Dangerous goods storage tank group safety monitoring method and system

Abstract

The application discloses a dangerous goods storage tank group safety monitoring method and system, and ingeniously monitors the corrosion condition, crack initiation and expansion condition of the storage tank bottom plate of the storage tank group through a corrosion monitoring sensing module and an acoustic monitoring sensing module, so that the thickness and acoustic abnormal condition of the bottom plate of the storage tank can be monitored in time during long-term application, and the corrosion condition and stress change condition of the storage tank body can be reliably known by the background management personnel in time. The application also fuses data of different dimensions, such as corrosion rate data, acoustic event features, environmental parameters, inter-tank position parameters and storage tank bottom plate thickness difference data, to generate multi-source fusion data, and on this basis, a corresponding algorithm is used to separately and combinedly judge the multi-source data, so as to analyze whether the storage tank bottom plate is abnormal, and a reliable mechanism guarantee is provided for the safety monitoring of the storage tank group.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of LNG storage tank monitoring, and particularly relates to a dangerous goods storage tank group safety monitoring method and system. BACKGROUND

[0002] Liquefied Natural Gas (LNG) as a clean and efficient energy occupies an increasingly important position in the global energy structure. LNG dangerous goods storage tanks, due to the storage medium having low temperature, flammable and potentially high risk characteristics, their safe operation is related to the stability of the energy supply chain and the safety of the surrounding environment.

[0003] At present, the safety monitoring of LNG storage tanks focuses on the integrity of the structure and the sealing protection. Traditional storage tank safety management relies on periodic manual inspection and offline non-destructive testing (NDT) techniques such as ultrasonic thickness measurement, magnetic powder detection, eddy current detection and penetration detection. These techniques can to some extent find corrosion, cracks, local thinning and potential structural defects in the tank bottom plate. However, due to the long inspection cycle and insufficient detection frequency, when corrosion or cracks develop rapidly between two detections, risk accumulation may occur, and early warning of accident signs cannot be achieved. In addition, manual or periodic offline detection is often time-consuming and labor-intensive, and has an impact on the normal operation of the storage tank.

[0004] With the development of sensors and Internet of Things technology, more and more online monitoring methods have been tried to be applied in the field of LNG storage tank safety. For example, electrochemical corrosion sensors, fiber Bragg grating sensors and ultrasonic thickness measurement sensors can be used to continuously monitor the corrosion condition of the tank bottom plate; multi-point arranged acoustic emission (AE) sensors can be used to capture acoustic event signals caused by micro-crack initiation, expansion, pitting damage and potential leakage. On the other hand, the risk of LNG storage tank group is not only affected by internal structural factors such as corrosion and crack expansion, but also by external environmental factors (such as temperature and humidity, corrosive atmosphere composition), inter-tank safety distance and layout, and overall operation conditions of the storage tank group. Traditional risk assessment is often based on a single factor or static conditions, and it is difficult to obtain accurate and dynamic risk assessment results in complex and variable actual operating environments.

[0005] In the case of multi-source data parallel generation and massive information accumulation, it is a trend to organically integrate corrosion monitoring, acoustic monitoring and environmental parameters, operating condition parameters of the storage tank, and to perform all-round and multi-level safety evaluation on the LNG storage tank group by using big data analysis, machine learning and probabilistic risk evaluation models (such as Bayesian network or risk benchmark evaluation model). The technical route can realize online monitoring of the corrosion of the bottom plate of the storage tank, dynamically predict the corrosion rate according to real-time data of the sensor, identify acoustic abnormal events (such as crack propagation, pitting, and micro-leakage), and fully consider external environmental factors such as gas composition, relative humidity, and wind speed and direction, so as to give a risk index and early warning information of the storage tank group. The multi-factor comprehensive risk evaluation method helps to realize closed-loop control of the safety state of the storage tank, that is, timely early warning when a safety hazard is detected, and feedback of maintenance and repair data to the model for parameter correction and continuous optimization.

[0006] However, the existing monitoring and evaluation methods are mostly fragmented, and it is difficult to organically integrate corrosion monitoring, acoustic monitoring, environmental parameter detection and risk evaluation. There is a lack of a systematic technical solution to integrate hardware sensing devices, data acquisition and transmission platforms, data analysis models and visual decision support tools in a unified framework, so as to dynamically and closed-looply grasp the safety status of the LNG storage tank group in the running process.

[0007] Therefore, it is necessary to propose a safety monitoring method and system capable of integrating various sensing technologies, data processing methods and intelligent risk analysis models, to realize real-time monitoring, comprehensive analysis and dynamic early warning of the corrosion of the bottom plate of the LNG dangerous goods storage tank group and the overall risk of the tank group, and to provide strong technical support for the long-period safe operation of the storage tank. SUMMARY

[0008] Therefore, it is necessary to propose a safety monitoring method and system capable of integrating various sensing technologies, data processing methods and intelligent risk analysis models, to realize real-time monitoring, comprehensive analysis and dynamic early warning of the corrosion of the bottom plate of the LNG dangerous goods storage tank group and the overall risk of the tank group, and to provide strong technical support for the long-period safe operation of the storage tank.

[0009] In order to achieve the above technical purpose, the technical scheme adopted by the present application is as follows:

[0010] A dangerous goods storage tank group safety monitoring method is applied to the monitoring of an LNG dangerous goods storage tank group. Corrosion monitoring sensing modules and acoustic monitoring sensing modules are arranged in the bottom plate area of the LNG dangerous goods storage tank group. An environmental sensor group is arranged in the placement site of the LNG dangerous goods storage tank group, which comprises:

[0011] S01, the corrosion monitoring sensing modules and the acoustic monitoring sensing modules are used to monitor the corrosion of the bottom plate area of the storage tank in real time, and the crack initiation and propagation, to generate corrosion monitoring data and acoustic monitoring data, respectively;

[0012] S02, monitoring the environmental corrosion factors and weather conditions around the storage tank and between the storage tanks by the environmental sensor group to generate environmental monitoring data;

[0013] S03, obtaining the corrosion monitoring data, acoustic monitoring data and environmental monitoring data, performing data cleaning, filtering and feature extraction, and then performing multi-source data fusion processing to generate multi-source fusion data;

[0014] S04, obtaining the multi-source fusion data, using a risk assessment algorithm based on machine learning and statistical model to assess the corrosion degree of the storage tank and the overall risk of the storage tank group, and generating an assessment result;

[0015] S05, obtaining the assessment result, and generating warning information according to the assessment result for the storage tank and the dangerous goods storage tank group;

[0016] S06, constructing a digital twin model based on the dangerous goods storage tank group, and presenting the assessment result and the warning information in a visual form.

[0017] As a possible implementation, further, in the present scheme S01, the corrosion monitoring sensor module includes a plurality of electrochemical sensors for monitoring the corrosion condition of the storage tank bottom plate; the acoustic monitoring sensor module includes a plurality of ultrasonic sensors for measuring the thickness, crack and crack propagation of the storage tank bottom plate; part of the ultrasonic sensors are arranged in the form of a sensor array in the preset monitoring area of the storage tank bottom plate.

[0018] As a possible implementation, further, the corrosion monitoring data of the present scheme includes the corrosion current density and corrosion rate of the corresponding monitoring points on the storage tank bottom plate; the acoustic monitoring data includes the bottom plate thickness of the corresponding monitoring points on the storage tank bottom plate and the corresponding received acoustic wave data and sound source position data.

[0019] As a possible implementation, further, the method for real-time monitoring and collecting data of the bottom plate area of the storage tank by the corrosion monitoring sensor module and the acoustic monitoring sensor module of the present scheme includes:

[0020] a plurality of monitoring points are arranged on the storage tank bottom plate, and the position information set is defined as , n for the number of monitoring points, for each monitoring point , the thickness thereof is periodically obtained by the acoustic monitoring sensor module , and the corrosion current density thereof is measured by the corrosion monitoring sensor module ;

[0021] based on the basic parameters of the storage tank bottom plate and the measured corrosion current density , the corrosion rate of the storage tank bottom plate is calculated, which is defined as follows:

[0022]

[0023] wherein, is the instantaneous corrosion rate at the monitoring point , is the unit conversion constant, is the molar mass of the metal used for the tank floor, is the number of electrons lost in the corrosion reaction, is the density of the metal used for the tank floor, is the corrosion current density at the monitoring point ;

[0024] The floor thickness at the monitoring point is obtained by the acoustic monitoring sensor module to assist in verifying the reliability of the calculation of the corrosion rate, which can be fitted into a mathematical relationship of thickness change over time, which is defined as follows:

[0025]

[0026] wherein, is the floor thickness measured by the acoustic monitoring sensor module at the monitoring point at time , is the floor thickness measured by the acoustic monitoring sensor module at the monitoring point at time , is the thickness difference;

[0027] The collection of sensing points of multiple ultrasonic sensors is defined as , n is the number of sensing points;

[0028] For acoustic source positioning, the time difference between the arrival of the acoustic signal emitted by the acoustic source at each sensing point of the ultrasonic sensor is recorded , and the acoustic source position coordinates are located, under the condition that the positions of the sensing points of the ultrasonic sensor and the acoustic velocity are known, the acoustic source position coordinates are solved according to the least square method, and the formula is defined as follows:

[0029]

[0030] The acoustic source position coordinates are obtained by solving the algorithm of the least square method, wherein, is the corresponding , x , y coordinate position of the sensing point

[0031] As a preferred implementation option, preferably, in the scheme S03, when extracting the acoustic monitoring data, the sound wave data collected by the ultrasonic sensor is subjected to wavelet transform, which is defined as follows:

[0032]

[0033] wherein, is the sound wave data is the wavelet coefficient at scale and translation ; is the wavelet base function; is the scale factor, is the translation factor, is the complex conjugate;

[0034] By selecting the wavelet base function and the characteristic coefficient statistics, the characteristic parameters of the acoustic event are extracted, which are set as acoustic event characteristics, including one or more of energy , duration , ring count ;

[0035] wherein the obtained characteristic parameters are associated with the sound source position coordinates together and stored, and collected into the acoustic monitoring data for multiplexing.

[0036] As a preferred implementation option, preferably, the environmental sensor group of the scheme includes one or more of a gas component sensor, a temperature sensor, a humidity sensor, a wind speed and direction sensor; the environmental corrosion factors include hydrogen sulfide, sulfur dioxide, chloride ions;

[0037] In S02, the environmental monitoring data is collected in the form of an environmental parameter set to form an environmental parameter vector , which is defined as follows:

[0038]

[0039] wherein, are respectively the gas concentration or ion concentration data of the corrosive gases hydrogen sulfide, sulfur dioxide and chloride ions measured by the gas component sensor; is the environmental temperature data, is the relative humidity data, are respectively the wind direction data and the wind speed data.

[0040] As a preferred implementation option, preferably, the scheme S02 further includes:

[0041] For a group of storage tanks, assuming there is a group of storage tanks , the center positions of the storage tanks are obtained by a laser ranging module or a GPS positioning module , wherein the inter-tank distance may be represented as follows:

[0042]

[0043] wherein, is the coordinate position of the storage tank i , and is the coordinate position of the storage tank j ;

[0044] all the inter-tank distances of the storage tank group are collected to form the inter-tank position parameter .

[0045] As a preferred implementation option, preferably, the scheme S03 comprises:

[0046] obtaining corrosion monitoring data, acoustic monitoring data and environmental monitoring data, and setting them as multi-source data;

[0047] detecting and interpolating the outliers in the multi-source data to realize data cleaning, and then using a sliding window average method or a median filtering method to filter the multi-source data, and then normalizing the data features to map all the data features to the interval [0, 1] so that data of different dimensions can be fused;

[0048] establishing a multi-dimensional feature vector to fuse data of different dimensions to generate multi-source fusion data , which includes corrosion rate data , acoustic event features , environmental parameters , inter-tank position parameters and storage tank bottom plate thickness difference data , which are defined as follows:

[0049]

[0050] wherein the multi-source fusion data is used to evaluate the corrosion degree of the storage tank and the overall risk of the storage tank group.

[0051] As a preferred implementation option, preferably, the scheme S04 comprises:

[0052] S041, obtaining multi-source fusion data, and extracting corrosion rate data , acoustic event features , environmental parameters , inter-tank position parameters Thickness difference between tank bottom plate and ;

[0053] S042. Use Random Forest RF or LSTM Time Series Network to calculate the corrosion rate of future moments Make predictions to obtain predicted corrosion rates;

[0054] Wherein, the input features include corrosion rate data, as well as environmental parameters and / or acoustic event characteristics;

[0055] S043. Classify the acoustic event features using a classifier including a SVM algorithm or a random forest algorithm, where the classification categories include: microcrack extension event (Crack), pitting event (Pitting), local leakage event (Leakage), or mechanical noise event (Noise); and obtain an acoustic event classification result.

[0056] Among them, the input feature vector of the classifier is , the output is the event category label ;

[0057] S044. Comprehensively analyze multi-source fusion data through Bayesian network (BN). The analysis factors include corrosion rate data, acoustic event characteristics, environmental parameters, and tank location parameters.

[0058] Among them, the Bayesian network nodes include:

[0059] Base Plate Corrosion Status Node C ;

[0060] Acoustic Event Node A ;

[0061] Environmental Corrosion Factor Node E ;

[0062] Distance between storage tanks and linkage risk nodes D ;

[0063] Construct the joint probability distribution of the Bayesian network BN, which is defined as follows:

[0064]

[0065] By inferring the Bayesian network BN, when the current corrosion rate increases, the frequency of acoustic events increases, or the environmental parameters are unfavorable, the posterior inference of the Bayesian network BN will give a high-probability risk state, and finally the result of the Bayesian network BN is mapped to a risk index with a value between 0 and 1. R , which is defined as follows:

[0066]

[0067] in, to normalize the mapping function;

[0068] S045, collecting the predicted corrosion rate, the acoustic event classification result and the risk index R as an evaluation result output.

[0069] As a preferred implementation option, preferably, the present scheme S05 comprises:

[0070] obtaining an evaluation result, comparing the predicted corrosion rate, the acoustic event classification result and the risk index R with a preset threshold or a reference classification, when it exceeds, generating a warning information and outputting, to realize the early warning of the storage tank and the dangerous goods storage tank group.

[0071] Based on the above, the present scheme also provides a dangerous goods storage tank warehouse area management method, which comprises the above-mentioned dangerous goods storage tank group safety monitoring method.

[0072] Based on the above, the present scheme also provides a dangerous goods storage tank group safety monitoring system, which comprises:

[0073] The corrosion monitoring sensor module is multiple, which is arranged on the inner wall and / or outer wall of the storage tank bottom plate, and is used for real-time monitoring of the corrosion condition of the storage tank bottom plate area, and generating corrosion monitoring data;

[0074] The acoustic monitoring sensor module is multiple, which is arranged on the inner wall and / or outer wall of the storage tank bottom plate, and is used for real-time monitoring of the crack initiation and expansion condition of the bottom plate area of the storage tank, and generating acoustic monitoring data;

[0075] The environmental sensor group is used for monitoring the environmental corrosion factors and weather conditions around the storage tank and between the storage tanks, and generating environmental monitoring data;

[0076] The data processing module is used for obtaining the corrosion monitoring data, the acoustic monitoring data and the environmental monitoring data, and performing data cleaning, filtering and feature extraction, and then performing multi-source data fusion processing to generate multi-source fusion data;

[0077] The data evaluation module is used for obtaining the multi-source fusion data, and evaluating the corrosion degree of the storage tank and the overall risk of the storage tank group by using a risk evaluation algorithm based on machine learning and statistical model, to generate an evaluation result;

[0078] The warning information unit is used for obtaining the evaluation result, and warning the storage tank and the dangerous goods storage tank group according to the evaluation result, to generate a warning information;

[0079] The digital twin module is used for constructing a digital twin model based on the dangerous goods storage tank group, and presenting the evaluation result and the warning information in a visual form.

[0080] By adopting the above technical solution, the present invention has the following beneficial effects compared with the prior art: the present solution cleverly monitors the corrosion, crack initiation and expansion of the bottom plates of the storage tanks of the storage tank group through the corrosion monitoring sensor module and the acoustic monitoring sensor module, so that when the storage tanks are used for a long time, the thickness of the bottom plates and the acoustic abnormalities can be monitored in time, so that the back-end management personnel can timely and reliably know the corrosion status and stress change status of the storage tank bodies. At the same time, the present solution also combines the environmental sensor group to monitor the environmental corrosion factors and meteorological conditions around and between the storage tanks, and generates environmental monitoring data, so that the placement environment of the storage tank group can be recorded, providing data basis and support for the maintenance and management of the storage tanks. In addition, the present solution also uses the corrosion rate data , acoustic event characteristics , environmental parameters , tank location parameters Thickness difference between tank bottom plate and Fusion of data of different dimensions to generate multi-source fusion data On this basis, the corresponding algorithms are used to make separate and combined judgments on the multi-source data to analyze whether there are any abnormalities in the tank bottom plate, providing a reliable mechanism guarantee for the safety monitoring of the tank group. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] 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 or the description of the prior art. 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.

[0082] Figure 1 This is a brief implementation flow chart of the monitoring method of this program;

[0083] Figure 2 This is a schematic diagram of the unit module connection of the monitoring system of this solution. DETAILED DESCRIPTION

[0084] The present invention will be described in further detail below with reference to the accompanying drawings and examples. It is particularly noted that the following examples are intended only to illustrate the present invention and are not intended to limit the scope of the present invention. Similarly, the following examples are only some embodiments of the present invention and are not intended to be exhaustive. All other embodiments obtained by those of ordinary skill in the art without creative effort are intended to fall within the scope of protection of the present invention.

[0085] like Figure 1As shown, the embodiment scheme is a dangerous goods storage tank group safety monitoring method, applied to the monitoring of an LNG dangerous goods storage tank group, wherein a corrosion monitoring sensor module and an acoustic monitoring sensor module are arranged in a bottom plate area of the LNG dangerous goods storage tank group, and an environment sensor group is arranged in a placement site of the LNG dangerous goods storage tank group, which comprises:

[0086] S01, the corrosion condition, crack initiation and expansion condition of the bottom plate area of the storage tank are monitored in real time by the corrosion monitoring sensor module and the acoustic monitoring sensor module, and corrosion monitoring data and acoustic monitoring data are respectively generated;

[0087] S02, the environmental corrosion factors and weather conditions around the storage tank and between the storage tanks are monitored by the environment sensor group, and environment monitoring data are generated;

[0088] S03, corrosion monitoring data, acoustic monitoring data and environment monitoring data are obtained, and after data cleaning, filtering and feature extraction, multi-source data fusion processing is performed to generate multi-source fusion data;

[0089] S04, the multi-source fusion data are obtained, and a risk assessment algorithm based on machine learning and statistical model is used to assess the corrosion degree of the storage tank and the overall risk of the storage tank group, and an assessment result is generated;

[0090] S05, the assessment result is obtained, and the storage tank and the dangerous goods storage tank group are warned according to the assessment result, and warning information is generated;

[0091] S06, a digital twin model based on the dangerous goods storage tank group is constructed, and the assessment result and the warning information are presented in a visual form.

[0092] The scheme ingeniously monitors the corrosion condition, crack initiation and expansion condition of the bottom plate of the storage tank group by the corrosion monitoring sensor module and the acoustic monitoring sensor module, so that the thickness of the bottom plate and the acoustic abnormal condition of the storage tank can be monitored in time during long-term application, and the background management personnel can timely and reliably know the corrosion condition and stress change condition of the storage tank body. Meanwhile, the scheme also monitors the environmental corrosion factors and weather conditions around the storage tank and between the storage tanks by the environment sensor group, generates environment monitoring data, and records the placement environment of the storage tank group, thereby providing data basis and support for the maintenance and management of the storage tank.

[0093] As a possible implementation, further in the scheme S01, the corrosion monitoring sensor module comprises a plurality of electrochemical sensors for monitoring the corrosion condition of the tank bottom plate; the acoustic monitoring sensor module comprises a plurality of ultrasonic sensors for measuring the thickness, crack and crack propagation of the tank bottom plate; part of the ultrasonic sensors in the plurality of ultrasonic sensors are arranged in the form of a sensor array in the preset monitoring area of the tank bottom plate.

[0094] As a possible implementation, further in the scheme, the corrosion monitoring data comprises the corrosion current density and corrosion rate of the corresponding monitoring point on the tank bottom plate; the acoustic monitoring data comprises the bottom plate thickness of the corresponding monitoring point on the tank bottom plate and the corresponding received acoustic wave data and sound source position data.

[0095] Correspondingly, as a possible implementation, further in the scheme, the method for the corrosion monitoring sensor module and the acoustic monitoring sensor module to monitor and collect data in real time in the bottom plate area of the tank comprises:

[0096] A plurality of monitoring points are arranged on the tank bottom plate, and the position information set is defined as , n For the number of monitoring points, for each monitoring point , the thickness thereof is obtained periodically by the acoustic monitoring sensor module , and the corrosion current density thereof is measured by the corrosion monitoring sensor module ;

[0097] Based on the basic parameters of the tank bottom plate and the measured corrosion current density , the corrosion rate of the tank bottom plate is calculated, which is defined as follows:

[0098]

[0099] Among them, is the instantaneous corrosion rate at the monitoring point , is the unit conversion constant, is the molar mass of the metal used in the tank bottom plate, is the number of lost electrons in the corrosion reaction, is the density of the metal used in the tank bottom plate, is the corrosion current density at the monitoring point ;

[0100] The bottom plate thickness of the monitoring point is obtained by the acoustic monitoring sensor module to assist in verifying the reliability of the calculation of the corrosion rate, which can be fitted into a mathematical relationship of the thickness change with time, which is defined as follows:

[0101]

[0102] in, Acoustic monitoring sensor module Monitoring points at all times The measured thickness of the base plate, Acoustic monitoring sensor module Monitoring points at all times The measured thickness of the base plate, Thickness difference;

[0103] The sensing point set of multiple ultrasonic sensors is defined as , n is the number of sensing points;

[0104] For sound source positioning, the time difference between the sound signal emitted by the sound source and the sensing point of each ultrasonic sensor is recorded. Locate the sound source coordinates , at the sensing point of the known ultrasonic sensor Position and speed of sound In the case of, the coordinates of the sound source position are solved according to the least square method ( ), whose formula is defined as follows:

[0105]

[0106] By solving the least squares equation, the coordinates of the sound source position are obtained ,in, The sensing point of the ultrasonic sensor Corresponding x 、 y Coordinate location.

[0107] In order to improve the reliability of the data, reduce the signal noise contained therein and facilitate the extraction of characteristic signals in the data, as a better implementation option, preferably, in this solution S03, when extracting the acoustic monitoring data, the sound wave data collected by the ultrasonic sensor is converted into Perform wavelet transform, which is defined as follows:

[0108]

[0109] in, For sound wave data In scale and pan The wavelet coefficients under ; is the wavelet basis function; , which is the scale factor, is the translation factor, is complex conjugate;

[0110] By selecting wavelet base function and characteristic coefficient statistics, the characteristic parameters of the acoustic event are extracted, which are set as acoustic event characteristics, including one or more of energy , duration , ring count .

[0111] Wherein, the obtained characteristic parameters are stored in association with the sound source position coordinates , and are collected into acoustic monitoring data for multiplexing.

[0112] In the aspect of environmental monitoring, as a preferred implementation selection, preferably, the environmental sensor group described in the scheme includes one or more of a gas component sensor, a temperature sensor, a humidity sensor, a wind speed and direction sensor; the environmental corrosion factor includes hydrogen sulfide, sulfur dioxide, chloride ion.

[0113] Correspondingly, in S02 of the scheme, the environmental monitoring data is collected in the form of an environmental parameter set to form an environmental parameter vector , which is defined as follows:

[0114]

[0115] Wherein, are the gas concentration or ion concentration data of the corrosive gases hydrogen sulfide, sulfur dioxide and chloride ion measured by the gas component sensor; is the environmental temperature data, is the relative humidity data, are the wind direction data and wind speed data, respectively.

[0116] In the aspect of tank spacing, as a preferred implementation selection, preferably, S02 of the scheme further includes:

[0117] For a tank group, assuming there is a tank group set , the center positions of each tank are obtained by a laser ranging module or a GPS positioning module , wherein the inter-tank distance can be represented as follows:

[0118]

[0119] Wherein, is the coordinate position of the tank i , and is the coordinate position of the tank j .

[0120] All inter-tank distances of the tank group are collected to form the inter-tank position parameter .

[0121] For the data fusion analysis aspect, as a preferred implementation option, the present solution S03 preferably comprises:

[0122] Obtaining corrosion monitoring data, acoustic monitoring data and environmental monitoring data, and setting them as multi-source data;

[0123] Detecting and interpolating the outliers in the multi-source data to realize data cleaning, and then using a sliding window average method or a median filtering method to filter the multi-source data, and then normalizing the data features to map all data features to the interval [0, 1] so that data of different dimensions can be fused and processed;

[0124] Establishing a multi-dimensional feature vector to fuse data of different dimensions and generate multi-source fusion data , which includes corrosion rate data , acoustic event features , environmental parameters , inter-tank position parameters and tank bottom plate thickness difference data , which are defined as follows:

[0125]

[0126] Among them, the multi-source fusion data is used to assess the corrosion degree of the storage tank and the overall risk of the storage tank group.

[0127] As a preferred implementation option, the present solution S04 preferably comprises:

[0128] S041, obtaining multi-source fusion data, and extracting corrosion rate data , acoustic event features , environmental parameters , inter-tank position parameters and tank bottom plate thickness difference data from the multi-source fusion data;

[0129] S042, using a random forest RF or LSTM time series network to predict the corrosion rate at a future time to obtain a predicted corrosion rate;

[0130] Among them, the input features include corrosion rate data, and environmental parameters and / or acoustic event features;

[0131] S043, classifying acoustic event features by a classifier containing an SVM algorithm or a random forest algorithm, the classification categories including: micro-crack propagation event Crack, pitting event Pitting, local leakage event Leakage or mechanical noise event Noise; obtaining acoustic event classification results;

[0132] wherein the input feature vector of the classifier , the output is the event class label ;

[0133] S044, comprehensive analysis is performed on the multi-source fusion data by the Bayesian network BN, and analysis factors include corrosion rate data, acoustic event features, environmental parameters, and inter-tank position parameters;

[0134] wherein the Bayesian network nodes include:

[0135] a bottom plate corrosion state node C ;

[0136] an acoustic event node A ;

[0137] an environmental corrosion factor node E ;

[0138] a tank distance and linkage risk node D ;

[0139] a joint probability distribution of the Bayesian network BN is constructed, and the definition is as follows:

[0140]

[0141] by inference of the Bayesian network BN, when the current corrosion rate is observed to increase, the acoustic event frequency rises, or the environmental parameters are unfavorable, the posterior inference of the Bayesian network BN will give a high probability of the risk state, and finally the result of the Bayesian network BN is mapped to a risk index with a value between 0 and 1 R , and the definition is as follows:

[0142]

[0143] wherein, is a normalized mapping function;

[0144] S045, the predicted corrosion rate, the acoustic event classification result, and the risk index R are collected as evaluation results and output.

[0145] On the basis of the above, as a preferred implementation option, preferably, the scheme S05 includes:

[0146] the evaluation results, the predicted corrosion rate, the acoustic event classification result, and the risk index R are compared with a preset threshold or a reference classification, when they exceed, a warning information is generated and output, so as to realize early warning of the tank and the dangerous goods tank group.

[0147] The scheme ingeniously fuses data of different dimensions, such as corrosion rate data , acoustic event features , environmental parameters , inter-tank position parameters , and tank bottom plate thickness difference data , to generate multi-source fusion data . On this basis, corresponding algorithms are used to make individual and combined judgments on the multi-source data to analyze whether there is an abnormality in the tank bottom plate; and a reliable mechanism is provided for the safety monitoring of the tank group.

[0148] Based on the above, the scheme further provides a dangerous goods tank storage area management method, which includes the dangerous goods tank group safety monitoring method described above.

[0149] As shown in Figure 2 , based on the above, the scheme further provides a dangerous goods tank group safety monitoring system, which includes:

[0150] The corrosion monitoring sensor modules are multiple, are disposed on the inner wall and / or outer wall of the tank bottom plate, and are used to monitor the corrosion condition of the tank bottom plate area in real time to generate corrosion monitoring data;

[0151] The acoustic monitoring sensor modules are multiple, are disposed on the inner wall and / or outer wall of the tank bottom plate, and are used to monitor the crack initiation and expansion condition of the tank bottom plate area in real time to generate acoustic monitoring data;

[0152] The environmental sensor group is used to monitor the environmental corrosion factors and meteorological conditions around the tank and between the tanks to generate environmental monitoring data;

[0153] The data processing module is used to acquire the corrosion monitoring data, acoustic monitoring data, and environmental monitoring data, and after data cleaning, filtering, and feature extraction, multi-source data fusion processing is performed to generate multi-source fusion data;

[0154] The data evaluation module is used to acquire the multi-source fusion data, and a risk evaluation algorithm based on machine learning and statistical models is used to evaluate the corrosion degree of the tank and the overall risk of the tank group to generate an evaluation result;

[0155] The early warning information unit is used to acquire the evaluation result, and according to the evaluation result, the tank and the dangerous goods tank group are warned to generate early warning information;

[0156] The digital twin module is used to construct a digital twin model based on the dangerous goods tank group, and the evaluation result and the early warning information are presented in a visual form.

[0157] In addition, each function unit in each embodiment of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software function unit.

[0158] If the integrated unit is realized in the form of a software function unit and sold or used as an independent product, it can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application, essentially or in the form of a contribution to the prior art, or all or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor to perform all or part of the steps of the methods in each embodiment of the present application. The foregoing storage medium includes various media that can store program codes, such as a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk.

[0159] The above only describes some embodiments of the present application, and does not limit the protection scope of the present application. Any equivalent device or equivalent process transformation, or direct or indirect application in other related technical fields, which is made by using the content of the specification and drawings of the present application, is also included in the protection scope of the present application.

Claims

1. A dangerous goods storage tank group safety monitoring method, applied to the monitoring of an LNG dangerous goods storage tank group, a corrosion monitoring sensor module and an acoustic monitoring sensor module are deployed in the bottom plate area of the LNG dangerous goods storage tank group, and an environmental sensor group is deployed in the placement site of the LNG dangerous goods storage tank group, characterized in that, It comprises: S01, by corrosion monitoring sensor module and acoustic monitoring sensor module, the corrosion condition, crack initiation and expansion condition of the bottom plate area of the storage tank are monitored in real time, and corrosion monitoring data and acoustic monitoring data are generated respectively; the corrosion monitoring data includes the corrosion current density and the corrosion rate of the corresponding monitoring point on the tank bottom plate; the acoustic monitoring data includes the bottom plate thickness of the corresponding monitoring point on the tank bottom plate and the corresponding received acoustic wave data and sound source position data; S02, by the environment sensor group, the environmental corrosion factors and weather conditions around the storage tank and between the storage tanks are monitored, and environment monitoring data is generated; S03, corrosion monitoring data, acoustic monitoring data and environment monitoring data are obtained, data cleaning, filtering and feature extraction are performed, and then multi-source data fusion processing is performed, and multi-source fusion data is generated; S04, the multi-source fusion data is obtained, the risk assessment algorithm based on machine learning and statistical model is used to evaluate the corrosion degree of the storage tank and the overall risk of the storage tank group, and the evaluation result is generated; S05, the evaluation result is obtained, the storage tank and the dangerous goods storage tank group are warned according to the evaluation result, and warning information is generated; Wherein, S03 comprises: Obtain corrosion monitoring data, acoustic monitoring data and environment monitoring data, and set them as multi-source data; The outliers in the multi-source data are detected and interpolated to realize data cleaning, and then the sliding window average method or the median filtering method is used for data filtering, and then the data features are normalized to map all data features to the interval [0, 1], so that data of different dimensions can be fused; A multi-dimensional feature vector is established to fuse data of different dimensions to generate multi-source fusion data F(t) including corrosion rate data {v corr，i (t)}, acoustic event features {E ae , D ae , N ring , x s , y s}, environmental parameters E(t), inter-tank position parameters {D ij (t)} and tank bottom thickness difference data {ΔT i (t)}, which are defined as follows: F(t) = [{v corr，i (t)}, {E ae , D ae , N ring , x s , y s}, E(t), {D ij (t)}, {ΔT i (t)}] Wherein, the multi-source fusion data F(t) is used to evaluate the corrosion degree of the storage tank and the overall risk of the storage tank group; S04 comprises: S041、Obtain multi-source fusion data, and extract corrosion rate data from the data{v corr，i (t)}、acoustic event features{E ae , D ae , N ring , x s , y s}、environmental parameters E(t), inter-tank position parameters{D ij (t)} and tank bottom thickness difference data{ΔT i (t)}; S042、using random forest, RF, or LSTM time series network on future time corrosion rate make a prediction, obtaining a predicted corrosion rate; Wherein, the input features include corrosion rate data, and environmental parameters and / or acoustic event features; S043, the acoustic event features are classified by a classifier containing SVM algorithm or random forest algorithm, the classified categories include: micro crack propagation event Crack, pitting event Pitting, local leakage event Leakage or mechanical noise event Noise; acoustic event classification result is obtained; where the input feature vector f = [E ae , D ae , N ring , x s , y s ] of the classifier, and the output is the event class label L ∈ {Crack, Pitting, Leakage, Noise}. S044, the multi-source fusion data is analyzed by Bayesian network BN, and the analysis factors include corrosion rate data, acoustic event features, environmental parameters and inter-tank position parameters.

2. The method of claim 1, wherein It further comprises: S06, a digital twin model based on the dangerous goods storage tank group is constructed, and the evaluation result and the warning information are presented in a visual form.

3. The method for monitoring safety of a group of dangerous product storage tanks according to Claim 2, wherein In S01, the corrosion monitoring sensor module comprises a plurality of electrochemical sensors for monitoring the corrosion condition of the tank bottom plate; the acoustic monitoring sensor module comprises a plurality of ultrasonic sensors for thickness measurement, crack and crack propagation monitoring of the tank bottom plate; part of the ultrasonic sensors in the plurality of ultrasonic sensors are arranged in a sensor array in the preset monitoring area of the tank bottom plate; The method for real-time monitoring and collecting data of the corrosion monitoring sensor module and the acoustic monitoring sensor module on the bottom plate area of the storage tank comprises: A plurality of monitoring points are arranged on the tank bottom plate, and the position information set is defined as P1, P2, …, Pn n , n is the number of monitoring points, for each monitoring point P i , periodically obtain its thickness T i (t) through the acoustic monitoring sensor module, and measure the corrosion current density I corr，i (t) through the corrosion monitoring sensor module; based on the basic parameters of the tank floor and the measured corrosion current density I corr，i (t) calculating the corrosion rate of the tank floor, which is defined as follows: where v corr，i (t) is the instantaneous corrosion rate at the monitoring point P i , k is the unit conversion constant, M is the molar mass of the metal used for the tank floor, n is the number of electrons lost in the corrosion reaction, p is the density of the metal used for the tank floor, I corr，i (t) is the corrosion current density at the monitoring point P i ; The bottom plate thickness of the monitoring point is acquired by the acoustic monitoring sensor module to assist in verifying the reliability of the calculation of the corrosion rate, which can be fitted into a mathematical relationship of thickness change over time, which is defined as follows: ΔT i (t) = T i (t0) - T i (t) where T i (t0) is the bottom plate thickness measured by the acoustic monitoring sensor module at time t0 i at time t0 i (t) is the bottom plate thickness measured by the acoustic monitoring sensor module at time t i at time t i (t) is the thickness difference; A set of sensing points of a plurality of ultrasonic sensors is defined as A1, A2, …, An n , n is the number of sensing points; For sound source positioning, the time difference Δτ between the sound signal emitted by the sound source and the sensing point of each ultrasonic sensor is recorded. i,j Position the sound source coordinates (x s ,y s ), at the sensing point A of the known ultrasonic sensor i In the case of the position and sound speed c, the coordinates of the sound source position (x s ,y s ), whose formula is defined as follows: The sound source position coordinates (x s , y s ) are obtained by solving the equation of least squares, wherein x j , y j are the x, y coordinate positions corresponding to the sensing point A i of the ultrasonic sensor.

4. The method for monitoring safety of a group of dangerous product storage tanks according to Claim 3, wherein In S03, when the acoustic monitoring data is extracted, the wavelet transform is performed on the acoustic data s(t) collected by the ultrasonic sensor, which is defined as follows: where W s (a, b) are the wavelet coefficients of the acoustic data s(t) at scale a and translation b; ψ(t) is the wavelet basis function; a > 0 is the scale factor and b is the translation factor; * is the complex conjugate; By selecting a wavelet base function and characteristic coefficient statistics, a characteristic parameter of an acoustic event is extracted, which is set as an acoustic event characteristic, including one or more of energy E ae , duration D ae , and ring count N ring . Wherein, the extracted feature parameters are stored in association with the sound source position coordinates (x s , y s ) and are aggregated into acoustic monitoring data for multiplexing.

5. The method for monitoring safety of a group of dangerous product storage tanks according to claim 3 or 4, characterized by, The environmental sensor group includes one or more of a gas component sensor, a temperature sensor, a humidity sensor, and a wind speed and direction sensor; and the environmental corrosion factors include hydrogen sulfide, sulfur dioxide, and chloride ions. In S02, the environmental monitoring data is collected in the form of an environmental parameter set to form an environmental parameter vector E(t), which is defined as follows: E(t) = {H2S(t), SO2(t), Cl - (t), T env (t), RH(t), W d (t), W s (t)} Among them, H2S(t), SO2(t), Cl - (t) are the gas concentration or ion concentration data of the corrosive gases hydrogen sulfide, sulfur dioxide, and chloride ions measured by the gas composition sensors; T env (t) is the ambient temperature data, RH(t) is the relative humidity data, W d (t), W s (t) wind direction data and wind speed data respectively; S02 also includes: For the tank group, suppose there is a tank group set {Tank1, Tank2, … Tank n}, the center position of each tank is obtained by a laser ranging module or a GPS positioning module Wherein, the inter-tank distance D ij (t) can be expressed as follows: wherein, is the coordinate position of tank i, is the coordinate position of tank j; D - all tank-to-tank distances D of the group of tanks ij (t), forming the tank-to-tank position parameter {D ij (t)}.

6. The method for monitoring safety of a group of dangerous product storage tanks according to Claim 5, wherein The Bayesian network nodes include: A bottom plate corrosion state node C; An acoustic event node A; An environmental corrosion factor node E; A distance between storage tanks and linkage risk node D; The joint probability distribution of the Bayesian network BN is constructed, which is defined as follows: P(C,A,E,D)=P(C|A,E,D)P(A|E,D)P(E)P(D) Through inference of the Bayesian network BN, when the current corrosion rate increases, the acoustic event frequency rises, or the environmental parameters are unfavorable, the posterior inference of the Bayesian network BN will give a high probability of a risk state. Finally, the result of the Bayesian network BN is mapped to a risk index R with a value between 0 and 1, which is defined as follows: R=f(P(C=Severe|A,E,D) Where f(·) is a normalization mapping function; S045, the predicted corrosion rate, acoustic event classification result, and risk index R are collected as evaluation results and output.

7. The method for monitoring safety of a group of dangerous product storage tanks according to Claim 6, wherein S05 includes: The evaluation results, the predicted corrosion rate, the acoustic event classification result, and the risk index R are compared with the preset threshold or the reference classification. When they exceed, a warning information is generated and output to realize the early warning of the storage tank and the dangerous goods storage tank group.

8. A dangerous article storage tank storage area management method characterized by, It includes the dangerous goods storage tank group safety monitoring method of any one of claims 1 to 7.

9. A dangerous article storage tank group safety monitoring system which applies the dangerous article storage tank storage area management method according to claim 8, characterized by, It includes: The corrosion monitoring sensor module is multiple, which is deployed on the inner wall and / or outer wall of the storage tank bottom plate, and is used for real-time monitoring of the corrosion condition of the storage tank bottom plate area to generate corrosion monitoring data; The acoustic monitoring sensor module is multiple, which is deployed on the inner wall and / or outer wall of the storage tank bottom plate, and is used for real-time monitoring of the crack initiation and expansion of the bottom plate area of the storage tank to generate acoustic monitoring data; The environmental sensor group is used to monitor the environmental corrosion factors and meteorological conditions around the storage tank and between the storage tanks to generate environmental monitoring data; The data processing module is used to acquire the corrosion monitoring data, acoustic monitoring data, and environmental monitoring data, and performs data cleaning, filtering, and feature extraction, and then performs multi-source data fusion processing to generate multi-source fusion data; The data evaluation module is used to acquire the multi-source fusion data, and uses a risk evaluation algorithm based on machine learning and statistical model to evaluate the corrosion degree of the storage tank and the overall risk of the storage tank group to generate evaluation results; The early warning information unit is configured to obtain the evaluation result, perform early warning on the storage tank and the dangerous goods storage tank group according to the evaluation result, and generate early warning information; The digital twin module is configured to construct a digital twin model based on the dangerous goods storage tank group, and present the evaluation result and the early warning information in a visual form.

Images