maximum radial displacement is selected as the characterization of the bulge amount: .

[0036] Next, we calculate the effective pressure Peff, and the liquid pressure distribution at different heights z: ; Where: ρ is the liquid density (kg / m³) 3 g is the acceleration due to gravity (9.81 m / s²). 2 H represents the actual liquid level height (m).

[0037] Calculate the equivalent stiffness: ; Calculate the relative stiffness degradation rate: ; Monthly average stiffness degradation rate η m : ;where K eq，t Let K be the equivalent stiffness of the t-th measurement. eq，0 Let t be the initial equivalent stiffness, and t be time. For time intervals.

[0038] Furthermore, the fatigue state data includes the cumulative amount of fatigue damage, and the calculation process for the cumulative amount of fatigue damage includes: The tank wall curvature is calculated based on the height difference and tilt angle difference between adjacent tilt sensors, and the bending strain is calculated based on the tank wall curvature and tank wall thickness.

[0039] The stress is calculated based on the bending strain and the material's elastic modulus, and the amplitude and number of stress cycles are statistically analyzed.

[0040] The equivalent stress amplitude is obtained by correcting the mean stress.

[0041] Calculate the fatigue life corresponding to each stress level based on the stress-fatigue life curve of the material, and calculate the cumulative fatigue damage.

[0042] The calculation of cumulative fatigue damage is a progressive process from geometric deformation to life assessment: First, the curvature of the tank wall is calculated using the height and tilt angle differences between adjacent tilt sensors, and then the bending strain is obtained by combining this with the tank wall thickness. Next, the strain is converted into stress using the elastic modulus, and the stress cycle amplitude and number of cycles experienced by the tank during operation are statistically analyzed. Then, considering the influence of average stress on fatigue life, the stress amplitude is corrected to obtain the equivalent stress amplitude. Finally, the fatigue life corresponding to each stress level is found based on the material's SN curve (stress-fatigue-life curve), and the cumulative fatigue damage is calculated using Miner's linear cumulative damage rule. This value reflects how much of the tank's lifespan has been consumed; when the cumulative damage approaches 1, it indicates that the tank is about to reach the critical state of fatigue failure, requiring timely warning and maintenance.

[0043] In this embodiment, the dynamic response monitoring chain is deployed along the tank wall. By observing the change in inclination angle between adjacent monitoring points, the bending strain of the tank wall is inferred: Assuming that within a certain height range, the monitoring heights of the upper and lower inclinometers are h1 and h2 respectively, then the average curvature K of the tank wall in this range is: K= ; in, and These are the tilt angle values ​​measured by two adjacent tilt sensors, and L is the height difference between the two sensors.

[0044] Bending strain It satisfies the following relationship with curvature K: ; Where t is the tank wall thickness, and t / 2 is the distance from the outer surface of the tank wall to the neutral axis.

[0045] Next, according to Hooke's Law , will strain Converted into stress , where E(T) is the elastic modulus of the tank wall material after temperature correction.

[0046] The time-series stress data were processed using the "rainflow counting method" to identify the maximum stress in each stress cycle. Minimum stress Statistical analysis of different stress amplitudes =( - ) / 2 and mean stress =( + / 2), corresponding to N loop counts. i .

[0047] Since mean stress affects the fatigue strength of materials, the nominal stress amplitude is corrected using the Goodman formula to obtain the effective stress amplitude: = / (1- / );in, This refers to the tensile strength of the tank wall material.

[0048] By consulting fatigue performance standards for tank wall materials (such as GB / T 3075-2008) or conducting fatigue tests, the stress-fatigue life curve (SN) of the material can be obtained. Its expression is as follows: ×N=C; where m is the material fatigue index and C is the material fatigue constant.

[0049] For each effective stress amplitude The fatigue life N is calculated based on the SN curve. i Damage D caused by this level of stress i For: D i =n i / N i ; where n i This represents the actual number of cycles for this stress level.

[0050] According to Miner's theory, the total fatigue damage D is the sum of the damage at each stress level: When D=1, the tank wall is determined to have reached its fatigue limit, and there is a risk of failure.

[0051] Based on monitoring data from the past 12 months, the monthly average cumulative fatigue damage rate D was calculated. 月 =D 总 / 12 (D represents the total damage over the past 12 months); assuming future operating conditions are consistent with historical operating conditions, the remaining fatigue life L is: L 剩余 =(1-D 总 ) / D 月。 Among them, D 总 For the current cumulative fatigue damage, D 月 This represents the monthly average cumulative rate of fatigue damage.

[0052] Furthermore, the intelligent prediction model is an LSTM neural network model.

[0053] Input features include the overall tilt angle of historical data, maximum relative settlement difference, local maximum curvature change rate, maximum deflection of tank wall, equivalent stiffness, stiffness degradation rate, cumulative fatigue damage, monthly average liquid level, and liquid level fluctuation amplitude.

[0054] The output prediction results include the overall settlement, local deformation, equivalent stiffness, fatigue damage, and overall structural health index of the storage tank in the future period.

[0055] To comprehensively reflect the key aspects of the storage tank's structural behavior and its influencing factors, the input features used in this application have been carefully selected, covering the tank's geometric deformation, material performance degradation, and external load conditions. Among these, the overall tilt angle, maximum relative settlement difference, local maximum rate of curvature change, and maximum deflection of the tank wall, based on historical monitoring data, primarily characterize the macroscopic and local geometric deformation of the tank. These features are obtained through raw data collected by tilt sensors and level sensors in the monitoring chain, combined with geometric calculations and fitting methods. The equivalent stiffness and stiffness degradation rate reflect the dynamic changes in the tank's structural load-bearing capacity and the degree of material performance degradation; their calculations are based on tilt angle data, level data, and corresponding mechanical models. The cumulative fatigue damage quantifies the degree of fatigue damage to the tank under long-term loads, typically calculated through stress-strain analysis and fatigue life theory. Furthermore, the monthly average level and level fluctuation amplitude, as important external load characteristics, are directly obtained from the level sensor data and are used to reflect the changes in internal pressure experienced by the tank. These multi-dimensional, multi-scale features are usually normalized before being input into the LSTM model to eliminate the differences in scale between different features and ensure the stability and convergence speed of model training.

[0056] Before inputting feature parameters, normalization processing is required to transform all input features to the [0, 1] interval to avoid affecting model training due to differences in units. For a small amount of missing data, such as temporary data interruptions caused by extreme weather, linear interpolation is used to supplement it. Historical monitoring data is divided into a training set (for model training), a validation set (for hyperparameter tuning), and a test set (for model performance verification) according to a certain proportion.

[0057] Model structure design such as Figure 2 As shown, the model consists of 3 LSTM hidden layers and 1 fully connected output layer; the first hidden layer has 27 neurons, the second has 26, and the third has 25; the tanh activation function is used.

[0058] The intelligent pre-model consists of 3 LSTM hidden layers and 1 fully connected output layer, using the tanh activation function.

[0059] Specifically, the intelligent prediction model adopts a deep learning architecture, consisting of three LSTM (Long Short-Term Memory) hidden layers and one fully connected output layer. The LSTM layer is responsible for learning the time-series features of the tank monitoring data, which can capture the evolution patterns and long-term dependencies of parameters such as deformation and stress over time. The fully connected layer maps the time-series features extracted by the LSTM to the final health index output. The model uses the tanh activation function (hyperbolic tangent function, with an output range between -1 and 1) to introduce nonlinear transformation capabilities, enabling the model to learn the complex nonlinear relationship of tank structure degradation.

[0060] The model training employs an early stopping method: training is stopped and the optimal model parameters are saved when the validation set loss fails to decrease after multiple consecutive iterations. Mean absolute error (MAE) and root mean square error (RMSE) are used to evaluate prediction accuracy. The model is periodically incrementally trained using newly added real-world monitoring data to ensure that prediction accuracy continuously optimizes over time.

[0061] The overall structural health index (HI) is divided into the following levels: When HI ≥ 80, the storage tank is in a healthy state and requires no special maintenance. This indicates that the tank's structural performance is excellent, all indicators are within safe ranges, and no additional intervention is needed. In this state, the system can maintain a regular monitoring frequency to ensure long-term stability.

[0062] When 60 ≤ HI < 80, the storage tank is in good condition and routine monitoring should be performed. This condition indicates that the tank's structural performance is good, but there may be slight fluctuations or potential risks, which have not yet reached the level requiring immediate intervention. The system recommends continuing routine monitoring and closely monitoring data trends to prevent risk escalation.

[0063] When 40 ≤ HI < 60, the storage tank is in an early warning state, and the monitoring frequency is increased while a detailed inspection is conducted. When the health index enters this range, it means that the structural behavior of the storage tank has become abnormal to a certain extent, posing a potential risk. At this time, the system will trigger an early warning and recommend immediately increasing the monitoring frequency, while simultaneously arranging for professional personnel to conduct a detailed on-site inspection to identify the specific problem and assess the risk level.

[0064] When 20 ≤ HI < 40, the storage tank is in a hazardous state, requiring restricted use and a maintenance plan to be developed. This state indicates that the tank's structural characteristics have significantly deteriorated, posing a high risk that may affect its safe operation. The system will issue a high-level warning and recommend immediate measures to restrict use, such as reducing the liquid level or limiting operating conditions. Simultaneously, a detailed maintenance plan must be developed quickly to repair structural defects.

[0065] When HI < 20, the storage tank is in a critically dangerous state and must be immediately shut down and emergency measures implemented. This is the highest level of warning, indicating that the tank's structural characteristics have reached a critical point, posing a serious safety hazard and a potential accident at any time. The system will forcibly trigger an emergency shutdown command and require immediate emergency response measures, such as draining the stored liquid and isolating the site, to minimize personal injury and property damage.

[0066] By refining the overall structural health index (HI) into a categorized level and clearly defining the corresponding status description, warning level, and maintenance decisions for each level, this application transforms the abstract health index into concrete and actionable guidelines. This allows managers to intuitively understand the current structural state of the storage tank, avoiding decision-making errors caused by vague judgments. For example, when the storage tank is in a healthy or good condition, unnecessary over-maintenance can be avoided, saving resources; while when the storage tank enters a warning or dangerous state, the system can promptly and accurately trigger the corresponding level of warning and provide clear handling suggestions, such as increasing monitoring frequency, restricting use, or immediate shutdown, thereby ensuring effective intervention measures are taken before the risk escalates, greatly improving the safety of storage tank operation, the targeted nature of maintenance, and the efficiency of management.

[0067] Example 2 This embodiment provides an intelligent monitoring and early warning system for the structural performance of storage tanks. Please refer to [link / reference]. Figure 2 ,include: The monitoring module includes a monitoring chain deployed along the tank wall 113. The monitoring chain includes multiple tilt sensors and liquid level sensors for collecting tilt data and liquid level data of the tank wall 113.

[0068] The calculation and evaluation module calculates multiple characteristic data representing the structural state of the storage tank based on the tilt angle data and the liquid level data; based on the multiple characteristic data, it outputs the overall structural health index of the storage tank using an intelligent prediction model.

[0069] The early warning module triggers multi-level early warnings and maintenance decisions based on the overall structural health index.

[0070] Furthermore, it also includes a data transmission module and an application terminal module.

[0071] The data transmission module is used to transmit the monitoring data to the cloud platform, and the calculation and evaluation module and the early warning module are deployed on the cloud platform.

[0072] The application terminal module is used to receive and display the early warning signals and maintenance decision suggestions output by the early warning module.

[0073] The intelligent monitoring and early warning system for the structural behavior of storage tanks proposed in this embodiment adopts a four-layer architecture of perception-transmission-analysis-application, constructing a layered and zoned monitoring module on the tank body. Specifically, the monitoring module includes a top monitoring ring 101 and a bottom monitoring ring 102 deployed on the tank top 114 and tank wall 113, as well as a dynamic monitoring chain 103 and locally reinforced monitoring grids 104 for key parts, forming a three-dimensional monitoring system covering the entire height of the storage tank. Each monitoring node is equipped with an inclinometer 111, a liquid level gauge 112, and a temperature sensor to collect multi-dimensional data such as inclinometer angle, liquid level, and temperature in real time. The collected monitoring data is transmitted to the cloud platform's data acquisition gateway via LoRa / 4G / 5G wireless communication. The cloud platform's calculation and evaluation module first preprocesses and stores the raw data in time series. Then, based on the tilt angle data and liquid level data, it calculates multi-dimensional characteristic parameters such as structural bearing capacity and fatigue damage accumulation using mechanical algorithms. These characteristic data are then input into an intelligent prediction model composed of a multi-layer LSTM neural network, which comprehensively outputs a quantitative overall structural health index. Based on the value range of this health index, the safety level of the storage tank is automatically determined, and corresponding early warning signals and maintenance decision suggestions are generated. Finally, the monitoring data, health index, and early warning decision information are pushed to management personnel in real time through application terminals (monitoring center large screen, mobile app, automatic alarm, reporting system). This achieves a technological leap from discrete sensor measurement points to intelligent perception of overall structural health, and from passive periodic detection to proactive real-time early warning.

[0074] In summary, by deploying a monitoring chain of tilt and level sensors on the tank wall 113, real-time data is collected and characteristic parameters such as overall settlement, local deformation, structural bearing capacity, and fatigue damage accumulation are calculated. A 3-layer LSTM deep learning model is then used to output a quantified health index, triggering early warnings and maintenance decisions based on the graded health index. This method overcomes the limitations of traditional monitoring technologies, achieving full-dimensional structural behavior perception and proactive real-time early warning. It can provide early warnings weeks or even months before a serious accident occurs in the tank, reducing safety risks and avoiding casualties, environmental pollution, and economic losses caused by sudden failures. Furthermore, by accurately assessing the remaining lifespan, it guides maintenance plans, avoiding production losses due to overly conservative maintenance and safety hazards caused by overly risky use.

[0075] The above examples illustrate the present invention only to aid in understanding it and are not intended to limit the scope of the invention. Those skilled in the art can make various simple deductions, modifications, or substitutions based on the principles of this invention.

Claims

1. A method for intelligent monitoring and early warning of the structural state of a storage tank, characterized in that, include: A monitoring chain is installed on the tank wall, which includes multiple tilt sensors and liquid level sensors to collect tilt angle data and liquid level data of the tank wall in real time. Based on the tilt angle data and the liquid level data, multiple characteristic data characterizing the structural behavior of the storage tank are calculated, including structural load-bearing capacity data and fatigue state data. Based on the aforementioned feature data, an intelligent prediction model is used to output the overall structural health index of the storage tank. Based on the numerical range of the overall structural health index, the structural condition of the storage tank is divided into different levels, and corresponding early warning and maintenance decisions are triggered according to different levels.

2. The method of claim 1, wherein the method further comprises: The structural bearing capacity data includes equivalent stiffness and stiffness degradation rate, and the calculation of the structural bearing capacity data includes: Based on the installation height of the tilt sensor and the measured tilt angle value, a deformation curve of the tank wall is constructed. The slope of the deformation curve is integrated, and the maximum radial displacement in the integration result is selected as the deformation amount. Calculate the effective pressure of the liquid on the tank wall based on the liquid level and liquid density; The equivalent stiffness is calculated based on the ratio of the effective pressure to the deformation. The stiffness degradation rate is calculated based on the comparison between the equivalent stiffness and the initial stiffness.

3. The intelligent monitoring and early warning method for the structural performance of storage tanks as described in claim 2, characterized in that, The stiffness degradation rate includes the relative stiffness change rate and the monthly average stiffness degradation rate; The relative stiffness change rate = 00%； The monthly average stiffness change rate 00%； where K eq，t is the equivalent stiffness at the tth measurement, K eq，0 is the initial equivalent stiffness, t is time, is the time interval.

4. The method of claim 1, wherein the method further comprises: The fatigue state data includes the cumulative amount of fatigue damage, and the calculation process for the cumulative amount of fatigue damage includes: The tank wall curvature is calculated based on the height difference and tilt angle difference between adjacent tilt sensors, and the bending strain is calculated based on the tank wall curvature and tank wall thickness. The stress is calculated based on the bending strain and the material's elastic modulus, and the amplitude and number of stress cycles are statistically analyzed. The equivalent stress amplitude is obtained by correcting the average stress. Calculate the fatigue life corresponding to each stress level based on the stress-fatigue life curve of the material, and calculate the cumulative fatigue damage.

5. The intelligent monitoring and early warning method for the structural performance of storage tanks as described in claim 4, characterized in that, The monthly average fatigue damage accumulation rate was calculated based on historical monitoring data, and the remaining fatigue life was also calculated. L 剩余 = (1 - D 总 ) / D 月 ; where D 总 is the current accumulated fatigue damage, D 月 is the monthly average fatigue damage accumulation rate.

6. The method for structural state intelligent monitoring and early warning of the storage tank according to claim 1, characterized in that, The intelligent prediction model is an LSTM neural network model; Input features include the overall tilt angle, maximum relative settlement difference, local maximum curvature change rate, maximum deflection of tank wall, equivalent stiffness, stiffness degradation rate, cumulative fatigue damage, monthly average liquid level, and liquid level fluctuation amplitude of historical data. The output prediction results include the overall settlement, local deformation, equivalent stiffness, fatigue damage, and overall structural health index of the storage tank in the future period.

7. The method of claim 6, wherein the method further comprises: The training process of the LSTM neural network model includes: The input features are normalized and transformed to the [0, 1] interval; Historical data is divided into training set, validation set, and test set; Training is performed, and training is stopped when the validation loss does not decrease for several consecutive iterations. The accuracy of predictions is assessed using mean absolute error and root mean square error. 8.The method of claim 1, wherein, The overall structural health index (HI) is divided into the following levels: When HI≥80, the storage tank is in good condition and requires no special maintenance. When 60≤HI<80, the storage tank is in good condition and routine monitoring should be performed. When 40≤HI<60, the storage tank is in an early warning state, the monitoring frequency is increased and a detailed inspection is carried out; When 20≤HI<40, the storage tank is in a dangerous condition, its use should be restricted, and a maintenance plan should be developed. When HI < 20, the storage tank is in a critically dangerous state and should be shut down immediately for emergency treatment.

9. A storage tank structural state intelligent monitoring and early warning system, characterized in that, include: The monitoring module includes a monitoring chain deployed along the tank wall, the monitoring chain including multiple tilt sensors and liquid level sensors, used to collect tilt angle data and liquid level data of the tank wall; The calculation and evaluation module calculates multiple characteristic data representing the structural state of the storage tank based on the tilt angle data and the liquid level data; based on the multiple characteristic data, it outputs the overall structural health index of the storage tank using an intelligent prediction model. The early warning module triggers multi-level early warnings and maintenance decisions based on the overall structural health index.

10. The structural state intelligent monitoring and early warning system of the storage tank according to claim 9, characterized in that, It also includes a data transmission module and an application terminal module; The data transmission module is used to transmit the monitoring data to the cloud platform, and the calculation and evaluation module and the early warning module are deployed on the cloud platform; The application terminal module is used to receive and display the early warning signals and maintenance decision suggestions output by the early warning module.

Images