Massive concrete water storage curing process multi-parameter fusion early warning system

Abstract

This invention relates to the field of construction monitoring technology, specifically disclosing a multi-parameter fusion early warning system for the water-curing process of large-volume concrete. The system includes a multi-source heterogeneous sensor network, an edge computing gateway, a cloud-based fusion analysis platform, and an early warning execution terminal. The sensor network collects multi-dimensional parameters such as temperature, humidity, and stress of the concrete and its environment. The cloud platform, based on a multi-physics coupling model and a fusion early warning algorithm, assesses the curing status and generates tiered early warning commands. The early warning terminal drives on-site equipment to perform differentiated interventions. This invention achieves comprehensive perception, accurate assessment, and proactive control of the multi-physics state during the curing process, improving the reliability of early warnings and the quality of curing.

Description

Technical Field

[0001] This invention belongs to the field of construction monitoring technology, specifically involving a multi-parameter fusion early warning system for the water storage and curing process of large-volume concrete. Background Technology

[0002] In the field of large-scale civil engineering and infrastructure construction, the construction quality and long-term durability of concrete structures are the core guarantees for their safe service. The final performance of concrete is not formed immediately after pouring, but depends on a key post-treatment stage—curing. This process directly affects the strength development, volume stability, and crack resistance of concrete by controlling the environmental conditions of hydration reaction.

[0003] Water curing of large-volume concrete components, as an efficient and widely used method for temperature and humidity control, aims to regulate the temperature gradient inside the concrete caused by the heat of hydration through the heat capacity and evaporation of external water bodies, and to keep the surface moist to prevent plastic shrinkage cracks. The basic principle of this technology is to create a relatively stable microenvironment that is conducive to the continuous hydration reaction.

[0004] Existing technologies typically rely on independent monitoring and threshold alarms for single or a few parameters, such as water level in curing tanks, water temperature, or concrete surface temperature. However, the performance evolution of concrete is a complex result of highly coupled and dynamically interacting multi-physics fields, including its internal micro-hydration process, macro-temperature and humidity fields, and the resulting self-generated and constrained stress fields. Relying solely on changes in a single parameter for early warning is prone to misjudgment due to neglecting the synergistic or antagonistic effects between parameters. For example, suitable water temperature but insufficient water level may lead to localized dehydration and cracking, while sufficient water level but abnormally high water temperature may exacerbate internal thermal stress. This fragmented monitoring method cannot comprehensively and accurately reflect whether the curing status is truly within the optimal or safe range, making it difficult to provide timely and reliable early warnings for potential quality risks such as early cracking and insufficient strength development. This has become a critical technical challenge that urgently needs to be addressed to improve the intelligent level of quality control in large-volume concrete construction. Summary of the Invention

[0005] The purpose of this invention is to provide a multi-parameter fusion early warning system for the water storage curing process of large-volume concrete, so as to solve the technical contradiction in the prior art that relies on independent monitoring of a single or a few parameters, which cannot comprehensively and accurately reflect the multi-physical field coupling state during the water storage curing process of large-volume concrete, resulting in low early warning reliability and difficulty in effectively preventing quality risks.

[0006] To achieve the above objectives, the technical solution adopted in this invention is a multi-parameter fusion early warning system for the water-curing process of large-volume concrete. This system includes a multi-source heterogeneous sensor network, an edge computing gateway, a cloud-based fusion analysis platform, and an early warning execution terminal. The multi-source heterogeneous sensor network is distributed across the surface, interior, and water body of the concrete component for real-time acquisition of multi-dimensional physical parameters characterizing the concrete hydration process, temperature field, humidity field, and stress field state. The edge computing gateway is connected to the multi-source heterogeneous sensor network via wired or wireless means to preprocess, locally cache, and perform preliminary fusion calculations on the acquired raw sensor data. The cloud-based fusion analysis platform communicates with the edge computing gateway via a wide area network to receive the preprocessed multi-dimensional data, perform in-depth multi-parameter fusion analysis and state assessment, and generate tiered early warning commands. The early warning execution terminal receives and executes the tiered early warning commands from the cloud-based fusion analysis platform, driving on-site audible and visual alarm devices, automatic water supply valves, circulating water pumps, or shading and insulation facilities to intervene.

[0007] Furthermore, the multi-source heterogeneous sensor network specifically includes an embedded strain-temperature composite sensor, a surface temperature and humidity sensor array, a multi-parameter water monitoring probe, and an environmental meteorological monitoring station. The embedded strain-temperature composite sensor is pre-embedded in a three-dimensional grid pattern within key sections of the concrete component to simultaneously measure temperature and strain at specific points inside the concrete. The surface temperature and humidity sensor array is attached to each exposed surface of the concrete component in an equally spaced matrix to measure surface temperature and relative humidity distribution. The multi-parameter water monitoring probe is submerged in a water storage and curing tank to continuously monitor water temperature, water level, pH value, and conductivity. The environmental meteorological monitoring station is set up near the curing area to collect data on ambient temperature, relative humidity, wind speed, and solar radiation intensity.

[0008] Furthermore, the data preprocessing process of the edge computing gateway includes the following steps: First, the raw data from each sensor is timestamped and parsed. Second, the parsed data undergoes outlier removal and missing value imputation. Outlier removal uses a dynamic thresholding method based on the Laida criterion, and missing value imputation uses a linear interpolation method based on time series. Finally, the processed data is standardized and normalized, mapping parameters of different dimensions and magnitudes to a numerical range of 0 to 1, forming a standardized multi-dimensional parameter vector.

[0009] Furthermore, the cloud-based fusion analysis platform includes a data warehouse module, a multiphysics coupling model module, a health assessment engine, and an early warning decision module. The data warehouse module stores historical maintenance data, current real-time data, and inherent property data such as concrete material mix proportions and design strength. The multiphysics coupling model module is a digital twin model built based on the finite element method and the principles of thermodynamics and hydration kinetics. It dynamically calibrates model parameters by receiving real-time data and simultaneously calculates the spatiotemporal evolution distribution of the three-dimensional temperature field, humidity field, and stress field inside the concrete. The health assessment engine calculates multiple key performance indicators based on the output of the multiphysics coupling model module.

[0010] Furthermore, the key performance indicators include the maximum internal temperature difference, surface evaporation rate, equivalent age, crack resistance safety factor, and hydration completion rate. The maximum internal temperature difference is obtained by extracting the temperature values ​​of all nodes in the digital twin model at the same moment and calculating the difference between the maximum and minimum values. The surface evaporation rate is calculated by combining surface temperature and humidity sensor data, environmental meteorological data, and an evaporation model based on mass transfer principles. The equivalent age is obtained by integrating the temperature history of each point inside the concrete and converting it according to maturity theory. The crack resistance safety factor is obtained by comparing the maximum principal tensile stress calculated by the digital twin model with the tensile strength of the concrete at the current age based on the equivalent age. The hydration completion rate is calculated by comparing the theoretical completion rate of the hydration reaction at the current equivalent age with the final designed hydration degree.

[0011] Furthermore, the early warning decision module runs a multi-level fusion early warning algorithm based on fuzzy reasoning and evidence theory. This algorithm first defines a membership function for each key performance indicator, mapping the measured or calculated values ​​of the indicator to a confidence assignment for three fuzzy states: safe, concerning, and risky. Second, using the Dempster-Shafer composition rule in evidence theory, it fuses the confidence assignments of the fuzzy states from different key performance indicators to obtain a joint confidence score for the overall maintenance status, classifying it as safe, concerning, or risky. Finally, two confidence thresholds are set: when the joint confidence score for the risky state exceeds the first threshold, a level-one early warning instruction is generated; when the joint confidence score for the risky state does not exceed the first threshold but the joint confidence score for the concerning state exceeds the second threshold, a level-two early warning instruction is generated; when the joint confidence score for all states is below the corresponding threshold, the state is determined to be safe, and no early warning instruction is generated.

[0012] Furthermore, the early warning execution terminal executes preset differentiated response strategies for different levels of early warning commands. Upon receiving a Level 1 early warning command, the terminal immediately triggers the highest-level audible and visual alarm and automatically activates all preset intervention measures, including opening the emergency water supply valve to maximum flow, starting all circulating water pumps for forced cooling, and deploying sunshade and insulation facilities. Upon receiving a Level 2 early warning command, the terminal triggers a medium-intensity audible and visual alarm and activates some intervention measures, such as opening the regulating valve for compensatory water supply or starting some circulating water pumps. The status and parameters of all executed actions are fed back to the data warehouse module of the cloud-based fusion analysis platform, forming a control closed loop.

[0013] Furthermore, the system also includes a maintenance effect traceability and knowledge base update module. After the concrete curing period ends, this module summarizes all monitoring data, model calculation records, early warning events, and intervention records from the entire curing process, generating a complete electronic maintenance log. Simultaneously, this module performs correlation analysis between the final strength test results, apparent quality assessment results, and multi-parameter sequences from the curing process. It utilizes machine learning algorithms to uncover potential patterns between early warning indicators, intervention measures, and the final maintenance quality, thereby optimizing and updating parameters in the multiphysics coupling model, indicator weights in the health assessment engine, and reliability thresholds and membership functions in the early warning decision module, achieving system self-learning and continuous performance improvement.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0015] 1. This invention constructs a multi-source heterogeneous sensor network that integrates embedded, surface, water-based, and environmental sensors, enabling comprehensive and three-dimensional synchronous monitoring of concrete hydration process, temperature, humidity, stress, and environmental conditions. This fundamentally overcomes the limitations of single-parameter monitoring and provides a complete data foundation for accurately assessing curing status.

[0016] 2. By introducing a digital twin model based on the finite element method and multiphysics theory, this invention can fuse discrete sensor data into a continuous spatiotemporal distribution of temperature, humidity and stress fields inside concrete, realizing a leap from point monitoring to field cognition, making it possible to identify potential internal risks such as temperature gradients, drying shrinkage and tensile stress concentration at an early stage.

[0017] 3. The multi-level fusion early warning algorithm based on fuzzy reasoning and evidence theory proposed in this invention effectively handles the uncertainty, fuzziness and conflict information among multiple parameters. It makes decisions through confidence fusion rather than simple threshold judgment, which significantly improves the accuracy and reliability of early warning and greatly reduces the false alarm rate caused by parameter noise or single parameter anomalies.

[0018] 4. This invention realizes a closed-loop linkage between early warning and execution, automatically triggering differentiated and precise intervention measures according to different risk levels, upgrading the traditional passive alarm to active control, greatly improving the control efficiency and intelligence level of the maintenance process, and effectively ensuring the construction quality and long-term durability of large-volume concrete.

[0019] 5. This invention has self-learning and optimization capabilities. Through the maintenance effect traceability and knowledge base update module, it can continuously accumulate engineering experience and optimize the core parameters and models of the system, so that the early warning performance of the system can continue to evolve with the increase of application cases, and has long-term practical value and adaptability. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the overall technical solution architecture of the present invention;

[0021] Figure 2 This is a schematic diagram of the core principle framework of the multi-level fusion early warning algorithm based on fuzzy reasoning and evidence theory in this invention;

[0022] Figure 3 This is a flowchart illustrating the data acquisition and preprocessing logic of the multi-source heterogeneous sensor network and edge computing gateway in this invention.

[0023] Figure 4 This is a schematic diagram of the interaction relationship and data flow between the multiphysics coupling model and the health assessment engine within the cloud-based fusion analysis platform of this invention;

[0024] Figure 5 This is a schematic diagram of the hierarchical early warning command closed-loop response between the early warning execution terminal and the cloud platform in this invention. Detailed Implementation

[0025] Example 1: Please refer to the appendix Figure 1 To be continued Figure 5 This invention proposes a multi-parameter fusion early warning system for the water-curing process of large-volume concrete. The system consists of four main parts: a multi-source heterogeneous sensor network, an edge computing gateway, a cloud-based fusion analysis platform, and an early warning execution terminal. These parts communicate efficiently via wired or wireless communication links, forming a closed-loop intelligent curing system integrating perception, computation, decision-making, and execution. The overall system design aims to provide comprehensive, high-precision, and real-time monitoring and assessment of the hydration reaction process, temperature field evolution, humidity migration behavior, stress development state, and external environmental disturbances involved in the water-curing stage of large-volume concrete. Based on this, it enables early risk identification, tiered early warning, and automatic intervention.

[0026] A multi-source heterogeneous sensor network forms the foundational sensing layer of this system, covering four key areas: the interior and surface of the concrete component, the water body, and the surrounding environment. Specifically, embedded strain-temperature composite sensors are pre-embedded in key sections within the concrete component using a three-dimensional grid structure. The grid spacing is set according to the component's geometry and the distribution of thermodynamically sensitive areas, with a typical value of 1 to 2 meters.

[0027] Each embedded strain-temperature composite sensor can simultaneously output the temperature and axial strain values ​​at its location. The sampling frequency is set to once every 5 minutes. The temperature measurement range is 0 to 80 degrees Celsius with an accuracy of ±0.2 degrees Celsius; the strain measurement range is -2000 to +2000 microstrains with an accuracy of ±5 microstrains. These sensors are deployed before concrete pouring and are led out to the outside of the component via dedicated wires to establish a physical connection with the edge computing gateway.

[0028] The surface temperature and humidity sensor array is attached to all exposed surfaces of the concrete structure in an equally spaced matrix, with a typical row and column spacing of 1.5 meters, ensuring effective capture of the spatial distribution characteristics of the surface temperature and humidity field. Each surface temperature and humidity sensor unit integrates dual sensing elements for temperature and relative humidity. The temperature measurement range is -10 to 60 degrees Celsius with an accuracy of ±0.3 degrees Celsius; the relative humidity measurement range is 0% to 100% with an accuracy of ±2%. The array is bonded with waterproof adhesive and supplemented with a protective cover to withstand long-term immersion in water-storage environments and mechanical impact. The data acquisition cycle is also set to 5 minutes, and the data is transmitted to the edge computing gateway via a wireless radio frequency module.

[0029] The multi-parameter water monitoring probe is submerged and installed in the central area and near the concrete surface of the water storage and maintenance tank. It continuously monitors four core water quality parameters: water temperature, water level, pH value, and conductivity. The water temperature measurement range is 0 to 50 degrees Celsius, with an accuracy of ±0.1 degrees Celsius; the water level is measured using an ultrasonic or pressure level gauge, with a range of 0 to 2 meters and an accuracy of ±1 millimeter; the pH value measurement range is 4 to 10, with an accuracy of ±0.1; and the conductivity measurement range is 0 to 2000 microsiemens per centimeter, with an accuracy of ±5 microsiemens per centimeter. All parameters are acquired at a high frequency with a 1-minute cycle to capture instantaneous changes in the water's condition.

[0030] The environmental meteorological monitoring station is erected in an open area no more than 10 meters away from concrete structures to avoid obstruction. It is used to collect ambient temperature, relative humidity, wind speed, and solar radiation intensity. The performance specifications of the ambient temperature and humidity sensors are consistent with those of the surface temperature and humidity sensors; the wind speed sensor has a range of 0 to 30 meters per second and an accuracy of ±0.3 meters per second; the solar radiation intensity sensor has a range of 0 to 1200 watts per square meter and an accuracy of ±5 watts per square meter. The meteorological data acquisition cycle is 2 minutes, and the data is uploaded to the edge computing gateway via LoRa or NB-IoT wireless protocols.

[0031] The data flow and preprocessing logic of the multi-source heterogeneous sensor network composed of all the above sensing units are shown in the attached figure. Figure 3 As shown, the edge computing gateway, acting as the local data processing hub, first performs timestamp alignment on the received raw data packets. Since the sampling periods of each sensor differ, the gateway incorporates a high-precision real-time clock module, using Coordinated Universal Time (UTC) as the reference, to assign a unified time stamp to all data, ensuring time consistency for subsequent fusion analysis. Subsequently, the gateway performs data packet parsing, extracts the payload fields of each sensor, and verifies data integrity and checksums.

[0032] During the data cleaning phase, the edge computing gateway employs a dynamic thresholding method based on the Laida criterion to remove outliers. This method first calculates the mean of a parameter within the current sliding time window (e.g., the last 10 sampling points). with standard deviation If the current sampled value x satisfies If a value is missing at time t, it is considered an outlier and is removed. For missing values ​​caused by communication interruption or sensor failure, the gateway uses a time-series-based linear interpolation method to fill in the missing values. Specifically, if data is missing at time t, and... and If the time-based data is valid, then the missing values... If multiple consecutive points are missing, a forward or backward filling strategy is used, but the data is marked as low-confidence data.

[0033] After cleaning, the data enters the standardization and normalization stage. Due to the significant differences in dimensions and magnitudes of the parameters (e.g., temperature is in the tens of degrees Celsius, strain is on the order of 1 / 1000, and pH is dimensionless), direct fusion would lead to a numerical dominance effect. Therefore, the gateway independently applies the minimum-maximum normalization formula to each type of parameter:

[0034]

[0035] in, and This refers to the theoretical extreme value or engineering safety boundary value of this parameter obtained statistically from historical engineering databases. For example, the internal temperature of concrete... Take 5 degrees Celsius. Take 75 degrees Celsius; surface evaporation rate Take 0 grams per square meter per hour. The value is 10 grams per square meter per hour. All normalized parameters are encapsulated into a standardized multi-dimensional parameter vector, the dimension of which is equal to the total number of effective sensing channels. This vector is uploaded to the cloud-based fusion analysis platform via 4G / 5G or fiber optic links.

[0036] The cloud-based fusion analytics platform is the core decision engine of this system, and its internal structure is shown in the attached figure. Figure 4 As shown, the system comprises four main functional units: a data warehouse module, a multiphysics coupling model module, a health assessment engine, and an early warning decision module. The data warehouse module employs a distributed time-series database architecture to store three types of data: first, real-time standardized parameter vector streams from edge computing gateways; second, similar project data accumulated from historical maintenance projects, including complete monitoring sequences under different mix proportions and environmental conditions; and third, inherent attribute data of the current project, such as concrete design strength grade C40, water-cement ratio 0.38, cement type P·O 42.5, aggregate type, and admixture type. All data is indexed by timestamp and supports millisecond-level query response.

[0037] The multiphysics coupling model module constructs a high-fidelity digital twin, based on a three-dimensional finite element mesh model. The mesh density is adaptively refined according to the expected geometric complexity and thermodynamic gradient of the components, with a typical element size of 0.5 meters. This model integrates the heat conduction equation, the moisture diffusion equation, and the constitutive relations of elasticity. Its governing equations are as follows:

[0038]

[0039] in, For concrete density, For specific heat capacity, For temperature, For time, Thermal conductivity, This represents the hydration heat release rate. The hydration kinetics model, derived from the Arrhenius type, allows for dynamic inversion calibration of parameters (such as activation energy and maximum heat release rate) based on real-time measured internal temperature history. The coupling between the humidity field and the stress field is achieved by considering the superposition of contractile strain caused by humidity and thermal strain caused by temperature, with the total strain... ,in, For mechanical strain, For thermal strain, For contraction strain, the model receives a normalized parameter vector from the edge gateway every 10 minutes, and uses this as the boundary and initial conditions to update the finite element solver, driving it to recalculate the spatiotemporal evolution of field variables over the next 24 hours.

[0040] The health assessment engine calculates five key performance indicators in real time based on the output of a multiphysics coupling model. The maximum internal temperature difference indicator is determined by iterating through the temperature values ​​of all nodes in the model at the same time point to find the global maximum value. and minimum value The difference This is the indicator. According to the specifications, when... Above 25 degrees Celsius, there is a significant risk of cracking. The surface evaporation rate was calculated using a mass transfer model.

[0041]

[0042] in, This is the saturated water vapor pressure on the concrete surface (obtained from a table based on the surface temperature). This is the ambient air water vapor pressure (calculated from ambient temperature and humidity). For wind speed, This is the wind speed correction function. This is an empirical coefficient. This index reflects the rate of moisture loss; excessively high values ​​will lead to plastic shrinkage cracks. The equivalent age index is determined by the temperature history at various points within the concrete. Integrate points:

[0043]

[0044] in, For activation energy, The gas constant is... The reference temperature is typically 20 degrees Celsius. The equivalent age is used to equate strength development under varying temperature conditions to the age under standard curing conditions. The crack resistance safety factor is defined as the current tensile strength of the concrete. The maximum principal tensile stress calculated by the model The ratio, that is .when At that time, it was determined to be a high-risk state. The hydration completion rate was then compared with the theoretical hydration rate at the current equivalent age. Compared with the final hydration degree of the design The ratio (usually taken as 0.95) is obtained, that is This is used to assess whether the hydration process is progressing as expected.

[0045] The early warning decision module runs a multi-level fusion early warning algorithm based on fuzzy reasoning and evidence theory. Its core principle is shown in the attached figure. Figure 2As shown, the algorithm first defines a triangular or trapezoidal membership function for each key performance indicator, mapping the indicator value to a confidence assignment of three fuzzy states: safe, concerned, and risk. For example, for the maximum internal temperature difference... ,like If the temperature is in Celsius, then the safety reliability is 1; if Then, the safety and concern confidence levels transition linearly; if If the probability is 1, then the risk confidence level is 1. Similarly, membership functions are constructed for the other four indicators.

[0046] Subsequently, using the composition rule in Dempster-Shafer evidence theory, the reliability assignments of the five indicators were orthogonally fused. Let... For the first Each indicator relates to the state. The basic reliability assignment for ∈{safety, concern, risk} is then the joint reliability. Iterative synthesis using the following formula:

[0047]

[0048] This process sequentially integrates all indicators to ultimately obtain the joint reliability of the overall maintenance status belonging to each fuzzy state. The system presets two reliability thresholds: a first-level early warning threshold and a second-level early warning threshold. Level II warning threshold .

[0049] When the joint confidence of risk status (Risk) When, a Level 1 warning instruction is generated; when (Risk)≤ but (Attention)> If the condition is met, a Level 2 warning command will be generated; otherwise, the condition will be deemed safe.

[0050] The early warning execution terminal receives tiered early warning instructions from the cloud platform and executes differentiated response strategies. Its closed-loop response mechanism is shown in the attached figure. Figure 5 As shown, upon receiving a Level 1 warning command, the terminal immediately activates the on-site audible and visual alarm device to emit a high-decibel alarm and flashing red lights. Simultaneously, it sends control signals to the actuators: the emergency water supply valve opens to 100%, all circulating water pumps start and operate at maximum speed, and the shading and insulation facilities automatically deploy to cover the entire concrete surface. The operational status of all actuators (such as valve opening, pump current, and facility deployment angle) is transmitted in real-time to the cloud data warehouse via feedback sensors for verification of the intervention effect.

[0051] Upon receiving a Level 2 warning command, the terminal triggers a moderate-intensity audible and visual alarm (yellow light, intermittent buzzer) and initiates some intervention measures: adjusting the water supply valve to 30% to 50% opening for compensatory water replenishment, or simply activating the circulating water pump located above the high-temperature area for localized heat dissipation. These measures aim to provide gentle regulation and avoid excessive intervention that could disrupt the normal hydration process.

[0052] In addition, the system includes a maintenance effect traceability and knowledge base update module. After the concrete curing period (usually 28 days) ends, this module automatically summarizes the data from the entire process, including all original sensor records, model calculation logs, early warning event timestamps, and details of intervention actions, generating a structured electronic maintenance log that conforms to engineering file management standards. Simultaneously, the module acquires the 28-day compressive strength test results (e.g., 48.5 MPa) and on-site appearance quality scores (e.g., no cracks, no spalling, score of 95 points) provided by the laboratory, and correlates them with the multi-parameter sequence during the curing process. Machine learning algorithms such as random forests or gradient boosting trees are used to explore the nonlinear relationships between features such as the duration of the maximum internal temperature difference, the peak surface evaporation rate, and the number of first-level early warning triggers, and the final strength and the number of cracks. Based on this, the system automatically adjusts the hydration heat parameters in the multiphysics coupling model, the weight coefficients of each indicator in the health assessment engine (e.g., increasing the weight of the crack resistance safety factor from 0.25 to 0.30), and the inflection point and confidence threshold of the membership function in the early warning decision module (e.g., increasing the weight of the crack resistance safety factor from 0.25 to 0.30). (From 0.7 to 0.68), thereby achieving adaptive optimization of the model and strategy, ensuring that the system has higher early warning accuracy and intervention effectiveness in subsequent engineering.

[0053] Example 2: In another implementation scenario, this system is deployed for the water storage and maintenance of large-volume concrete foundations in port terminals within a marine environment. Due to the high concentration of salt spray, large diurnal temperature range, and strong sea winds, the system's multi-source heterogeneous sensor network configuration has been specifically adjusted. The embedded strain-temperature composite sensor housing is made of 316L stainless steel, possessing IP68 protection and resistance to chloride ion corrosion. The surface temperature and humidity sensor array is coated with a hydrophobic nano-coating to prevent salt crystallization from clogging the humidity sensing elements. The multi-parameter water monitoring probe has been enhanced with dissolved oxygen and turbidity monitoring functions to assess the impact of seawater infiltration on the maintenance water. The environmental meteorological monitoring station is equipped with a salt spray deposition rate sensor with a range of 0 to 5 milligrams per square centimeter per day.

[0054] The data preprocessing algorithm for edge computing gateways has also been enhanced accordingly. During the outlier removal stage, a sliding window midpoint filter is introduced as an auxiliary criterion to address instantaneous meteorological data jumps caused by sudden changes in sea breezes. Missing value imputation uses cubic spline interpolation instead of linear interpolation to better fit the slow water level fluctuations caused by tidal cycles. During the standardization and normalization process, the pH value of the water body... and The values ​​were adjusted to 6.5 and 8.5 to adapt to the slightly alkaline background of seawater.

[0055] The multiphysics coupling model of the cloud-based fusion analysis platform has added a chloride ion diffusion sub-model, whose governing equations are coupled to the humidity field equations for assessing the risk of rebar corrosion. The health assessment engine has added two new indicators: chloride ion penetration depth and rebar cover saturation. The early warning decision module has expanded to include a fourth fuzzy state of corrosion risk and redefined the membership function and confidence synthesis rules. When the joint confidence score of corrosion risk exceeds 0.6, an additional anti-corrosion intervention command is triggered, such as activating a freshwater flushing system to dilute surface salts.

[0056] The intervention measures implemented by the early warning execution terminal also include corrosion prevention options. In addition to conventional measures, a Level 1 early warning system automatically activates a freshwater spray system, rinsing the concrete surface at a flow rate of 0.5 liters per square meter per minute for 30 minutes. The status feedback from all newly added sensors and actuators is incorporated into the control loop to ensure controllable maintenance quality under the unique marine conditions. During retrospective analysis, the knowledge base update module focuses on linking salt spray deposition rate, chloride ion penetration depth, and the steel reinforcement potential test results after 28 days, continuously optimizing the corrosion prevention early warning strategy.

Claims

1. A multi-parameter fusion early warning system for the water-curing process of large-volume concrete, characterized in that, include: A multi-source heterogeneous sensor network is distributed and deployed on the surface, inside and in the water body for water storage and curing of concrete components to collect multi-dimensional physical parameters in real time that characterize the hydration process, temperature field, humidity field and stress field of concrete. An edge computing gateway is connected to the multi-source heterogeneous sensor network via wired or wireless means to preprocess, cache locally, and perform preliminary fusion calculations on the collected raw sensor data to form a standardized multi-dimensional parameter vector. The cloud-based fusion analysis platform communicates with the edge computing gateway via a wide area network to receive the standardized multi-dimensional parameter vector, perform deep multi-parameter fusion analysis and status assessment, and generate hierarchical early warning instructions. The early warning execution terminal receives and executes hierarchical early warning instructions from the cloud-based fusion analysis platform, driving on-site audible and visual alarm devices, automatic water supply valves, circulating water pumps, or sunshade and heat preservation facilities to intervene. The cloud-based fusion analysis platform includes a data warehouse module, a multiphysics coupling model module, a health assessment engine, and an early warning decision module. The data warehouse module is used to store historical maintenance data, current real-time data, and inherent property data of concrete materials. The multiphysics coupling model module is a digital twin model built based on the finite element method and the principles of thermodynamics and hydration kinetics. It dynamically calibrates the model parameters by receiving real-time data and simultaneously calculates the spatiotemporal evolution distribution of the three-dimensional temperature field, humidity field and stress field inside the concrete. The health assessment engine is used to calculate multiple key performance indicators based on the output of the multiphysics coupling model module; The early warning decision module is used to run a multi-level fusion early warning algorithm based on fuzzy reasoning and evidence theory. It defines a membership function for each key performance indicator to map to the confidence assignment of fuzzy states, and uses the synthesis rules in evidence theory to fuse the confidence assignments of fuzzy states from different key performance indicators to obtain the joint confidence for the overall maintenance status. Then, it generates a graded early warning instruction according to the preset confidence threshold. The multi-source heterogeneous sensor network includes an embedded strain-temperature composite sensor, a surface temperature and humidity sensor array, a water body multi-parameter monitoring probe, and an environmental meteorological monitoring station. The embedded strain-temperature composite sensor is pre-embedded in a three-dimensional grid pattern inside key sections of the concrete component to simultaneously measure the temperature and strain at specific points inside the concrete. The surface temperature and humidity sensor array is attached to each exposed surface of the concrete component in an equally spaced matrix to measure the surface temperature and relative humidity distribution. The water body multi-parameter monitoring probe is immersed in a water storage and curing tank to continuously monitor water temperature, water level, pH value, and conductivity. The environmental meteorological monitoring station is set up near the curing area to collect data on ambient temperature, relative humidity, wind speed, and solar radiation intensity. The key performance indicators include the maximum internal temperature difference, surface evaporation rate, equivalent age, crack resistance safety factor, and hydration completion rate. The maximum internal temperature difference is obtained by extracting the temperature values ​​of all nodes in the digital twin model at the same moment and calculating the difference between the maximum and minimum values. The surface evaporation rate is calculated by combining surface temperature and humidity sensor data, environmental meteorological data, and an evaporation model based on mass transfer principles. The equivalent age is obtained by integrating the temperature history of each point inside the concrete and converting it according to maturity theory. The crack resistance safety factor is obtained by comparing the maximum principal tensile stress calculated by the digital twin model with the tensile strength of the concrete at the current age based on the equivalent age. The hydration completion rate is obtained by comparing the theoretical completion rate of the hydration reaction at the current equivalent age with the final designed hydration degree. The early warning execution terminal executes preset differentiated response strategies for different levels of early warning commands; when a first-level early warning command is received, the early warning execution terminal immediately triggers the highest level of audible and visual alarm, and automatically starts all preset intervention measures, including opening the emergency water supply valve to the maximum flow rate, starting all circulating water pumps for forced heat dissipation, and deploying sunshade and heat preservation facilities. When a Level 2 warning instruction is received, the warning execution terminal triggers a medium-intensity audible and visual alarm and initiates some intervention measures, including opening the regulating valve for compensatory water replenishment or starting some circulating water pumps. The system also includes a maintenance effect traceability and knowledge base update module. After the concrete curing period ends, the maintenance effect traceability and knowledge base update module summarizes all monitoring data, model calculation records, early warning events and intervention records of the entire curing process to generate a complete electronic curing log. At the same time, the maintenance effect traceability and knowledge base update module performs correlation analysis on the final strength test results and apparent quality assessment results of this curing and the multi-parameter sequence during the curing process. It uses machine learning algorithms to mine the potential patterns between early warning indicators, intervention measures and final curing quality, and uses this to optimize and update the parameters in the multiphysics coupling model module, the indicator weights in the health assessment engine and the reliability threshold and membership function in the early warning decision module.

2. The multi-parameter fusion early warning system for the water-curing process of large-volume concrete according to claim 1, characterized in that, The data preprocessing process of the edge computing gateway includes the following steps: aligning the timestamps and parsing the raw data from each sensor; removing outliers and imputing missing values ​​in the parsed data, wherein outlier removal adopts a dynamic thresholding method based on the Laida criterion, and missing value imputation adopts a linear interpolation method based on time series; and standardizing and normalizing the processed data, mapping parameters of different dimensions and magnitudes to a numerical range of 0 to 1, forming the standardized multi-dimensional parameter vector.

3. The multi-parameter fusion early warning system for the water-curing process of large-volume concrete according to claim 2, characterized in that, The multi-level fusion early warning algorithm running in the early warning decision module sets two confidence thresholds. When the joint confidence of the risk status exceeds the first threshold, a first-level early warning instruction is generated. When the joint confidence of the risk status does not exceed the first threshold but the joint confidence of the concern status exceeds the second threshold, a second-level early warning instruction is generated. When the joint confidence of all statuses is lower than the corresponding threshold, it is determined to be a safe status and no early warning instruction is generated.

4. The multi-parameter fusion early warning system for the water-curing process of large-volume concrete according to claim 3, characterized in that, In the multiphysics coupling model module, the hydration heat release rate is given by the hydration kinetic model, and its parameters are dynamically inverted and calibrated based on the real-time measured internal temperature history; the coupling between the humidity field and the stress field is achieved by considering the superposition of the contraction strain caused by humidity and the thermal strain caused by temperature.

5. The multi-parameter fusion early warning system for the water-curing process of large-volume concrete according to claim 4, characterized in that, The calculation process for the equivalent age is as follows: the temperature history at each point inside the concrete is integrated, where the integration kernel function is an exponential function with activation energy and gas constant as parameters, and the reference temperature is 20 degrees Celsius.

Images