Concrete life prediction method based on surface monitoring and physical-data model

Abstract

The invention relates to a concrete service life prediction method based on surface monitoring and a physical-data model, and the method achieves the correlation of an external environment dynamic state and an internal ion transport mechanism through a machine learning model, and enables environment time sequence data which is collected through a non-invasive surface sensor and is easy to obtain to be transmitted to the internal ion transport mechanism. The real-time correction values are intelligently converted into real-time correction values of target chloride ion transportation key parameters which are difficult to measure and dynamically change in a chloride ion diffusion physical model, so that the physical model is driven to carry out high-fidelity life prediction. According to the method, high-precision, lossless and dynamic prediction of the service life of the concrete structure is realized through a collaborative mechanism with a physical model as a framework and machine learning as an engine, and the method has both physical interpretability and intelligence and has important engineering application value.

Description

TECHNICAL FIELD

[0001] The embodiment of the present application relates to the technical field of concrete life prediction, and particularly relates to a concrete life prediction method based on surface monitoring and physical-data model. BACKGROUND

[0002] Reinforced concrete is the most widely used structural material in modern civil engineering, which is widely used in the construction of bridges, ports, tunnels and industrial and civil buildings, and its durability is directly related to the safety of the structure and the national economy. However, in marine environments, cold deicing salt areas and industrial pollution environments, chloride ions in the environment will gradually penetrate into the interior through the micro-pore structure of the concrete. When the invading chloride ions accumulate to a certain critical concentration on the surface of the steel bar, they will destroy the passivation film on the surface of the steel bar and induce electrochemical corrosion of the steel bar. After the steel bar corrodes, the volume expands, which will produce tension on the surrounding concrete, causing the concrete protective layer to crack and peel off, and eventually causing the structural bearing capacity to decrease, and in severe cases, can cause catastrophic accidents.

[0003] Therefore, it is a core task in the field of structural health monitoring and disaster prevention and mitigation to evaluate the durability of concrete structures and scientifically predict their remaining service life in a chloride ion erosion environment. Through accurate life prediction, managers can develop the most cost-effective maintenance and reinforcement strategies to avoid resource waste caused by excessive maintenance and safety hazards caused by disrepair, which has great engineering significance and economic value.

[0004] At present, researchers have proposed various life prediction implementation schemes, which can be mainly divided into two categories: methods based on optimization of traditional physical models and emerging pure data-driven methods, but either are constrained by the static parameter assumption of the physical model and the destructive detection means, resulting in distorted prediction results; or are limited by the "black box" nature and insufficient generalization ability of pure data-driven models, making it difficult to be widely applied in actual engineering. These technical solutions cannot well balance the accuracy of prediction, the convenience of detection and the universality of the model. SUMMARY

[0005] The following is a summary of the subject matter described in detail herein. This summary is not intended to limit the scope of the claims.

[0006] The main purpose of the embodiment of the present disclosure is to propose a concrete life prediction method based on surface monitoring and physical-data model, which can realize non-destructive, high-precision, strong explanatory and adaptive concrete life prediction.

[0007] The first aspect of the present application provides a concrete life prediction method based on surface monitoring and physical-data model, the method comprising: acquire environment parameters of the reinforced concrete within a preset time period; the environment parameters are environment data collected based on a plurality of non-invasive sensors deployed on a structural surface area of the reinforced concrete; construct training data according to the environment parameters, and train a preset machine learning model according to the training data, to train a parameter prediction model capable of learning a nonlinear mapping relationship between time-varying environment parameters of the reinforced concrete and key parameters of chloride ion transport; acquire target environment parameters of a target reinforced concrete; the target environment parameters are environment data collected based on the plurality of non-invasive sensors deployed on a structural surface area of the target reinforced concrete; input the target environment parameters into the parameter prediction model, to obtain time-varying target chloride ion transport key parameters of the target reinforced concrete output by the parameter prediction model; input the time-varying target chloride ion transport key parameters as boundary conditions and model coefficients into a numerical solver based on a chloride ion diffusion physical model, to obtain a chloride ion concentration distribution function at any depth inside the target reinforced concrete by solving the numerical solver; predict a remaining service life of the target reinforced concrete according to the chloride ion concentration distribution function.

[0008] The method provided by the embodiment has at least the following beneficial effects: The method collects time series data on the surface of the reinforced concrete by using non-invasive sensors, and learns a nonlinear mapping relationship between time-varying environment parameters of the reinforced concrete and key parameters of chloride ion transport by using a machine learning model, to obtain a trained parameter prediction model. The parameter prediction model can convert time series environment parameters collected on the surface of target concrete into time-varying target chloride ion transport key parameters. These time-varying dynamic parameters are input into a numerical solver based on a chloride ion diffusion physical model, to calculate an internal chloride ion concentration distribution function, and finally evaluate the remaining service life of the structure.

[0009] The method associates external environment dynamics with internal ion transport mechanisms by using a machine learning model. Time series environment data collected by non-invasive surface sensors and easily obtained are intelligently converted into real-time correction values of target chloride ion transport key parameters that are difficult to measure and dynamically change in the chloride ion diffusion physical model, to drive the physical model to perform high-fidelity service life prediction. Through the collaborative mechanism of the physical model as a skeleton and the machine learning as an engine, high-precision, non-destructive, and dynamic prediction of the service life of the concrete structure is achieved, which has important engineering application value.

[0010] In a second aspect, the application provides a concrete life prediction system based on surface monitoring and physical-data model, comprising: a sensing monitoring subsystem configured to acquire environmental parameters of the reinforced concrete within a preset time period, wherein the environmental parameters are based on data collected by a plurality of non-invasive sensors deployed on a structural surface area of the reinforced concrete, and to acquire target environmental parameters of a target reinforced concrete, wherein the target environmental parameters are based on data collected by the plurality of non-invasive sensors deployed on a structural surface area of the target reinforced concrete; a central processing subsystem configured to construct training data based on the environmental parameters, to train a preset machine learning model based on the training data, to train a parameter prediction model capable of learning a nonlinear mapping relationship between time-varying changes of the environmental parameters of the reinforced concrete and key parameters of chloride ion transport, to input the target environmental parameters into the parameter prediction model to obtain time-varying target chloride ion transport key parameters of the target reinforced concrete output by the parameter prediction model, to input the time-varying target chloride ion transport key parameters as boundary conditions and model coefficients into a numerical solver based on a chloride ion diffusion physical model to obtain a chloride ion concentration distribution function at any depth inside the target reinforced concrete by solving the numerical solver, and to predict a remaining life of the target reinforced concrete based on the chloride ion concentration distribution function.

[0011] In a third aspect, the application provides an electronic device comprising at least one controller and a memory in communication connection with the controller, wherein the memory stores instructions executable by the at least one controller, and the instructions are executed by the at least one controller to cause the at least one controller to perform a concrete life prediction method based on surface monitoring and physical-data model as described above.

[0012] In a fourth aspect, the application provides a computer-readable storage medium storing computer-executable instructions for causing a computer to perform a concrete life prediction method based on surface monitoring and physical-data model as described above.

[0013] Additional aspects and advantages of the application will be made apparent by the following description. BRIEF DESCRIPTION OF DRAWINGS

[0014] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiments or related description will be briefly introduced. Obviously, the drawings in the following description only constitute some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort.

[0015] Figure 1 is a flowchart of a concrete life prediction method based on surface monitoring and physical-data model provided by the embodiments of the present application; Figure 2 is a schematic diagram of a reinforced concrete according to a sensor provided by the embodiments of the present application; Figure 3 is a structural schematic diagram of a concrete life prediction system based on surface monitoring and physical-data model provided by the embodiments of the present application; Figure 4 is a schematic diagram of two stages of training and inference of a machine learning model provided by the embodiments of the present application; Figure 5 is a flowchart of a concrete life prediction method based on surface monitoring and physical-data model provided by another embodiment of the present application; Figure 6 is a prediction diagram of residual life provided by another embodiment of the present application; Figure 7 is a structural schematic diagram of an electronic device provided by another embodiment of the present application. DETAILED DESCRIPTION

[0016] In order to make the purpose, technical solutions and advantages of the present application more clear, the present application will be further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0017] In the description of the present application, if the first, second, etc. are described, it is only for the purpose of distinguishing technical features, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features or the sequence of indicated technical features.

[0018] In the description of the present application, it should be understood that the position description, such as the position or location relationship indicated by up, down, etc. is based on the position or location relationship shown in the drawings, which is only for the convenience of describing the present application and simplifying the description, and is not used to indicate or imply that the indicated device or element must have a specific position, be constructed in a specific position and function, and therefore cannot be understood as a limitation of the present application.

[0019] First, before introducing the embodiments of the present application, necessary background introduction is made: At present, researchers have proposed various implementation schemes for predicting the residual life of reinforced concrete structures, which can be mainly divided into two categories: The first category is based on the optimization of traditional physical models. The core is to use the theoretical framework of Fick's second law, but through on-site core sampling, layered powder grinding and chemical titration to obtain the chloride ion concentration profile of the concrete structure at a specific service time point. Then, through curve fitting, the two key parameters, surface chloride ion concentration (Cs) and apparent chloride ion diffusion coefficient (Dc), are calculated. Input these two parameters into the Fick's law formula, the chloride ion distribution at any future time can be predicted, and the life can be evaluated.

[0020] The defects of this scheme are: (1) This method essentially uses the detection results at a certain moment to predict the changes in the future tens or even hundreds of years. It regards the surface chloride ion concentration (Cs) and apparent chloride ion diffusion coefficient (Dc) calculated at the detected moment as constant constants. However, in reality, these two parameters are time-varying dynamic quantities. For example, the diffusion coefficient Dc fluctuates with seasonal temperature and humidity changes; the surface concentration Cs changes dramatically due to rain season erosion and winter deicing salt spraying. Simplifying the dynamic process into static parameters is the root cause of the large deviation between the prediction results of this method and the actual situation.

[0021] (2) Destructive testing and local representativeness limitation: core sampling is a destructive test that will cause irreversible damage to the structure, and the detection cost is high and the process is complicated. More importantly, whether the data obtained from a few sampling points can represent the true state of the entire component or even the entire structure is highly uncertain. This point-to-surface method ignores the non-uniformity of deterioration caused by differences in light, wind, dry-wet cycle, etc. in different parts of the structure.

[0022] The second category is the emerging pure data-driven method. That is, directly build an "end-to-end" machine learning model (such as artificial neural network , support vector machine , etc.) to predict the residual life of reinforced concrete. This method collects a large amount of existing experimental data or engineering cases, and through training, the machine learning model learns the nonlinear mapping relationship between the input and output independently.

[0023] The defects are: (1) Lack of physical meaning and poor generalization ability; (2) Over-reliance on high-quality big data.

[0024] To solve the defects of the above two methods, an embodiment of the present application is as follows: Figure 1As shown, a concrete life prediction method based on surface monitoring and physical-data model is provided, and the method comprises: In step S110, the environmental parameters of the reinforced concrete in a preset time period are acquired; the environmental parameters are environmental data collected based on a plurality of non-invasive sensors deployed on the structural surface area of the reinforced concrete; In step S120, training data are constructed according to the environmental parameters, and a preset machine learning model is trained according to the training data, so as to train a parameter prediction model capable of learning the nonlinear mapping relationship between the time sequence variation of the environmental parameters of the reinforced concrete and the chloride ion transport key parameters; In step S130, target environmental parameters of a target reinforced concrete are acquired; the target environmental parameters are environmental data collected based on a plurality of non-invasive sensors deployed on the structural surface area of the target reinforced concrete; In step S140, the target environmental parameters are input into the parameter prediction model, so as to obtain time-varying target chloride ion transport key parameters of the target reinforced concrete output by the parameter prediction model; In step S150, the time-varying target chloride ion transport key parameters are input into a numerical solver based on a chloride ion diffusion physical model as boundary conditions and model coefficients, so as to obtain a chloride ion concentration distribution function at any depth inside the target reinforced concrete by the numerical solver; In step S160, the residual life of the target reinforced concrete is predicted according to the chloride ion concentration distribution function.

[0025] Steps S110 and S120 are a model training process, and the purpose is to train a parameter prediction model capable of predicting chloride ion transport key parameters having linear correlation with the time sequence variation of the environmental parameters based on the environmental parameters of the structural surface area of the reinforced concrete. The parameters are used in the subsequent physical-level prediction process, which will be described in detail in the subsequent embodiment.

[0026] As shown, Figure 2 Firstly, the collection of the environmental parameters is required, the reinforced concrete is determined, then a plurality of non-invasive sensors are deployed on the structural surface area of the reinforced concrete (here, the sensors are arranged on the surface area), and then the environmental parameters are collected by the plurality of non-invasive sensors. For example, the surface temperature is required to be collected, so a temperature sensor is arranged, or the surface humidity is required to be collected, so a humidity sensor is arranged. The plurality of sensors can be integrated sensors or independent sensors. It should be noted that the data in a period of time are required to be collected to explore the time sequence variation of the parameters.

[0027] Then a machine learning model (such as a long short-term memory network) is selected, training data is constructed through the environmental parameters, and corresponding labels are set. Finally, the machine learning model is trained based on the training data, so that the machine learning model can learn the nonlinear mapping relationship between the time-varying changes of the environmental parameters of the reinforced concrete and the chloride ion transport key parameters. Finally, after the training is completed, the trained machine learning model is used as a parameter prediction model.

[0028] In subsequent steps S130 to S160, it is the use process of the model. First, the target reinforced concrete refers to the reinforced concrete structure that needs to be predicted for the remaining life. The collection method of the target environmental parameters of the target reinforced concrete is similar to that of step S110.

[0029] In step S140, the target environmental parameters are input into the parameter prediction model, and the time-varying target chloride ion transport key parameters of the target reinforced concrete can be obtained. The time-varying target chloride ion transport key parameters are used as key parameters for subsequent physical level prediction.

[0030] In step S150, the numerical solver is a mathematical model in which a chloride ion diffusion physical model is set. By solving the chloride ion diffusion physical model, the law of chloride ion diffusion can be known, which is convenient for subsequent remaining life calculation. Here, the time-varying target chloride ion transport key parameters are input into the numerical solver based on the chloride ion diffusion physical model as boundary conditions and model coefficients, so as to obtain the chloride ion concentration distribution function at any depth inside the target reinforced concrete by solving the numerical solver.

[0031] In step S160, the remaining life of the target reinforced concrete is predicted according to the chloride ion concentration distribution function. For example, there is a critical chloride ion concentration threshold value for existing steel bar corrosion. When the threshold value is reached in the chloride ion concentration distribution function solution, a time node corresponding to the threshold value can be considered as the total life length of the structure. Then, the use length of the target reinforced concrete is queried, and the final remaining life can be obtained by subtracting the used time from the total life length.

[0032] The concrete life prediction method based on surface monitoring and physical-data model provided in the embodiment has the following advantages: The method collects time series data of the reinforced concrete surface through non-invasive sensors, and learns the nonlinear mapping relationship between the time series changes of the environmental parameters of the reinforced concrete and the key parameters of chloride ion transport by using a machine learning model, to obtain a trained parameter prediction model. Through the parameter prediction model, the time series data collected on the target concrete surface area can be converted into time-varying target chloride ion transport key parameters. These time-varying dynamic parameters are input into a numerical solver based on the chloride ion diffusion physical model to calculate the internal chloride ion concentration distribution function, and finally evaluate the remaining life of the structure.

[0033] The method realizes the association of external environmental dynamics and internal ion transport mechanism by using a machine learning model. The easily obtained environmental time series data collected by non-invasive surface sensors are intelligently converted into real-time correction values of the target chloride ion transport key parameters (such as chloride ion diffusion coefficient and surface concentration) which are difficult to measure and dynamically change in the chloride ion diffusion physical model, thereby driving the physical model to perform high-fidelity life prediction. Through the cooperative mechanism of physical model as the skeleton and machine learning as the engine, high-precision, non-destructive and dynamic prediction of the life of the concrete structure is realized, which has important engineering application value.

[0034] In some embodiments of the present application, the plurality of non-invasive sensors at least include: a surface temperature sensor, a surface relative humidity sensor, and a surface chloride ion concentration sensor. The environmental parameters at least include: surface temperature, surface relative humidity, and surface chloride ion concentration.

[0035] In this embodiment, it has been found that there is an important linear relationship between the surface temperature, the surface relative humidity, and the surface chloride ion concentration and the key parameters of chloride ion transport. Therefore, the environmental parameters at least include: surface temperature, surface relative humidity, and surface chloride ion concentration, which can improve the accuracy of the parameter prediction model in predicting the key parameters of chloride ion transport.

[0036] In some embodiments, the surface temperature sensor, the surface relative humidity sensor, and the surface chloride ion concentration sensor can be integrated sensors or independent sensors.

[0037] In some embodiments, the most representative positions of the reinforced concrete surface area with the worst environmental effects are selected for deployment, such as the windward surface of the building, the splash zone, or the tidal zone in the offshore platform structure.

[0038] In some embodiments of the present application, the surface temperature sensor is a platinum resistance (e.g., PT1000) or an NTC thermistor, and the measurement range of the surface temperature sensor covers -40 degrees Celsius to 80 degrees Celsius. The surface relative humidity sensor is a capacitive humidity sensor, and the measurement range of the surface relative humidity sensor covers 0 to 100% relative humidity. The surface chloride ion concentration sensor is a solid, maintenance-free ion-selective electrode. The free chloride ion concentration in the liquid film can be measured in situ and in real time.

[0039] The sensor can be fixed to the surface of the reinforced concrete by high-strength weather-resistant adhesive or mechanical anchoring in a non-destructive manner. The selection of the sensor in the present method can provide more accurate environmental parameters to improve the accuracy of subsequent model prediction of chloride ion transport key parameters.

[0040] In some embodiments of the present application, the machine learning model is a long short-term memory network.

[0041] Compared with conventional machine learning models, the long short-term memory network (LSTM) is a special recurrent neural network (RNN) whose unique gating mechanism (input gate, forget gate, and output gate) enables it to effectively learn and remember long-term dependencies in time series data. This is crucial for the application scenario of the present method, because the chloride ion diffusion performance of concrete is not only affected by the current environment, but also by the cumulative effects of the humidity history of the past period (several days or even weeks) (such as the water saturation degree of the pore structure and the evolution of the hydration products).

[0042] In some embodiments of the present application, the chloride ion transport key parameters include time-varying effective chloride ion diffusion coefficient and time-varying surface chloride ion concentration. In step S120, the training data is constructed according to the environmental parameters in the preset time period. Before training the preset machine learning model according to the training data, the method further comprises: In step S210, a rapid chloride ion migration coefficient test is performed on the reinforced concrete to obtain an instantaneous migration coefficient that reflects the compactness of the reinforced concrete. In step S220, the instantaneous migration coefficient is converted into an instantaneous effective chloride ion diffusion coefficient. In step S120, the training data is constructed according to the environmental parameters in the preset time period. Before training the preset machine learning model according to the training data, the method further comprises: In step S1210, the instantaneous effective chloride ion diffusion coefficient is paired with the surface temperature, the surface relative humidity, and the surface chloride ion concentration to form a training sample. In step S1220, based on the training samples, the long short-term memory network is trained using supervised learning, and the backpropagation algorithm is used to adjust the weights of the long short-term memory network until the error between the predicted result output by the long short-term memory network and the true value is minimized.

[0043] In this embodiment, a rapid chloride ion migration test (RCM) was performed on reinforced concrete, following the NT BUILD 492 standard. The RCM method accelerates chloride ion migration by applying an external electric field, enabling rapid measurement of an instantaneous migration coefficient that reflects the current density of the concrete, which can then be converted into an effective chloride ion diffusion coefficient.

[0044] In some embodiments of this application, the physical model for chloride ion diffusion is a one-dimensional Fick's second law partial differential equation with time-varying coefficients; Step S140, which involves obtaining the chloride ion concentration distribution function at any depth inside the target reinforced concrete using a numerical solver, includes: Step S1410: Set initial conditions and boundary conditions; Step S1420: Based on the finite difference method, the partial differential equation of the first-dimensional Fick's second law is iteratively calculated with discrete time steps to obtain the chloride ion concentration distribution function at any depth inside the target reinforced concrete.

[0045] In this embodiment, the physical simulation can be solving a one-dimensional Fick's second law partial differential equation with time-varying coefficients: ; in, It is depth In time chloride ion concentration, It is the time-varying effective chloride ion diffusion coefficient obtained.

[0046] The Finite Difference Method (FDM) is preferably used to numerically solve the above equations. Specifically, the concrete cover is considered in the depth direction ( (axis) and time direction ( Discretize along the axis to form a two-dimensional grid.

[0047] Boundary conditions and initial conditions, for example: Initial conditions: ,in It is the initial chloride ion concentration in reinforced concrete (usually 0 or a very small value).

[0048] Boundary condition 1 (surface): ,in It is the time-varying surface chloride ion concentration obtained.

[0049] Boundary Condition 2 (Internal): Assuming the depth of the protective layer is much smaller than the structural dimensions, a zero-flux boundary condition can be set at the other end of the computational domain, i.e. .

[0050] Time-stepping solution process: at each time step At that moment and Then, Update the coefficients of the finite difference scheme, and... As The Dirichlet boundary conditions at that point are determined; subsequently, a difference iteration is performed to solve for the boundary conditions at the end of that time step. ), all depth nodes The new chloride ion concentration By repeating this process, the complete dynamic process of chloride ion concentration distribution within concrete evolving over time can be simulated.

[0051] In some embodiments of this application, step S160, predicting the remaining life of the target reinforced concrete based on the chloride ion concentration distribution function, includes: Step S1610: Determine the thickness of the concrete cover for the target reinforced concrete and the critical chloride ion concentration threshold for steel corrosion. Step S1620: Calculate the chloride ion concentration corresponding to the thickness of the steel reinforcement protective layer according to the chloride ion concentration distribution function, and determine the corresponding time when the chloride ion concentration first exceeds or equals the critical chloride ion concentration threshold for steel reinforcement corrosion. Step S1630: Determine the remaining life of the target reinforced concrete based on the time and the service time of the target reinforced concrete.

[0052] In this embodiment, the thickness of the reinforcing bar protective layer refers to the distance from the outer surface of the reinforcing bar to the exposed concrete surface; the critical chloride ion concentration threshold for reinforcing bar corrosion is a widely accepted empirical value.

[0053] At each time step of the physical model solution, the chloride ion concentration at the depth of the reinforcing steel is queried based on the chloride ion concentration distribution function. When this value first reaches or exceeds the preset critical chloride ion concentration threshold, the corresponding time is determined as the predicted total lifespan of the reinforced concrete.

[0054] Finally, determine the service life of the target reinforced concrete since its construction. Subtract the service life from the total service life of the reinforced concrete to obtain the remaining service life of the target reinforced concrete.

[0055] As Figure 3 shown in the drawings, one embodiment of the present application provides a concrete life prediction system based on surface monitoring and physical-data model, the system comprises: The sensing monitoring subsystem 100 is configured to acquire environmental parameters of the reinforced concrete within a preset time period; the environmental parameters are environmental data collected based on a plurality of non-invasive sensors deployed on the structural surface area of the reinforced concrete, and the target environmental parameters of the target reinforced concrete; the target environmental parameters are data collected based on a plurality of non-invasive sensors deployed on the structural surface area of the target reinforced concrete; The central processing subsystem 200 is configured to construct training data according to the environmental parameters, train a preset machine learning model according to the training data, to train a parameter prediction model capable of learning the nonlinear mapping relationship between the time-varying changes of the environmental parameters of the reinforced concrete and the key parameters of the chloride ion transport, and input the target environmental parameters into the parameter prediction model to obtain the time-varying target chloride ion transport key parameters of the target reinforced concrete output by the parameter prediction model, and input the time-varying target chloride ion transport key parameters as boundary conditions and model coefficients into a numerical solver based on a chloride ion diffusion physical model, to obtain the chloride ion concentration distribution function at any depth inside the target reinforced concrete by solving the numerical solver, and predict the remaining life of the target reinforced concrete according to the chloride ion concentration distribution function.

[0056] It should be noted that the concrete life prediction system based on surface monitoring and physical-data model provided in the embodiment and the concrete life prediction method based on surface monitoring and physical-data model described above are based on the same inventive concept, and therefore the content of the concrete life prediction system based on surface monitoring and physical-data model described in the embodiment is also applicable to the content of the concrete life prediction method based on surface monitoring and physical-data model described above, which will not be described here.

[0057] For ease of understanding, the present application provides a set of best embodiments, which provide a concrete life prediction system based on surface monitoring and physical-data model and a concrete life prediction method based on surface monitoring and physical-data model implemented by the system, specifically comprising the following: As Figure 3 shown, the system comprises: The sensing monitoring subsystem 100 is disposed on the structural surface of the concrete, and internally integrated with a surface temperature sensor 102, a surface humidity sensor 103 and a surface chloride ion concentration sensor 104, configured to perform a non-invasive data acquisition step; Central processing subsystem 200: in communication connection with the sensing monitoring subsystem 100, for receiving and processing the collected data. The central processing subsystem is internally configured with: Parameter prediction unit 210: embedded with a pre-trained machine learning model (named as parameter prediction model after training), for receiving the time series data set and performing the dynamic parameter correction step, outputting the dynamic And .

[0058] Physical simulation unit 220: embedded with a numerical solver of the second law of Fick, for receiving the dynamic parameters output by the parameter prediction unit 210 and performing the physical model solving step.

[0059] Storage unit 300: for storing historical monitoring data, pre-trained machine learning model and prediction result data.

[0060] Output and display unit 400: for showing the calculated structure residual life, chloride concentration distribution cloud map and other prediction results to the user, and can issue a warning according to the preset conditions, such as a Web application or a mobile APP.

[0061] The implementation method is as follows: Step S910, non-invasive time series environmental data collection.

[0062] It is executed by the sensing monitoring subsystem 100 laid on the surface area of the reinforced concrete structure. The installation method of the sensor: select the most representative position in the concrete where the environmental action is the worst, such as the windward surface of the building, the splash zone, or the tidal zone in the offshore platform structure. Through high-strength weather-resistant adhesive or mechanical anchoring method, it is fixed non-destructively on the concrete surface; The composition of the sensing monitoring subsystem 100: In an integrated, weather-resistant shell with an IP67 protection level, at least including: Surface temperature sensor 102: preferably high-precision platinum resistance (such as Pt1000) or NTC thermistor, ensuring that the measurement range covers -40℃ to +80℃, and the measurement accuracy is better than ±0.1℃.

[0063] Surface relative humidity sensor 103: preferably a capacitive humidity sensor, with a measurement range of 0-100% RH and an accuracy better than ±2% RH.

[0064] Surface chloride ion concentration sensor 104: Preferably, a solid-state, maintenance-free ion-selective electrode is adopted. The sensor is based on Ag / AgCl electrochemical principle, and its sensitive membrane is directly in contact with the trace moisture or thin liquid film attached to the concrete surface, so as to measure the free chloride ion activity (concentration) in the liquid film in situ and in real time.

[0065] Microcontroller unit MCU 105: A low-power MCU, such as the STM32L4 series, is selected. It is responsible for controlling the acquisition timing of each sensor according to the preset program, performing preliminary processing and digitization of the original signal, and attaching an accurate time stamp to each data point.

[0066] Power supply unit 106: To meet the long-term unattended monitoring requirements, a small solar panel combined with a rechargeable lithium battery (such as a 18650 type) is preferably adopted to achieve energy self-sufficiency.

[0067] Wireless communication unit 107: A low-power wide-area network (Low-Power Wide-Area Network, LPWAN) communication module, especially a LoRaWAN module, is preferably adopted. The reason for choosing LoRaWAN is its excellent long-distance transmission capability, extremely low power consumption, and strong signal penetration, making it very suitable for application in large, dispersed, and difficult-to-power and network-cabling infrastructure monitoring scenarios.

[0068] Data synchronization and transmission: The microcontroller unit wakes up all sensors at a preset time frequency, synchronously acquires data once. The data packet (including temperature, humidity, chloride ion concentration, and time stamp) is sent wirelessly to the remote central processing subsystem 200 through the wireless communication unit 107.

[0069] Step S920, dynamic parameter correction step based on machine learning model.

[0070] This step is performed by the parameter prediction unit 201 in the central processing subsystem 200 deployed in the cloud server or local gateway.

[0071] Step S921, selection and structure of machine learning model; Referring to Figure 4 As shown, a long short-term memory network (Long Short-Term Memory, LSTM) is preferably adopted as the core algorithm.

[0072] Input and output of the model: The parameter prediction unit 201 receives a continuous multi-dimensional time series data sequence from the sensor monitoring subsystem 100, such as a data point sequence within a sliding time window. The model processes the sequence and outputs two time points Corresponding instantaneous physical parameter value: time-varying effective chloride ion diffusion coefficient and time-varying surface chloride ion concentration .

[0073] Step S922, model training data generation.

[0074] This is the core link to ensure that the machine learning model has physical meaning and prediction ability. In order to make the model fully public and reproducible, its training process is as follows: Test piece preparation: According to the common proportion in engineering, prepare multiple groups of concrete cylinder test pieces, covering different water-cement ratios, and adding common mineral admixture fly ash or slag variables.

[0075] Controlled environment simulation: Place the test pieces in a large programmable environmental test box. In the box, a pre-designed, dynamically changing temperature and humidity cycle program is applied to simulate typical daily temperature difference, seasonal change, dry-wet alternation and other real environmental conditions. At the same time, through the automatic spraying system, periodic changes of chloride salt solution (such as NaCl solution) are applied on the surface of the test piece to simulate rainfall or tidal action. The temperature, humidity in the environmental box and the chloride salt concentration on the surface of the test piece are recorded continuously at high frequency to form an accurate environmental parameter history sequence.

[0076] Fast determination of instantaneous diffusion coefficient: At multiple preset time points during environmental simulation, a part of parallel test pieces are randomly taken out from the environmental box. Immediately, these test pieces are tested for rapid chloride ion migration (RCM), which follows the NT BUILD 492 standard.

[0077] Construct training data pairs: pair the instantaneous effective diffusion coefficient value measured at each time point by RCM method with a segment of environmental parameter history sequence before its measurement to form a complete training sample.

[0078] Model supervised learning: by repeating the above steps, a data set containing hundreds or even thousands of training samples is constructed. Using this data set, the LSTM network is trained in a supervised learning manner, and the network weights are adjusted through the back propagation algorithm until the error (such as root mean square error) between the predicted value output by the model and the true value measured by RCM method is minimized.

[0079] The LSTM model trained through the above process internalizes the complex, nonlinear mapping relationship between the easily measurable surface environmental time series data and the difficult to directly measure internal key physical parameters. For , the model also learns the correction relationship between the surface sensor measured value and the effective boundary concentration inside the surface layer of concrete that really drives diffusion.

[0080] Step S930: Obtain the environmental parameters of the target reinforced concrete and predict the dynamic parameters based on the parameter prediction model trained in the above steps.

[0081] Step S940: Solve the physical model based on dynamic parameters.

[0082] This step is performed by the physical simulation subunit 202 in the central processing subsystem 200.

[0083] Governing equation: The core of physical simulation is solving a one-dimensional Fick's second law partial differential equation with time-varying coefficients. ; in, It is depth In time chloride ion concentration, It is the time-varying diffusion coefficient obtained in real time from parameter prediction subunit 201.

[0084] Numerical solver: The finite difference method (FDM) is preferably used to solve the above equations numerically.

[0085] Boundary conditions and initial conditions: Initial conditions: ; Boundary condition 1 (surface): ; Boundary condition 2 (interior): ; Time-stepping solution process: at each time step At that moment and Then, Update the coefficients of the finite difference scheme, and... As The Dirichlet boundary conditions at that point are determined; subsequently, a difference iteration is performed to solve for the boundary conditions at the end of that time step. ), all depth nodes The new chloride ion concentration By repeating this process, the complete dynamic process of chloride ion concentration distribution within concrete evolving over time can be simulated.

[0086] Step S950, assessment of the remaining service life of the concrete structure.

[0087] The process is completed within the central processing subsystem 200, and the results are transmitted to the output and display module 400.

[0088] Input key thresholds: Based on the design drawings or relevant specifications of the structure to be evaluated, determine two key parameters: Reinforcement cover thickness : the distance from the outer surface of the steel bar to the exposed surface of the concrete.

[0089] Critical chloride concentration threshold for steel bar corrosion : This is a widely accepted empirical value, usually taken as 0.4% relative to the mass of cement or 0.05% relative to the mass of concrete.

[0090] Life prediction and result output: At each time step of the physical model solution, the system checks the chloride concentration at the depth of the steel bar . When this value first reaches or exceeds the preset critical threshold , the corresponding time is determined as the predicted total life of the structure . The remaining service life (RSL) of the structure is calculated by the following formula: ; Where, is the time used since the construction of the target reinforced concrete.

[0091] Finally, the predicted remaining life, risk warning information, and the dynamic distribution cloud of the chloride ion concentration in the concrete are presented to the user in a graphical manner through the output and display module 400.

[0092] The actual beneficial effects of the embodiments of the application are: The embodiments establish a nonlinear mapping relationship between the real-time environmental parameter sequence and the real-time physical parameters through the machine learning model. This means that when the sensor monitors the night cooling and the humidity rises, the machine learning model can immediately output a reduced diffusion coefficient value that matches it; when the summer high temperature or salt concentration increases, the model will output an increased diffusion coefficient value. This dynamic correction of parameters enables the physical model to perceive changes in the external environment in real time, and its simulation process infinitely approximates the real physical process, so the prediction accuracy is fundamentally improved; The embodiments use completely non-invasive surface sensors, completely avoiding detection methods such as core sampling that cause permanent damage to the structure, and the environmental parameters collected by the non-invasive surface area arrangement of the sensors are more representative of the true state of the entire reinforced concrete structure or even the entire reinforced concrete structure. Combined with low-power wireless communication technology, it can realize non-destructive, continuous, and automatic monitoring and evaluation of the health status of the in-service reinforced concrete structure, providing a technical foundation for the whole life cycle of structure management; Instead of replacing the classical physical laws with an uninterpretable black box model, the present embodiment uses the Fick's second law as the theoretical framework, and only uses the machine learning model as an advanced intelligent parameterization engine. The final prediction results (such as the chloride ion concentration distribution) have clear physical connotations; at the same time, the introduction of machine learning technology enables it to adaptively learn the complex effects under different environments and different material proportions, and has a high degree of flexibility that pure physical models do not have. The task target of the machine learning model in the present embodiment is more focused and clear, i.e., to predict the physical parameter and . Since the learning task is strongly constrained by the physical law, the model no longer needs to blindly find statistical rules from massive data, thereby significantly reducing the demand for the amount of training data. More importantly, since the output of the model is always constrained by a universal physical diffusion equation, its reliability for long-term prediction and ability to generalize to new working conditions far exceed pure data-driven models that lack physical constraints.

[0093] Table 1 below summarizes the comparative advantages of the present embodiment over two main existing technologies.

[0094] Table 1

[0095] To verify the technical effect of the present embodiment, a comparative test was conducted. Three methods were selected to predict the chloride ion penetration depth of C30 concrete specimens exposed for 180 days under the same dynamic environment (simulating the dry-wet, temperature and humidity, and salt spray cycle of ocean tides). The final depth was compared with the depth measured by the silver nitrate colorimetric method.

[0096] Pure physical model: using the classical Fick's law, the diffusion coefficient and the surface concentration According to the NT BUILD 492 standard rapid chloride ion migration coefficient test, the test follows the standard to set a constant after the initial 28-day test, which are =10.5×10-12 and =0.8%.

[0097] Pure data-driven model: using a pure data-driven ANN model, the average temperature, average humidity, total salt spray amount, and other macro-environmental parameters within 180 days are taken as inputs, and the final penetration depth is directly output. The model has been trained with a large amount of historical static data.

[0098] The present embodiment: using the physical-data hybrid model of the present embodiment, real-time acquisition of environmental time series data by sensors deployed on the concrete surface, dynamic correction of physical model parameters and simulation prediction.

[0099] After the test, the test piece was cut and sprayed with 0.1 mol / L silver nitrate solution, and the boundary of chloride ion penetration was clearly visible. The true penetration depth was obtained by multiple measurements with a vernier caliper and averaging.

[0100] Table 2

[0101] From the comparative data in Table 2, the following conclusions can be clearly drawn: The traditional physical model cannot capture the influence of dynamic environment on accelerated chloride ion erosion (for example, alternating temperature and humidity can significantly exacerbate erosion) due to the use of static parameters. The relative error of 35.2% between the predicted value of 8.1 mm and the measured value of 12.5 mm proves that this method is unreliable in real and variable environments.

[0102] Although the pure data model takes into account environmental factors, it loses the timing dynamic information in the process because it inputs macroscopic average parameters, and the generalization ability of the model is limited, resulting in a certain deviation in the predicted result of 14.2 mm, with a relative error of 13.6%. Although it is better than the traditional physical model, the precision still needs to be improved.

[0103] In this embodiment, the environmental dynamics are obtained through real-time surface monitoring, and the machine learning model is used to convert the dynamics into real-time corrections to the physical model parameters. The predicted result of 12.8 mm is highly consistent with the measured value of 12.5 mm, with an absolute error of only 0.3 mm and a relative error as low as 2.4%.

[0104] In summary, the experimental data strongly prove that the physical-data hybrid driving method proposed in this embodiment has obvious advantages in prediction accuracy compared to existing technologies (i.e., traditional physical model and pure data model), enabling high-fidelity prediction of concrete structure life under real chloride salt environment, and has high engineering application value.

[0105] The following provides two application scenarios of this embodiment: (1) Improve public safety: The durability failure of major infrastructure such as bridges, ports, and nuclear power plant containment structures can have catastrophic consequences. This embodiment provides a reliable "health barometer" that provides a scientific basis for the safe operation of these critical structures, ensuring the safety of life and property of the public.

[0106] (2) Promote industry intelligent transformation: This invention is a typical example of the combination of Internet of Things, artificial intelligence and traditional civil engineering, providing a feasible technical solution for the intelligent and automated health management of large infrastructure networks, and is an important part of building a smart city and smart transportation system.

[0107] For example,Figure 7 One embodiment of the present application also provides an electronic device, which comprises a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-mentioned method for predicting the service life of concrete based on surface monitoring and physical-data model. The electronic device comprises: at least one battery; at least one memory; at least one processor; at least one program; The program is stored in the memory, and the processor executes the at least one program to implement the above-mentioned method for predicting the service life of concrete based on surface monitoring and physical-data model.

[0108] The electronic device can be any intelligent terminal, such as a mobile phone, a tablet computer, a personal digital assistant (PDA), a vehicle-mounted computer, etc.

[0109] The electronic device of the embodiment of the present application will be described in detail below.

[0110] The processor 1600 can be implemented in the form of a general central processing unit (CPU), a microprocessor, an application specific integrated circuit (ASIC), or one or more integrated circuits, etc., and is used to execute related programs to implement the technical solutions provided by the embodiments of the present application. The memory 1700 can be implemented in the form of a read only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM), etc. The memory 1700 can store an operating system and other application programs. When the technical solutions provided by the embodiments of the present application are implemented by software or firmware, the related program codes are saved in the memory 1700 and are called and executed by the processor 1600 to implement the method for predicting the service life of concrete based on surface monitoring and physical-data model.

[0111] The input / output interface 1800 is used to realize information input and output. The communication interface 1900 is used to realize the communication interaction between the device and other devices. The communication can be realized by wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.). A bus 2000 transmits information between various components (for example, the processor 1600, the memory 1700, the input / output interface 1800, and the communication interface 1900) of the device. The processor 1600, the memory 1700, the input / output interface 1800, and the communication interface 1900 are communicatively connected to each other within the device through the bus 2000.

[0112] The disclosure embodiments also provide a storage medium, which is a computer-readable storage medium, and stores computer executable instructions for causing a computer to execute the above-mentioned concrete surface monitoring and physical-data model based concrete service life prediction method.

[0113] The memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer executable programs. In addition, the memory can include a high-speed random access memory, and can also include a non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state memory device. In some embodiments, the memory can optionally include a memory remotely arranged relative to the processor, and these remote memories can be connected to the processor through a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0114] The embodiments described in the disclosure embodiments are for more clearly illustrating the technical solutions of the disclosure embodiments, and do not constitute a limitation on the technical solutions provided by the disclosure embodiments. Those skilled in the art can know that, with the evolution of technology and the appearance of new application scenarios, the technical solutions provided by the disclosure embodiments are also applicable to similar technical problems.

[0115] Those skilled in the art can understand that the technical solutions shown in the figures do not constitute a limitation on the disclosure embodiments, and can include more or fewer steps than those shown in the figures, or combine certain steps, or different steps.

[0116] The device embodiments described above are only schematic, and the units described as separate components can or can not be physically separate, that is, can be located in one place, or can be distributed on multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiments.

[0117] Those skilled in the art can understand that all or some of the steps in the above-mentioned method, the functional modules / units in the system and the device can be implemented as software, firmware, hardware, and appropriate combinations thereof.

[0118] The terms "first", "second", "third", "fourth", and the like in the description of this application and in the claims hereof, if any, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the use of the terms so termed herein is solely for the convenience of the reader and does not limit the scope of the application. It is also to be understood that the description and examples in this application are intended to cover all possible combinations where any of the several elements can represent one or more elements.

[0119] It should be understood that, in this application, "at least one" means one or more, "multiple" means two or more. "And / or" is used to describe the relationship between associated objects, which means that there can be three relationships, for example, "A and / or B" can mean: only A, only B, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally represents an "or" relationship between the associated objects. "At least one of the following" or similar expressions means any combination of these items, including any combination of single or multiple items. For example, at least one of a, b or c, can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0120] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only illustrative, for example, the division of units is only a logical function division, and actual implementation can have another division manner, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed units can be indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0121] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or they can be distributed on multiple network units. According to actual needs, some or all of the units can be selected to achieve the purpose of the embodiment of the present application.

[0122] 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.

[0123] When 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 multiple instructions for causing an electronic device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of the embodiments of the present application. The aforementioned storage medium includes various media that can store program codes, such as a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0124] The above is a specific description of the preferred implementation of the embodiments of the present application, but the embodiments of the present application are not limited to the above implementation. Those skilled in the art can make various equivalent modifications or replacements without departing from the spirit of the embodiments of the present application, and these equivalent modifications or replacements are all included in the scope defined by the claims of the embodiments of the present application.

Claims

1. A method for concrete life prediction based on surface monitoring and physics-data model, characterized in that, The method comprises: obtaining environmental parameters of the reinforced concrete within a preset time period; the environmental parameters are environmental data collected based on a plurality of non-invasive sensors deployed on a structural surface area of the reinforced concrete; constructing training data according to the environmental parameters, and training a preset machine learning model according to the training data to obtain a parameter prediction model capable of learning a nonlinear mapping relationship between time-varying changes of the environmental parameters of the reinforced concrete and key parameters of chloride ion transport; obtaining target environmental parameters of a target reinforced concrete; the target environmental parameters are environmental data collected based on the plurality of non-invasive sensors deployed on a structural surface area of the target reinforced concrete; inputting the target environmental parameters into the parameter prediction model to obtain time-varying target chloride ion transport key parameters of the target reinforced concrete output by the parameter prediction model; inputting the time-varying target chloride ion transport key parameters as boundary conditions and model coefficients into a numerical solver based on a chloride ion diffusion physical model to obtain a chloride ion concentration distribution function at any depth inside the target reinforced concrete by solving the numerical solver; predicting the remaining life of the target reinforced concrete according to the chloride ion concentration distribution function.

2. The concrete life prediction method based on surface monitoring and physical-data model according to claim 1, characterized in that: the plurality of non-invasive sensors at least include a surface temperature sensor, a surface relative humidity sensor, and a surface chloride ion concentration sensor; the environmental parameters at least include surface temperature, surface relative humidity, and surface chloride ion concentration.

3. The method for concrete life prediction based on surface monitoring and physics-data model according to claim 2, characterized in that, the surface temperature sensor is a platinum resistance or an NTC thermistor, and the measurement range of the surface temperature sensor covers -40 degrees Celsius to 80 degrees Celsius; the surface relative humidity sensor is a capacitive humidity sensor, and the measurement range of the surface relative humidity sensor covers 0 to 100% relative humidity; the surface chloride ion concentration sensor is a solid, maintenance-free ion-selective electrode.

4. The method for predicting the service life of concrete based on surface monitoring and physical-data model according to claim 2, characterized in that, the machine learning model is a long short-term memory network.

5. The method for concrete life prediction based on surface monitoring and physics-data model according to claim 4, characterized in that, the chloride ion transport key parameters include time-varying effective chloride ion diffusion coefficient and time-varying surface chloride ion concentration; before the step of constructing training data according to the environmental parameters within the preset time period and training the preset machine learning model according to the training data, the method further comprises: performing a rapid chloride ion migration coefficient test on the reinforced concrete to obtain an instantaneous migration coefficient reflecting the compactness of the reinforced concrete; converting the instantaneous migration coefficient into an instantaneous effective chloride ion diffusion coefficient; the step of constructing training data according to the environmental parameters within the preset time period and training the preset machine learning model according to the training data comprises: pairing the instantaneous effective chloride ion diffusion coefficient with the surface temperature, the surface relative humidity, and the surface chloride ion concentration to form training samples; Based on the training sample, the long short-term memory network is trained in a supervised learning manner, and a back propagation algorithm is used to adjust the weight parameters of the long short-term memory network until the error between the prediction result and the true value output by the long short-term memory network is minimized.

6. The method for concrete life prediction based on surface monitoring and physics-data model according to claim 5, characterized in that, The chloride ion diffusion physical model is a one-dimensional Fick's second law partial differential equation with time-varying coefficients; The chloride ion concentration distribution function at any depth inside the target reinforced concrete is obtained by solving the numerical solver, including: Setting initial conditions and boundary conditions; Based on the finite difference method, the one-dimensional Fick's second law partial differential equation is iteratively calculated at discrete time steps to obtain the chloride ion concentration distribution function at any depth inside the target reinforced concrete.

7. A method for concrete life prediction based on surface monitoring and physics-data model according to claim 6, characterized in that, The residual life of the target reinforced concrete is predicted according to the chloride ion concentration distribution function, including: Determining the reinforcement cover thickness of the target reinforced concrete and the critical chloride ion concentration threshold of reinforcement corrosion; According to the chloride ion concentration distribution function, the chloride ion concentration corresponding to the reinforcement cover thickness is calculated, and the corresponding time is determined when the chloride ion concentration is greater than or equal to the critical chloride ion concentration threshold of reinforcement corrosion for the first time; Based on the time and the service time of the target reinforced concrete, the residual life of the target reinforced concrete is determined.

8. A surface monitoring and physics-data model based concrete life prediction system, characterized by, The system comprises: A sensing and monitoring subsystem for acquiring environmental parameters of reinforced concrete within a predetermined time period; the environmental parameters are environmental data collected by a plurality of non-invasive sensors deployed on the structural surface area of the reinforced concrete, and a target environmental parameter of a target reinforced concrete; the target environmental parameter is environmental data collected by the plurality of non-invasive sensors deployed on the structural surface area of the target reinforced concrete; A central processing subsystem for constructing training data based on the environmental parameters, training a predetermined machine learning model based on the training data to train a parameter prediction model capable of learning the nonlinear mapping relationship between the time series variation of the environmental parameters of the reinforced concrete and the chloride ion transport key parameters, inputting the target environmental parameters into the parameter prediction model to obtain the time-varying target chloride ion transport key parameters of the target reinforced concrete output by the parameter prediction model, and inputting the time-varying target chloride ion transport key parameters as boundary conditions and model coefficients into a numerical solver based on a chloride ion diffusion physical model to obtain a chloride ion concentration distribution function at any depth inside the target reinforced concrete by solving the numerical solver, and predicting the residual life of the target reinforced concrete according to the chloride ion concentration distribution function.

9. An electronic device, comprising: comprising at least one controller and a memory connected in communication with the controller; the memory storing instructions executable by the at least one controller, the instructions being executed by the at least one controller to cause the at least one controller to perform a method of concrete life prediction based on surface monitoring and physical-data models as claimed in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer readable storage medium stores computer executable instructions for causing a computer to perform a method of concrete life prediction based on surface monitoring and physical-data models as claimed in any one of claims 1 to 7.

Images