Intelligent temperature control maintenance system for mass concrete of bridge pier

Abstract

The application discloses a kind of bridge pier body mass concrete intelligent temperature control maintenance system, belong to mass concrete temperature control maintenance technical field.Solve the problem that concrete hydration heat parameter time-varying leads to prediction misalignment, temperature control lagging behind is prone to cracking, the system includes temperature sensing array, pre-buried cooling water pipe network, circulating pump and intelligent controller.Intelligent controller is identified by constructing heat balance prediction model with equivalent specific heat, equivalent thermal conductivity and hydration heat release rate as time-varying parameter;With the measured and predicted temperature residual, with recursive least square method or Kalman filtering on-line rolling identification parameter, so that model approximates actual;Based on updated model, calculate future temperature prediction bias as feedforward, superimposed measured deviation feedback generates composite control signal, dynamically adjusts circulating pump speed.Thereby realize the adaptive precision control of mass concrete temperature, mainly used for intelligent temperature control maintenance of bridge pier body and other mass concrete, prevent temperature crack.

Description

Technical Field

[0001] This invention belongs to the field of temperature control curing technology for large-volume concrete, specifically relating to an intelligent temperature control curing system for large-volume concrete in bridge piers. Background Technology

[0002] In the construction and curing of large-volume concrete, the large amount of heat released by the cement hydration reaction causes the internal temperature of the structure to rise rapidly, while the surface dissipates heat more quickly, resulting in a significant temperature difference between the inside and outside. When the tensile stress caused by this temperature difference exceeds the early tensile strength of the concrete, temperature cracks occur, directly affecting the load-bearing capacity and durability of the bridge piers. To control this process, cooling water pipes are often pre-embedded in the concrete and circulated to lower the temperature. However, existing temperature control methods have a series of technical shortcomings when dealing with the dynamic changes in hydration heat and the thermal inertia brought about by the structural volume, which restricts the timeliness and accuracy of temperature control.

[0003] First, the key to determining the cooling effect lies in the accurate prediction of the future trend of the internal temperature field of the concrete. However, the thermal parameters upon which this prediction relies are not constant. The equivalent specific heat capacity, equivalent thermal conductivity, and hydration heat release rate will continuously change over time with the hydration process, local temperature history, and even fluctuations in the actual mix proportions. Parameters measured by standard methods under laboratory conditions are difficult to truly reflect the evolution of the cement hydration environment after pouring, resulting in a non-negligible deviation between the temperature field simulation results based on fixed parameters and the actual monitoring data. If such inaccurate predictions are used directly to guide water supply decisions, it is very easy to lead to a mismatch between cooling intensity and actual needs. This could result in insufficient cooling leading to excessive temperatures, or excessive cooling increasing the internal temperature gradient of the structure, which in turn increases the risk of cracking.

[0004] Secondly, even with reliable real-time temperature monitoring data, simple feedback control methods are limited by the thermal inertia of large-volume concrete. The larger the concrete volume, the more significant the temperature response delay from the point of application of cooling energy to areas far from water pipes. Common feedback control loops adjust cooling water flow or temperature only after detecting an anomaly. By this time, the internal temperature field has often deviated significantly from the target trajectory, requiring a longer adjustment period to slowly bring it back, failing to effectively suppress the rapid temperature rise during the peak of hydration heat. This delayed response makes temperature control during curing a passive, follow-the-leader approach, unable to proactively predict and intervene, making it difficult to consistently keep the maximum temperature and maximum temperature difference within a safe range.

[0005] Attempts to overcome the problem of pure feedback lag by establishing mathematical models for online prediction and applying feedforward compensation theoretically exist, but in reality, this approach faces significant challenges in parameter identification. The time-varying parameters in the hydration process of large-volume concrete are coupled and cannot be directly measured, requiring indirect inference based on time-series data from a few internal temperature measurement points. Most commonly used online parameter identification methods are based on certain linear or quasi-linear assumptions, but the nonlinear characteristics of concrete temperature field evolution and the time-varying nature of parameters make identification algorithms prone to divergence or slow convergence during abrupt state changes. The limitations on the number and location of measurement points, coupled with measurement noise, further exacerbate the difficulty of extracting effective signals. To simultaneously achieve stable tracking of slowly varying parameters such as equivalent thermal conductivity and rapidly changing state variables such as heat release rate within a single algorithmic framework, while maintaining physical consistency of the identification results throughout the curing process, current technologies struggle to meet the real-time requirements of engineering projects both quickly and accurately.

[0006] Furthermore, even if relatively reliable current-moment parameters are obtained, there are still technical obstacles to integrating the predicted future temperature trend into a reliable feedforward control variable and coordinating it with feedback correction. The feedforward channel is highly sensitive to model accuracy; once the model has deviations, feedforward compensation can easily introduce additional disturbances. Meanwhile, a simple feedback loop, acting alone when the feedforward fails, reverts to the old problem of response lag. Finding a balance between these two aspects and ensuring smooth switching of control strategies under different hydration stages and environmental conditions has always lacked a simple and effective engineering solution. This insufficient coordination between prediction and decision-making often degenerates on-site temperature control into intermittent adjustments relying on human experience, resulting in significant dispersion in maintenance quality.

[0007] In summary, existing temperature-controlled curing technologies for large-volume concrete struggle to provide reliable closed-loop solutions for addressing the time-varying thermophysical parameters, hysteretic temperature response, and the challenges of predictive-control coordination. Inaccurate parameters lead to distorted predictions, which in turn render feedforward unusable. The lack of feedforward results in lagging control, ultimately causing visible or microscopic temperature-induced cracks to appear in some bridge pier structures during curing. Overcoming these deficiencies requires a comprehensive improvement across sensing, modeling, identification, and control. However, engineering solutions integrating real-time online identification, adaptive prediction, and composite control are rarely reported, making this a long-standing technical challenge that remains to be effectively addressed in this field. Summary of the Invention

[0008] One object of the embodiments of the present invention is to solve at least the above-mentioned problems and / or defects, and to provide at least the advantages described below.

[0009] Another objective of this invention is to provide an intelligent temperature control curing system for large-volume concrete in bridge piers.

[0010] The problem is that during the curing of large-volume concrete, the hydration heat parameters of cement change continuously over time. Simulation predictions using fixed thermal parameters deviate significantly from the actual temperature field. Furthermore, relying solely on feedback adjustment results in a severe lag in the water cooling response due to the enormous thermal inertia of concrete. This makes it difficult to suppress the temperature rise in time during the peak temperature rise stage and to smoothly control the temperature difference during the cooling stage. Ultimately, this leads to the temperature difference between the inside and outside of the structure frequently exceeding the limit, resulting in temperature cracks.

[0011] This paper addresses the problem that when identifying parameters of a concrete thermal balance model online, the equivalent specific heat capacity and equivalent thermal conductivity change slowly, while the hydration heat release rate changes drastically. These two parameters are coupled and cannot be directly measured. Most existing identification methods struggle to simultaneously achieve both computational stability and tracking speed, resulting in the identified parameters failing to consistently approximate the actual physical process.

[0012] The problem is that when generating the feedforward control signal, only the current measured temperature deviation is used for simple correction, which lacks comprehensive consideration of factors such as future changes in cooling water flow rate, ambient temperature fluctuations, and time-varying thermal parameters. As a result, it is impossible to effectively predict the temperature trend after the control cycle, and the feedforward compensation component lacks predictability and is difficult to offset the upcoming temperature deviation in advance.

[0013] To achieve the above-mentioned objectives, the present invention employs the following technical solution: A smart temperature-controlled curing system for large-volume concrete in bridge piers includes: Temperature sensing array, used to collect spatiotemporal temperature field data of the inside and surface of concrete in real time; Cooling water piping is embedded in the concrete. A circulating pump, connected between the water supply unit and the cooling water network, is used to drive the cooling water circulation; An intelligent controller is electrically connected to both the temperature sensing array and the circulating pump. The intelligent controller includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to perform the following steps: A prediction model based on concrete thermal balance is constructed, wherein the prediction model includes equivalent specific heat capacity, equivalent thermal conductivity and hydration heat release rate as time-varying parameters to be identified. An online adaptive parameter identification program is executed, using the residual between the measured temperature value collected by the temperature sensing array at the current moment and the temperature prediction value output by the prediction model at the previous moment as input. The recursive least squares method or Kalman filter algorithm is used to continuously update the time-varying parameters in the prediction model so that the temperature prediction value output by the updated prediction model at the current moment converges to the measured temperature value. A feedforward-feedback composite control signal is generated. Based on the updated prediction model, the predicted temperature value after a preset control period is calculated. The deviation between the predicted temperature value and the target temperature curve is used as the feedforward control component, and the deviation between the measured temperature value and the target temperature curve is used as the feedback control component. The composite control signal is generated by superimposing these components. Temperature control is performed, and the speed of the circulating pump is dynamically adjusted according to the composite control signal.

[0014] Preferably, in the aforementioned intelligent temperature-controlled curing system for large-volume concrete in bridge piers, when the processor executes the online adaptive parameter identification program, it is specifically used for: Based on the residuals, a Jacobian matrix is ​​constructed, and the equivalent specific heat capacity and equivalent thermal conductivity are identified online using the recursive least squares method. Alternatively, a state-space equation containing the heat release rate of hydration can be constructed, and the heat release rate of hydration can be estimated in real time using a Kalman filter.

[0015] Preferably, in the intelligent temperature control curing system for large-volume concrete in bridge piers, when the processor generates the feedforward control component, it is also used for: The current cooling water flow rate, ambient temperature, and the updated time-varying parameters are input into the prediction model to calculate the predicted sequence of the concrete internal temperature field within a preset time domain.

[0016] Preferably, in the intelligent temperature control curing system for large-volume concrete in bridge piers, when the processor performs temperature control, it is specifically used for: The composite control signal is converted into a PWM pulse width modulation signal or an analog voltage signal to control the speed of the circulating pump driven by the frequency converter.

[0017] Preferably, in the intelligent temperature control curing system for large-volume concrete of bridge piers, the processor is further configured as follows: At the moment the circulating pump is started, a thermal shock suppression period is set. During the thermal shock suppression period, the composite control signal is forced to output a preset low speed drive signal, and the shutdown logic based on the temperature difference threshold is blocked. The duration of the thermal shock suppression period is calculated by the processor based on the difference between the water temperature in the cooling water network and the current average temperature of the concrete, as well as the total volume of the cooling water network.

[0018] Preferably, in the intelligent temperature-controlled curing system for large-volume concrete of bridge piers, the system further includes a heating unit connected to the cooling water network; the processor is further configured to: Calculate the real-time temperature difference between the internal temperature and the surface temperature of the concrete; When the real-time temperature difference value is less than or equal to the preset temperature difference threshold, but the internal temperature of the concrete exceeds the preset peak temperature, the circulating pump is controlled to maintain operation at a constant low speed. When the internal temperature of the concrete drops below the preset curing temperature, the circulating pump is turned off and the heating unit is started to heat the cooling water, so that the internal temperature of the concrete drops to the ambient temperature at a preset cooling rate.

[0019] Preferably, in the intelligent temperature control curing system for large-volume concrete of bridge piers, the temperature sensing array is a distributed fiber optic temperature sensing array, which is buried at multiple points at different depths and on the surface of the concrete to obtain the spatiotemporal temperature field data.

[0020] Preferably, in the intelligent temperature control curing system for large-volume concrete of bridge piers, the temperature sensing array is a distributed fiber grating array, and the intelligent controller is also used to demodulate the temperature and strain data of each measuring point according to wavelength division multiplexing technology, and use the temperature data as the spatial boundary constraint condition of the prediction model.

[0021] Preferably, in the intelligent temperature control curing system for large-volume concrete of bridge piers, the cooling water pipe network is a multi-group serpentine pipe cluster arranged in layers, and the inlet of each group of pipe clusters is equipped with an electronic flow regulating valve controlled by the intelligent controller. The processor is also used to generate independent opening signals for each of the electronic flow control valves based on the spatial temperature field gradient output by the prediction model, so as to achieve differentiated cooling control in different regions.

[0022] Preferably, in the intelligent temperature-controlled curing system for large-volume concrete in bridge piers, when the processor executes the online adaptive parameter identification program, it specifically employs a recursive least squares method with a variable forgetting factor, and adjusts the magnitude of the forgetting factor in real time according to the temperature change rate collected by the temperature sensor array. When the temperature change rate exceeds a preset threshold, the forgetting factor is reduced to enhance the tracking sensitivity of the identification algorithm to the current data. When the temperature change rate is lower than the preset threshold, the forgetting factor is increased to suppress fluctuations in the identification results caused by measurement noise, thereby ensuring that the time-varying parameters of the prediction model can stably converge to the true physical characteristics throughout the entire curing cycle.

[0023] To facilitate accurate understanding of the present invention by those skilled in the art, the equivalent specific heat capacity and equivalent thermal conductivity are defined as follows: Equivalent specific heat capacity c eq ( tSpecific heat capacity (SHC) refers to the calculated specific heat capacity within a finite-volume discrete unit of large-volume concrete, taking into account the exothermic reaction of cement hydration, the difference in heat capacity between aggregates and cementitious materials, and the influence of microcracks or pores on heat storage capacity. Its dimensions are the same as conventional specific heat capacity, expressed as J / (kg·℃). During curing, as the degree of hydration increases, the cementitious material transforms from a liquid paste to solid hydration products. c eq ( t The trend is non-linear, and this invention tracks this change in real time by online identification.

[0024] Equivalent thermal conductivity k eq ( t Thermal conductivity (W / (m·℃)) refers to the calculated thermal conductivity in heterogeneous concrete media, considering the combined effects of aggregate distribution, interfacial transition zone, early microcracks, and moisture migration on heat conduction. Its dimensions are the same as conventional thermal conductivity. In the early stages of hydration, when the free water content is high, the thermal conductivity is relatively high; as hydration progresses and drying shrinkage occurs, the thermal conductivity decreases. k eq ( t The thermal conductivity of the model may decrease. This invention does not rely on empirical formulas, but rather identifies the parameter in real time using a recursive least squares method driven by temperature residuals, so that the thermal conductivity of the model matches the actual temperature response.

[0025] Both of the above parameters are time-varying state variables and cannot be directly measured, but they can be estimated using the measured data of the temperature sensing array through the online adaptive parameter identification program described in this invention.

[0026] Compared with the prior art, the advantages and beneficial technical effects of the present invention are: This invention combines a temperature sensing array, cooling water pipe network, and circulating pump with an intelligent controller that incorporates an online adaptive parameter identification and feedforward-feedback composite control algorithm. This enables the system to perceive the spatiotemporal temperature field inside the concrete in real time, and dynamically update the prediction model based on the identified equivalent heat capacity, thermal conductivity, and hydration heat rate. It also calculates temperature trends in advance, superimposes real-time deviation corrections, and actively adjusts the circulating pump speed to achieve precise control of the temperature field throughout the curing process. This effectively eliminates temperature difference cracks caused by inaccurate prediction and control lag, and significantly improves the curing quality and automation level of large-volume concrete.

[0027] This invention effectively separates slowly varying parameters from rapidly varying states by employing a recursive least squares method based on residual-constructed Jacobian matrices to identify equivalent specific heat capacity and thermal conductivity, or by estimating the hydration heat release rate using a Kalman filter based on state-space equations during online adaptive parameter identification. This results in an identification process that combines stability and rapid tracking capability, ensuring that the predicted model parameters closely approximate the true values ​​at each stage of concrete hydration. This provides a reliable thermal parameter basis for subsequent temperature prediction and control decisions.

[0028] This invention incorporates the current cooling water flow rate, ambient temperature, and real-time updated time-varying parameters into the prediction model when generating the feedforward control component. This model calculates the predicted temperature field sequence within a preset time domain, thereby enabling the early detection of temperature drift trends caused by cooling intensity or environmental changes. This allows feedforward compensation to accurately offset impending deviations, overcoming the problem of thermal response hysteresis in large-volume concrete and significantly enhancing the predictability and stability of the control.

[0029] This invention controls the circulating pump driven by the frequency converter by converting the composite control signal into a PWM pulse width modulation signal or an analog voltage signal. This allows the output of the intelligent controller to flexibly adapt to different pump drive interfaces, achieving high-precision, stepless speed regulation. It ensures that the cooling water circulation flow follows the control command quickly and accurately, eliminating the disconnect between control decision and physical execution, thereby improving the real-time control quality and engineering applicability of the temperature control system.

[0030] This invention sets a thermal shock suppression period at the moment the circulating pump starts. During this period, the circulating pump is forced to run at a preset low speed and the shutdown logic based on temperature difference is blocked. This allows the low-temperature cooling water to enter the pipeline slowly and steadily, avoiding local temperature stress and impact cracks caused by sudden cooling. This ensures the integrity of the concrete structure in the initial stage of water cooling and prevents secondary damage caused by operation during sensitive periods. It also establishes a safe starting point for the subsequent normal temperature control process.

[0031] This invention adds a heating unit and configures a control logic that automatically switches between cooling and heating based on the internal and external temperature difference and the internal temperature. When the internal temperature is too high but the temperature difference is still acceptable, it maintains low-speed cooling. When the internal temperature drops below the curing temperature, it automatically stops and starts heating, cooling down slowly at a preset rate. This achieves seamless coordination of cooling and heating throughout the curing process, avoiding tensile stress caused by excessively rapid cooling and preventing the internal and external temperature difference from exceeding the standard, thus significantly reducing the risk of cracking in the later stages of curing.

[0032] This invention employs a distributed fiber optic temperature sensing array, which is embedded at multiple points within concrete at different depths and on its surface. This allows for the acquisition of high-density, continuous spatial temperature field data. Compared to traditional point sensors, this provides rich and realistic spatiotemporal temperature information as input and verification data for predictive models. It strongly supports the refined reconstruction of the temperature field and parameter identification, thereby improving the targeting and effectiveness of temperature control strategies.

[0033] This invention utilizes distributed fiber grating arrays and wavelength division multiplexing (WDM) technology to demodulate temperature and strain, directly using temperature data as spatial boundary constraints for the prediction model. This makes the model boundary conditions closer to the real heat dissipation environment, improving the accuracy of simulation calculations. At the same time, the acquired strain information can provide auxiliary judgment for crack risk assessment, realizing the integration of temperature measurement and health monitoring, expanding the functional dimensions of the system, and providing additional safety assurance for concrete curing.

[0034] This invention designs the cooling water pipe network as a multi-group serpentine pipe cluster arranged in layers, and sets an electronic flow regulating valve at each inlet that is independently regulated by an intelligent controller. Based on the spatial temperature field gradient, it generates differentiated opening signals, realizing zoned differentiated cooling along different parts of the pier. This can uniformize the temperature distribution, avoid local overcooling or hot spot residue, effectively reduce the overall temperature gradient of the structure, and improve the uniformity and safety of temperature control across the entire cross section.

[0035] This invention employs a recursive least squares method with a variable forgetting factor to dynamically adjust the forgetting factor based on the rate of temperature change collected by the temperature sensing array. When the temperature changes drastically, the forgetting factor is reduced to improve the sensitivity to new data and enhance the tracking ability. When the temperature changes gradually, the forgetting factor is increased to suppress parameter fluctuations caused by measurement noise. This ensures that the time-varying parameters of the prediction model can stably converge to the true physical characteristics throughout the entire maintenance cycle, thereby maintaining the long-term reliability of the prediction model and eliminating erroneous adjustments.

[0036] Other advantages, objectives, and features of the embodiments of the present invention will be apparent in part from the following description, and in part will be understood by those skilled in the art through study and practice of the embodiments of the present invention. Detailed Implementation

[0037] To further illustrate the technical means and effects of this invention, the following embodiments are provided for further explanation. The specific embodiments described herein are merely illustrative and not intended to limit the scope of the invention.

[0038] It should be noted that, unless otherwise specified, the experimental methods described in the following implementation plan are all conventional methods, and the reagents and materials described are all commercially available unless otherwise specified.

[0039] According to one embodiment of the present invention, a smart temperature-controlled curing system for large-volume concrete in bridge piers includes: Temperature sensing array, used to collect spatiotemporal temperature field data of the inside and surface of concrete in real time; Cooling water piping is embedded in the concrete. A circulating pump, connected between the water supply unit and the cooling water network, is used to drive the cooling water circulation; An intelligent controller is electrically connected to both the temperature sensing array and the circulating pump. The intelligent controller includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to perform the following steps: A prediction model based on concrete thermal balance is constructed, wherein the prediction model includes equivalent specific heat capacity, equivalent thermal conductivity and hydration heat release rate as time-varying parameters to be identified. An online adaptive parameter identification program is executed, using the residual between the measured temperature value collected by the temperature sensing array at the current moment and the temperature prediction value output by the prediction model at the previous moment as input. The recursive least squares method or Kalman filter algorithm is used to continuously update the time-varying parameters in the prediction model so that the temperature prediction value output by the updated prediction model at the current moment converges to the measured temperature value. A feedforward-feedback composite control signal is generated. Based on the updated prediction model, the predicted temperature value after a preset control cycle is calculated. The deviation between the predicted temperature value and the target temperature curve is used as the feedforward control component, and the deviation between the measured temperature value and the target temperature curve is used as the feedback control component. These components are then superimposed to generate the composite control signal. Let the current state be the [number]th [cycle]. k One control cycle (control cycle Δ) t c Generally, the value is taken as an integer multiple of the temperature sampling period (e.g., 30 minutes). The processor generates the composite control signal according to the following steps: Step 1: Calculate the feedback control component The current measured temperature T meas ( k ) and target temperature curve T target ( k The deviation is denoted as The proportional-integral (PI) feedback mechanism is adopted. in u FB ( k ) represents the feedback control component (the dimension corresponds to a percentage of the pump speed command). K p ,K i The pre-calibrated proportional and integral coefficients can be obtained through closed-loop tuning based on the concrete volume and cooling water network characteristics.

[0040] Step 2: Calculate the feedforward control components Updated time-varying parameters i k Current cooling water flow rate v w ( k (Measured in real time by the flow meter) and ambient temperature T amb ( k Input a discrete prediction model and recursively calculate the future. H p One control cycle (prediction time domain, e.g.) H p Temperature prediction sequence (=4~6) Select the predicted temperature at the end of the first control cycle within the prediction time domain. With target temperature The deviation is used as the basis for feedforward control: The feedforward control component is designed as follows: in K FF For the feedforward gain, it can be approximately set based on the steady-state influence coefficient of cooling water flow rate on concrete temperature, and is generally taken as 0 < K FF ≤1.

[0041] Step 3: Superimpose to generate composite control signals Composite control signal u total ( k Assembled according to the following formula: Here, sat(⋅) is a saturation function that limits the output to the adjustable range of the circulating pump speed (0% corresponds to shutdown, 100% corresponds to rated speed). This composite signal, after digital-to-analog conversion or PWM modulation, drives the frequency converter to achieve coordinated control of feedforward rapid prediction and feedback steady-state correction.

[0042] Temperature control is performed, and the speed of the circulating pump is dynamically adjusted according to the composite control signal.

[0043] In one specific implementation, the intelligent temperature control curing system for the large-volume concrete of the bridge pier consists of a temperature sensor array, a cooling water pipe network, a circulating pump, and an intelligent controller. The temperature sensor array employs multiple sets of embedded digital temperature sensors, layered vertically along the pier and deployed at different depths in each section, with additional sensors on the surface, to acquire real-time spatiotemporal temperature field data of the concrete's interior and surface. The cooling water pipe network uses seamless steel pipes bent into multi-layered serpentine pipes, pre-embedded during the rebar binding stage. The circulating pump is a variable frequency centrifugal pump, with the inlet connected to a water supply tank and the outlet connected to each layer of cooling water pipes to form a closed-loop circulation. The intelligent controller is an industrial-grade embedded controller equipped with solid-state memory and a processor, connected to the frequency converters of the temperature sensor array and the circulating pump via signal lines.

[0044] The memory contains a computer program that implements the temperature control logic. When the processor executes this program, it first constructs a prediction model based on the thermal balance of concrete. This model sets the equivalent specific heat capacity, equivalent thermal conductivity, and hydration heat release rate as time-varying parameters to be identified. Subsequently, the program performs online adaptive parameter identification, reading the measured temperature value collected by the temperature sensor array in real time, calculating the residual between the measured value and the temperature prediction value output by the prediction model at the previous moment, and using the recursive least squares method or Kalman filter algorithm to continuously update the three time-varying parameters, so that the temperature prediction value output by the updated prediction model at the current moment continuously converges to the measured temperature value. Based on this, a feedforward-feedback composite control signal is generated: using the updated prediction model, combined with the current cooling water flow rate and ambient temperature, the temperature prediction value after a preset control period is calculated, and the deviation between this temperature prediction value and the target temperature curve is used as the feedforward control component; at the same time, the deviation between the current measured temperature value and the target temperature curve is used as the feedback control component, and the two parts are superimposed to generate a composite control signal. Finally, based on the magnitude and direction of the composite control signal, the corresponding analog voltage signal is output to adjust the frequency converter, thereby dynamically changing the speed of the circulating pump and controlling the cooling water flow.

[0045] In one specific embodiment of the present invention, the prediction model is established based on the three-dimensional unsteady thermal conductivity differential equation of concrete, and its expression is: in, T Temperature (unit: °C) t Time (unit: seconds) r Concrete density (unit: kg / m³) 3 ), c eq ( t ) represents the equivalent specific heat capacity to be identified (unit: J / (kg·℃)). k eq ( t) represents the equivalent thermal conductivity to be identified (unit: W / (m·℃)). q ( t The heat release rate of hydration (unit: W / m) 3 ), which is directly related to the hydration heat release rate parameter to be identified.

[0046] To enable online computation, the finite difference method is used to discretize the above equations spatially using either a seven-point difference (three-dimensional) or a five-point difference (two-dimensional cross-sectional approximation) method, while an implicit time scheme is employed to ensure numerical stability. The discretized nodal temperature prediction equations can be uniformly written as follows: In the formula, T k For the first k Temperature vector of each node at each sampling time, u k For control inputs (current cooling water flow rate, ambient temperature, etc.) The vector represents the time-varying parameters to be identified. This discretized model can be directly deployed in a smart controller, recursively pushing forward at fixed time steps (e.g., 5–10 minutes) to output predicted temperature field values ​​for one or more future control cycles.

[0047] Taking a solid rectangular bridge pier as an example, the pier's cross-sectional dimensions are 7m × 2.8m, the casting height is 22m, and the concrete strength grade is C50. Temperature sensors are layered and placed at distances of 0.5m, 2.5m, 5m, 8m, 12m, and 16m from the top surface. Each layer has measuring points at the center, 0.4m from the surface, and on the surface itself, forming a spatial temperature measurement array. Cooling water pipes are made of steel pipes with an outer diameter of 42mm and a wall thickness of 3mm, with a vertical spacing of 1.0m, laid out in a serpentine pattern in each layer. The target temperature curve is set so that the maximum temperature does not exceed 70℃, and the maximum internal and external temperature difference does not exceed 25℃. The circulating pump has a rated flow rate of 12m³ / h. 3 Monitoring and control began immediately after concrete pouring. Initially, the heat of hydration was released slowly, and the predicted heat release rate identified by the model was low. The composite control signal kept the circulating pump at a low speed and the water flow rate small. As the heat of hydration increased rapidly, the measured and predicted temperatures rose almost synchronously. The online identification program used residual information to quickly update parameters such as the heat release rate and thermal conductivity. The predicted model anticipated the temperature rise trend within the next hour, and the feedforward control component increased first, instructing the pump speed to increase in advance and increasing the cooling water flow rate. This effectively reduced the temperature peak as it approached its peak, while the feedback component eliminated minor static errors. During the cooling period, the heat of hydration declined, and the temperature drop trend reflected in the predicted model caused the feedforward component to weaken accordingly. The pump speed decreased smoothly, and the concrete temperature slowly decreased along the target curve. Throughout the curing process, the maximum measured internal and external temperature difference remained stable at around 21℃.

[0048] A common existing temperature control method uses the same pier dimensions, concrete mix proportions, and cooling water pipe layout, but lacks online parameter identification and feedforward-feedback composite control. The traditional method pre-sets a cooling water supply schedule based on experience, supplying water for 1.5 hours every 3 hours, with the water pump running at a constant speed during the supply period, at a flow rate of approximately 8 m³ / h. 3 The system operates at a rate of / h, relying on a small number of internally embedded temperature monitoring points for alarms. When the temperature at any monitoring point exceeds 65℃, the water pump is activated and stops when the temperature drops to 60℃. This method is sustainable in the early stages after pouring, but during the peak hydration heat stage, the fixed water flow cycle fails to effectively respond to the instantaneous increase in heat release rate. By the time the monitoring points detect overheating and activate the water pump, heat has already accumulated inside the structure. Combined with the significant thermal inertia of concrete, the temperature drops slowly even with increased cooling water flow, reaching a peak temperature of 76℃, with the internal and external temperature difference reaching 33℃ at one point. Inspection after demolding revealed several fine temperature cracks on the side of the pier, caused by thermal stress exceeding the early tensile strength.

[0049] The intelligent temperature control curing system for large-volume concrete bridge piers of this invention solves the problem of temperature cracking caused by inaccurate predictions due to time-varying hydration heat parameters and lag in simple feedback adjustment during the curing of large-volume concrete. This is achieved by running a feedforward-feedback composite control program in an intelligent controller, which includes online adaptive parameter identification and an updated model. The spatiotemporal temperature field data sensed by this invention can not only be used for real-time monitoring but also directly drive the continuous self-correction of the prediction model parameters, ensuring that the model remains close to the actual thermal state throughout the entire curing cycle. Feedforward control actions are generated in advance based on reliable predictions, compensating for the inherent defect of slow concrete thermal response. The feedback channel suppresses model errors and random disturbances. The composite control signal formed by the superposition of these two aspects ensures that every adjustment of the circulating pump speed is both predictive and corrective, thereby precisely constraining the internal maximum temperature and the internal-external temperature difference within a safe range, preventing temperature cracking, and improving the curing quality and structural durability of large-volume concrete bridge piers.

[0050] According to one embodiment of the present invention, a smart temperature-controlled curing system for large-volume concrete in bridge piers, wherein when the processor executes the online adaptive parameter identification program, it is specifically used for: Based on the residuals, a Jacobian matrix is ​​constructed, and the equivalent specific heat capacity and equivalent thermal conductivity are identified online using the recursive least squares method. Alternatively, a state-space equation containing the heat release rate of hydration can be constructed, and the heat release rate of hydration can be estimated in real time using a Kalman filter.

[0051] In one implementation, the processor within the intelligent controller performs online adaptive parameter identification processing to classify time-varying parameters. First, the processor acquires the measured temperature values ​​at each measuring point on the pier body using a temperature sensor array. This value is then subtracted point-by-point from the temperature prediction values ​​output by the prediction model in the previous iteration to obtain a temperature residual vector. For the equivalent specific heat capacity and equivalent thermal conductivity, the identification program calculates partial derivatives of these two parameters based on the thermal balance control equation, constructing a Jacobian matrix. Each row of the Jacobian matrix corresponds to the sensitivity of a temperature measuring location to parameter changes. Subsequently, a recursive least squares algorithm is invoked, starting with the parameter estimates from the previous moment. The gain vector is calculated using the newly acquired residuals and the Jacobian matrix, recursively correcting the equivalent specific heat capacity and equivalent thermal conductivity. Simultaneously, the covariance matrix of the parameter estimates is updated, ensuring that the identified values ​​approach the true physical values ​​one step closer after each sampling. For the rapidly changing time-varying heat release rate of hydration, the identification program treats it as a state that changes over time, establishing a state-space equation with the heat release rate as the state variable and the measured temperature as the observation. A Kalman filter is then used for real-time estimation. Within each sampling period, the Kalman filter sequentially performs two steps: state prediction and measurement update. It uses the temperature residual and Kalman gain to pull the estimated heat release rate closer to the actual heat release level. The outputs of both are merged at the end of each control period to form a complete set of model parameters for the current moment, which is then fed into the temperature prediction stage. This embodiment uses a recursive least squares method based on the Jacobian matrix to identify the equivalent specific heat capacity and equivalent thermal conductivity, ensuring the smoothness and physical rationality of the estimated two fundamental thermophysical parameters. Simultaneously, the Kalman filter independently captures the instantaneous changes in the heat release rate of hydration, giving the identification system both stability and rapid response capabilities. The prediction model accurately reflects the true thermal state of concrete at each hydration stage, providing a reliable basis for subsequent feedforward and feedback control.

[0052] According to one embodiment of the present invention, a smart temperature-controlled curing system for large-volume concrete in bridge piers, wherein when the processor generates the feedforward control component, it is further configured to: The current cooling water flow rate, ambient temperature, and the updated time-varying parameters are input into the prediction model to calculate the predicted sequence of the concrete internal temperature field within a preset time domain.

[0053] In the specific implementation of generating the feedforward control components, the processor reads three real-time inputs at each control decision point: the current cooling water flow rate provided by the flow meter installed on the main cooling water pipe circuit, the ambient temperature sensor value located in the louvered box outside the pier, and the equivalent specific heat capacity, equivalent thermal conductivity, and hydration heat release rate parameters just updated by the online identification program. The processor simultaneously inputs these values ​​into a prediction model built based on the concrete heat balance equation. This model discretizes the concrete into several nodes and recursively solves for the temperature field distribution along the time axis in the future direction. The recursion duration can be set to a preset time domain that matches the system control cycle, such as 30 minutes or 1 hour. Through this forward extrapolation, the model outputs a predicted sequence of the internal temperature field of the concrete at each time point and node location within this preset time domain. The controller then extracts the expected temperature value at the corresponding time from the target temperature curve, calculates the deviation between the predicted sequence and the expected value at each time step, and generates the feedforward control components for this control operation by weighted summation of these deviations. This invention calculates the predicted sequence of concrete temperature field within a preset time domain by inputting the current cooling water flow rate, ambient temperature, and updated time-varying parameters into the prediction model. Based on this, a feedforward control component is generated, enabling the control system to sense the future trend of the temperature field in advance and adjust the cooling water volume accordingly. This fundamentally overcomes the adjustment lag problem caused by the thermal response lag of large-volume concrete, thereby initiating additional cooling before the peak of hydration heat arrives and reducing the cooling load in advance during the temperature drop period, so that the temperature amplitude and temperature difference remain stable and controllable throughout the entire curing process.

[0054] According to one embodiment of the present invention, a smart temperature control curing system for large-volume concrete in bridge piers, wherein the processor, when performing temperature control, is specifically used for: The composite control signal is converted into a PWM pulse width modulation signal or an analog voltage signal to control the speed of the circulating pump driven by the frequency converter.

[0055] When performing temperature control, the intelligent controller's output module converts the composite control signal into a physical signal that can directly drive the circulating pump inverter. When analog voltage control is selected, the processor outputs a 0V to 10V DC voltage signal via a digital-to-analog converter based on the value of the composite control signal. This voltage corresponds to the rated speed at the inverter's set frequency upper limit. An increase in the composite control signal indicates a need for stronger cooling, causing the output voltage to rise linearly. Upon receiving this signal, the inverter correspondingly increases its power supply frequency, increasing the circulating pump speed and cooling water flow. Conversely, a decrease in the signal leads to a decrease in voltage and a smooth drop in speed. When using PWM (Pulse Width Modulation), the processor maps the composite control signal to a square wave signal with an adjustable duty cycle from 0% to 100%. The square wave frequency is typically set within the inverter's input allowable range. The inverter linearly adjusts the output frequency based on the duty cycle, achieving stepless adjustment of the pump speed. Both interface methods can be flexibly selected based on the inverter model used on-site. This implementation converts the composite control signal generated by the intelligent controller into a PWM pulse width modulation signal or an analog voltage signal to control the circulating pump driven by the frequency converter. This achieves continuous and smooth speed regulation from zero to rated speed, allowing the cooling water flow to follow the changes in control requirements delicately. It eliminates the inherent flow step and temperature disturbance of start-stop regulation, significantly improves the stability of the cooling process and the accuracy of temperature control, and also reduces the mechanical and thermal stress impact on the water pump and pipeline.

[0056] According to one embodiment of the present invention, a smart temperature-controlled curing system for large-volume concrete in bridge piers is provided, wherein the processor is further configured to: At the moment the circulating pump is started, a thermal shock suppression period is set. During the thermal shock suppression period, the composite control signal is forced to output a preset low speed drive signal, and the shutdown logic based on the temperature difference threshold is blocked. The duration of the thermal shock suppression period is calculated by the processor based on the difference between the water temperature in the cooling water network and the current average temperature of the concrete, as well as the total volume of the cooling water network.

[0057] In one specific implementation, when the intelligent controller issues a cooling circulation command for the first time—that is, the moment the circulation pump is started for the first time after concrete pouring—the processor does not immediately hand over control of the pump to the feedforward-feedback composite control loop. Instead, it first enters a specially designed thermal shock suppression period. The processor first reads the temperature of each measuring point inside the concrete from the temperature sensor array to calculate the average temperature of the concrete. Simultaneously, it reads the values ​​from the water temperature sensors installed at the inlet of the cooling water network to calculate the difference between the water temperature inside the pipe and the average temperature of the concrete. The larger the temperature difference, the more severe the cold shock caused by the sudden flow of water. The processor also retrieves the total volume data of the pre-embedded cooling water network from memory. This total volume was calculated and entered into the system before construction based on the pipe diameter and length. By substituting the temperature difference and the total volume of the pipeline into a preset conversion formula, the processor automatically determines the duration of this thermal shock suppression period. During this suppression period, the intelligent controller forcibly bypasses the drive signal originally output by the composite control algorithm and continuously outputs a pre-set low-speed drive signal to the frequency converter, causing the circulating pump to operate at an extremely low speed. For example, this allows cooling water to slowly fill the pipe network at a very low flow rate. At the same time, the processor locks out the automatic shutdown protection logic based on the temperature difference threshold to prevent the pump from stopping erroneously due to a brief temperature difference exceeding the limit. Once the suppression period has elapsed, the processor automatically releases the forced low-speed command and shutdown shield, smoothly handing control back to the normal composite control closed loop. Thereafter, the circulating pump speed is entirely regulated in real time by the feedforward-feedback signal. This invention sets a thermal shock suppression period at the moment the circulating pump starts, forcing the circulating pump to operate at a preset low speed and shielding the temperature difference shutdown logic. This allows cooling water to gradually fill the pipe network at an extremely slow rate. The process of cold energy being transferred to the concrete through the pipe wall is significantly lengthened and mitigated. The concrete around the pipe does not generate a concentrated temperature gradient during the process of adapting to the low temperature. Thus, at the very beginning and most vulnerable moment of water cooling, it effectively avoids micro-cracks and macro-cracks in the concrete caused by thermal shock, protects the integrity of the pier structure, and lays the foundation for temperature control safety throughout the curing process.

[0058] According to one embodiment of the present invention, a smart temperature-controlled curing system for large-volume concrete in bridge piers is provided, the system further comprising a heating unit connected to the cooling water pipe network; the processor is further configured to: Calculate the real-time temperature difference between the internal temperature and the surface temperature of the concrete; When the real-time temperature difference value is less than or equal to the preset temperature difference threshold, but the internal temperature of the concrete exceeds the preset peak temperature, the circulating pump is controlled to maintain operation at a constant low speed. When the internal temperature of the concrete drops below the preset curing temperature, the circulating pump is turned off and the heating unit is started to heat the cooling water, so that the internal temperature of the concrete drops to the ambient temperature at a preset cooling rate.

[0059] In one specific implementation, the intelligent temperature-controlled curing system connects a heating unit in series on the main loop of the cooling water network. This heating unit can be a resistance heater or a hot water heat exchanger, and its start / stop and power are controlled by the digital output port of the intelligent controller. After the curing process enters the hydration cooling stage, the processor continuously acquires temperature data of each layer of the pier from the temperature sensor array and calculates two key indicators in real time: one is the real-time temperature difference between the highest internal temperature and the lowest surface temperature of the concrete, and the other is the highest internal temperature of the concrete. When the program determines that the real-time temperature difference is less than or equal to the preset temperature difference safety threshold, such as 25°C as specified in the standard, but the highest internal temperature of the concrete still exceeds the preset peak control temperature, it indicates that there is still accumulated heat inside that needs to be discharged, but the internal and external temperature difference is not drastic. At this time, the processor issues an instruction to make the circulating pump run continuously at a constant low speed, introducing a small amount of cooling water to slowly and steadily remove the internal residual heat, avoiding excessive cooling water use at this time, which would widen the internal and external temperature difference. As the heat of hydration gradually diminishes, when the highest internal temperature of the concrete finally drops below the preset curing temperature, the processor simultaneously performs two actions: first, it shuts off the circulating pump, stopping the cooling water circulation to prevent the concrete from cooling down too quickly; second, it immediately starts the heating unit to heat the water in the pipeline system and restarts the circulating pump at a low temperature and low speed, pushing the warm water into the cooling water network. This ensures that the internal temperature of the concrete does not drop abruptly, but rather decreases gradually and smoothly according to the preset cooling rate set within the processor, until it finally matches the ambient temperature. The entire process requires no manual intervention; the intelligent controller automatically switches between cooling and heating modes based on temperature sensing and temperature difference conditions. This embodiment connects a heating unit to the cooling water network, and the processor automatically performs phased coordinated control of cooling and heating based on the real-time temperature difference and internal temperature. When the temperature difference is within a safe range but the internal temperature is still high, low-speed cooling is maintained. When the internal temperature has dropped below the curing temperature, it automatically switches to heating and slow cooling, so that the hydration cooling process proceeds smoothly at a predetermined rate. This not only eliminates the temperature difference stress caused by excessively rapid cooling, but also avoids late-stage cracking caused by allowing the temperature to run at its own pace after cooling stops, significantly improving the safety of the entire curing cycle, especially the later stages.

[0060] According to one embodiment of the present invention, a smart temperature control curing system for large-volume concrete in bridge piers is provided, wherein the temperature sensing array is a distributed fiber optic temperature sensing array, which is embedded at multiple points at different depths and on the surface of the concrete to obtain the spatiotemporal temperature field data.

[0061] In one specific implementation, the temperature sensing array is entirely composed of a distributed fiber optic temperature sensing system. During construction, after the pier reinforcement cage is tied and before the formwork is closed, one or more multimode temperature-sensing optical fibers are vertically layered and tied to the main reinforcement bars and stirrups of the reinforcement cage. The fiber routing is specially designed: in the central area of ​​the pier, the fiber travels back and forth through the core concrete in a spiral or serpentine path to ensure coverage of the highest internal temperature zone; in the near-surface area close to the formwork, the fiber is arranged along the perimeter of the cross-section, with the distance from the inner surface of the formwork precisely controlled to reflect the temperature gradient near the surface; in certain critical sections, the fiber is also directly adhered to the inner wall of the formwork, and after demolding, the fiber is attached to the concrete surface to directly collect surface temperature data. One end of the temperature-sensing fiber is led to a distributed fiber optic temperature measurement host outside the pier, which has a built-in pulsed laser source and Raman scattering demodulation module. During temperature measurement, the host emits nanosecond-level laser pulses into the fiber. As the pulse propagates in the fiber, backscattered Raman light is generated at various locations. Stokes light is insensitive to temperature, while the intensity of anti-Stokes light changes significantly with temperature. The host computer collects the intensity of both light and light using a high-speed photoelectric detector, and calculates the temperature value at approximately one meter interval along the optical fiber based on the relationship between light speed and return time. All temperature measurement points are connected in series to form a complete temperature profile curve, which is then combined to generate spatiotemporal temperature field data covering the entire height and cross-section of the pier. The data is transmitted to the intelligent controller in real time via a communication interface. This implementation method uses a distributed optical fiber temperature sensor array, which is buried at multiple depths and on the surface of the concrete. A single thin optical fiber replaces a large number of point sensors, acquiring high spatial density continuous temperature distribution information along the entire height and width of the pier without weakening the concrete cross-section. This makes the distribution of hot and cold spots, the direction and magnitude of temperature gradients in the internal temperature field readily apparent, providing sufficient spatial temperature data support for the prediction model. This makes the temperature field reconstruction and identification results more accurate and reliable, enabling zoned differentiated control and improving the scientific nature and accuracy of temperature control decisions.

[0062] According to one embodiment of the present invention, a smart temperature control curing system for large-volume concrete in bridge piers is provided, wherein the temperature sensing array is a distributed fiber grating array, and the smart controller is further used to demodulate the temperature and strain data of each measuring point according to wavelength division multiplexing technology, and to use the temperature data as the spatial boundary constraint condition of the prediction model.

[0063] In one implementation, a distributed fiber Bragg grating array is used for the temperature sensing array. During construction, fiber Bragg grating strings are tied at different elevations and cross-sectional locations within the pier's reinforcing steel skeleton. Dozens of Bragg gratings with different center wavelengths are written onto each fiber using phase masking technology, with each grating constituting an independent temperature and strain measurement point. The grating string layout covers the high-temperature zone in the pier core, the near-surface transition zone at different distances from the surface, and the surface area exposed to the atmosphere after demolding. The fiber Bragg grating array's output is connected to the wavelength demodulation module of an intelligent controller. This module incorporates a broadband swept-frequency light source and a Fabry-Perot filter or diffraction grating spectrometer. During demodulation, broadband light is injected into the fiber, and each grating reflects its own narrowband light at its center wavelength. The wavelength demodulation module uses wavelength division multiplexing (WDM) technology to distinguish different grating measurement points based on the differences in reflected wavelengths, accurately measuring the minute offset of the center wavelength of each grating. Since wavelength offset incorporates contributions from both temperature and strain variations, the processor incorporates a temperature-strain cross-sensitivity decoupling algorithm. This algorithm uses a reference grating or differential compensation structure to separate the actual temperature and strain values ​​at each measuring point from the total offset. The separated temperature data is directly used as the spatial boundary constraint for the prediction model: during each model solution, the processor embeds the node temperature value corresponding to the fiber grating measuring point as a defined boundary constraint into the equation solving process, replacing the previous indirect boundary setting based on ambient temperature and convective heat transfer coefficient. Strain data is recorded independently, and the controller continuously compares the strain value at each measuring point with the ratio of the ultimate tensile strain of the concrete at its current age. An early warning is issued when the strain approaches a dangerous threshold. This implementation uses wavelength division multiplexing (WDM) technology to simultaneously demodulate temperature and strain data from various measuring points, and uses the temperature data as the spatial boundary constraint condition for the prediction model. This allows the boundary conditions required for temperature field simulation to be directly driven by measured values ​​rather than rough assumptions, significantly improving the consistency between the prediction model's calculation results and the actual temperature field. Simultaneously, strain monitoring information is used to achieve synchronous perception of cracking risks. Without adding additional sensors, the function of the temperature control system is expanded from simple temperature control to a fusion of temperature control and structural safety monitoring, providing more comprehensive protection for bridge pier maintenance.

[0064] According to one embodiment of the present invention, a smart temperature control curing system for large-volume concrete of bridge piers is provided, wherein the cooling water pipe network is a multi-group serpentine pipe cluster arranged in layers, and the inlet of each group of pipe clusters is provided with an electronic flow regulating valve controlled by the smart controller. The processor is also used to generate independent opening signals for each of the electronic flow control valves based on the spatial temperature field gradient output by the prediction model, so as to achieve differentiated cooling control in different regions.

[0065] In one implementation, the cooling water network employs a layered, independently controlled serpentine pipe cluster layout. A set of serpentine cooling loops is installed every two to three meters vertically along the pier. Each set consists of a steel pipe repeatedly bent and coiled to form a flat, serpentine heat dissipation surface. The inlets of each pipe cluster converge at a distributor located on the outer side of the pier. Each outlet branch of the distributor is equipped with an electronic flow regulating valve. The actuator of this valve receives control signals from the intelligent controller, and the outlets converge at a collector before connecting to a circulating pump circuit. During temperature control, the intelligent controller runs a predictive model to solve for the temperature field across the entire pier height. This not only provides the temperature variation over time but also calculates and outputs the spatial temperature field gradient for each elevation layer, i.e., the temperature difference between adjacent layers and the temperature distribution difference between the center and edge within the same layer. Based on this gradient information, the processor automatically generates independent opening commands for the electronic flow control valves corresponding to each group of serpentine pipe clusters: for sections with the fastest temperature rise and the most severe core heat accumulation, the controller outputs a larger valve opening signal, increasing the cooling water flow rate in that layer to remove more heat of hydration; for sections near the pier top or bottom, with a large heat dissipation area or good natural ventilation, the controller reduces the corresponding valve opening, limiting the cooling water supply to prevent these areas from experiencing excessively low temperatures and excessive internal and external temperature differences due to overcooling. Each valve responds in real time and dynamically matches the actual cooling demand of its area. This implementation design, by using a multi-group independent serpentine pipe network arranged in layers and installing electronic flow control valves independently controlled by an intelligent controller at each group's inlet, achieves differentiated cooling control by region based on the spatial temperature field gradient output by the predictive model. This allows the cooling water volume to be automatically and optimally allocated according to the actual heat dissipation demand of each region, effectively smoothing out the temperature differences between the upper and lower parts of the pier and the internal and external parts, reducing the overall temperature gradient of the structure, thereby fundamentally reducing the risk of cracks caused by local temperature stress concentration and significantly improving the precision and uniformity of temperature control.

[0066] According to one embodiment of the present invention, a smart temperature-controlled curing system for large-volume concrete in bridge piers is provided. When the processor executes the online adaptive parameter identification program, it specifically employs a recursive least squares method with a variable forgetting factor, and adjusts the magnitude of the forgetting factor in real time based on the temperature change rate collected by the temperature sensing array. When the temperature change rate exceeds a preset threshold, the forgetting factor is reduced to enhance the tracking sensitivity of the identification algorithm to the current data. When the temperature change rate is lower than the preset threshold, the forgetting factor is increased to suppress fluctuations in the identification results caused by measurement noise, thereby ensuring that the time-varying parameters of the prediction model can stably converge to the true physical characteristics throughout the entire curing cycle.

[0067] In one implementation, when the processor of the intelligent controller executes an online adaptive parameter identification program to update the equivalent specific heat capacity and equivalent thermal conductivity, the core identification algorithm employs a recursive least squares method with a variable forgetting factor. The forgetting factor is a key parameter determining the degree to which the identification algorithm depends on historical data: the smaller the forgetting factor, the faster the old data decays, and the higher the algorithm's sensitivity to newly arrived measured temperature values; the larger the forgetting factor, the longer the historical data is retained, resulting in a smoother estimation but a slower response to new changes. In each identification cycle, the processor uses the current temperature value collected by the temperature sensor array at each measuring point and the temperature value of the previous cycle to calculate the average temperature change rate of a representative area of ​​the pier concrete, in °C / h. The processor compares this temperature change rate with two preset threshold values. When the hydration reaction is in a phase of rapid temperature rise, and the rate of temperature change exceeds a preset higher threshold, the processor automatically reduces the value of the forgetting factor, for example, lowering it from the default value to around 0.88. This allows the identification algorithm to assign greater weight to the latest temperature residual, enabling the estimated values ​​of equivalent specific heat capacity and thermal conductivity to converge more quickly to the actual thermophysical parameters of the concrete at this moment, avoiding distortion of the prediction model during the peak temperature rise period due to identification lag. When the hydration reaction slows down and the rate of temperature change falls below the preset lower threshold, the concrete enters a quasi-steady-state slow heat dissipation phase. The relative influence of measurement noise on temperature data increases. At this time, the processor automatically increases the forgetting factor, for example, raising it to around 0.99, allowing the identification algorithm to rely more on accumulated historical observation information, smoothing parameter estimation fluctuations caused by random noise, and keeping the equivalent specific heat capacity and equivalent thermal conductivity stable during slow changes, without significant jumps due to individual measurement outliers. The adjustment of the forgetting factor is automatically completed in each iteration step of the identification program, requiring no external intervention. This implementation method employs a recursive least squares method with a variable forgetting factor, and adjusts the forgetting factor in real time based on the temperature change rate collected by the temperature sensor array. During the rapid heating stage of hydration, the forgetting factor is reduced to enhance the sensitivity of the identification to real-time changes. During the stable cooling stage of hydration, the forgetting factor is increased to suppress parameter fluctuations caused by noise. This ensures that the time-varying parameters in the prediction model always converge stably to the true physical properties of concrete throughout the entire curing cycle, regardless of whether they are in a state of rapid or slow change. This guarantees the long-term reliability of the closed-loop chain of identification-prediction-control and avoids temperature control errors caused by inaccurate parameters.

[0068] Forgetting factor in recursive least squares algorithm l ( k The following rules will be applied for dynamic adjustment: First, the current rate of temperature change is calculated from representative measuring points (such as near the pier center) collected by the temperature sensing array: Where Δ t sTemperature sampling interval (unit: h). Two preset thresholds: low threshold... r low =0.5 / h, high threshold r high =2.0 / h.

[0069] when When the hydration reaction is intense or the temperature changes rapidly due to cold shock, the forgetting factor is taken. l ( k The value is 0.88~0.92 (the specific value is negatively correlated with the sampling noise level) to enhance the sensitivity to new data; when When this time, it indicates the start of a quasi-steady-state slow cooling period. l ( k () = 0.98~0.99, to suppress parameter fluctuations caused by measurement noise; when At that time, take l ( k )=0.95 is used as the default value.

[0070] The aforementioned threshold and forgetting factor values ​​can be fine-tuned on-site based on the peak hydration heat rate and sensor noise level of the actual project, but the adjustment range should follow the principle of rapid change. l Stable rise l The principle.

[0071] (II) Formula for converting the duration of thermal shock suppression period Assume the temperature difference between the cooling water network temperature and the current average temperature of the concrete is . (Unit: °C), the total volume of the cooling water pipe network is V (Unit: m) 3 The water flow rate corresponding to the preset low speed of the circulating pump during the inhibition period is... Q low (Unit: m) 3 / s). Then the duration of the thermal shock inhibition period. (Unit: s) Calculate using the following formula: in, α For safety margin (recommended value is 1.0 to 1.2), β This is a temperature difference correction factor (recommended value: 0.05–0.10℃). -1 When the temperature difference is large, the inhibition period should be extended to introduce cold water more gradually; when the temperature difference is small, the inhibition period can be appropriately shortened. Unless otherwise specified, a simplified method can be used. (That is, the time required to ensure that the cold water fills the pipe network once), and then add an additional time of no more than 50% linearly according to the temperature difference.

[0072] The processor reads the temperature difference Δ in real time. T and preset V , Q low Automatic calculation During this period, a low-speed drive signal is forcibly output, while the shutdown logic based on the temperature difference threshold is disabled. After the suppression period expires, it automatically switches to the normal composite control loop.

[0073] Although embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. It can be applied to various fields suitable for the embodiments of the present invention. Other modifications can be readily implemented by those skilled in the art. Therefore, without departing from the general concept defined by the claims and their equivalents, the embodiments of the present invention are not limited to the specific details.

Claims

1. A smart temperature-controlled curing system for large-volume concrete in bridge piers, characterized in that, include: Temperature sensing array, used to collect spatiotemporal temperature field data of the inside and surface of concrete in real time; Cooling water piping is embedded in the concrete. A circulating pump, connected between the water supply unit and the cooling water network, is used to drive the circulation of cooling water; The intelligent controller is electrically connected to the temperature sensor array and the circulating pump, respectively. The intelligent controller includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to perform the following steps: A prediction model based on concrete thermal equilibrium is constructed, which includes equivalent specific heat capacity, equivalent thermal conductivity and hydration heat release rate as time-varying parameters to be identified. An online adaptive parameter identification program is executed, taking the residual between the measured temperature value collected by the temperature sensing array at the current moment and the temperature prediction value output by the prediction model at the previous moment as input. The recursive least squares method or Kalman filter algorithm is used to continuously update the time-varying parameters in the prediction model so that the temperature prediction value output by the updated prediction model at the current moment converges to the measured temperature value. A feedforward-feedback composite control signal is generated. Based on the updated prediction model, the predicted temperature value after a preset control period is calculated. The deviation between the predicted temperature value and the target temperature curve is used as the feedforward control component, and the deviation between the measured temperature value and the target temperature curve is used as the feedback control component. The composite control signal is generated by superimposing these components. Temperature control is implemented, and the speed of the circulating pump is dynamically adjusted according to the composite control signal.

2. The intelligent temperature-controlled curing system for large-volume concrete in bridge piers as described in claim 1, characterized in that, When the processor executes the online adaptive parameter identification program, it is specifically used for: Based on the residuals, the Jacobian matrix is ​​constructed, and the equivalent specific heat capacity and equivalent thermal conductivity are identified online using the recursive least squares method. Alternatively, a state-space equation containing the heat release rate of hydration can be constructed, and the heat release rate of hydration can be estimated in real time using a Kalman filter.

3. The intelligent temperature-controlled curing system for large-volume concrete in bridge piers as described in claim 1, characterized in that, When the processor generates the feedforward control components, it is also used for: Input the current cooling water flow rate, ambient temperature, and updated time-varying parameters into the prediction model to calculate the predicted sequence of the concrete internal temperature field within a preset time domain.

4. The intelligent temperature-controlled curing system for large-volume concrete in bridge piers as described in claim 1, characterized in that, When the processor performs temperature control, it is specifically used for: The composite control signal is converted into a PWM pulse width modulation signal or an analog voltage signal to control the speed of the circulating pump driven by the frequency converter.

5. The intelligent temperature-controlled curing system for large-volume concrete in bridge piers as described in claim 1, characterized in that, The processor is also configured as follows: At the moment the circulating pump is started, a thermal shock suppression period is set. During the thermal shock suppression period, the composite control signal is forced to output a preset low speed drive signal, and the shutdown logic based on the temperature difference threshold is disabled. The duration of the thermal shock suppression period is calculated by the processor based on the difference between the water temperature in the cooling water network and the current average temperature of the concrete, as well as the total volume of the cooling water network.

6. The intelligent temperature-controlled curing system for large-volume concrete in bridge piers as described in claim 1, characterized in that, The system also includes a heating unit connected to the cooling water network; the processor is also configured to: Calculate the real-time temperature difference between the internal temperature and the surface temperature of the concrete; When the real-time temperature difference is less than or equal to the preset temperature difference threshold, but the internal temperature of the concrete exceeds the preset peak temperature, the circulating pump is controlled to maintain operation at a constant low speed. When the internal temperature of the concrete drops below the preset curing temperature, the circulation pump is turned off and the heating unit is started to heat the cooling water, so that the internal temperature of the concrete drops to the ambient temperature at a preset cooling rate.

7. The intelligent temperature-controlled curing system for large-volume concrete in bridge piers as described in claim 1, characterized in that, The temperature sensing array is a distributed fiber optic temperature sensing array, which is buried at multiple points at different depths and on the surface of the concrete to obtain spatiotemporal temperature field data.

8. The intelligent temperature-controlled curing system for large-volume concrete in bridge piers as described in claim 1, characterized in that, The temperature sensing array is a distributed fiber grating array. The intelligent controller is also used to demodulate the temperature and strain data of each measuring point according to wavelength division multiplexing technology, and to use the temperature data as the spatial boundary constraint condition of the prediction model.

9. The intelligent temperature-controlled curing system for large-volume concrete in bridge piers as described in claim 1, characterized in that, The cooling water network consists of multiple serpentine pipe clusters arranged in layers, and each cluster's inlet is equipped with an electronic flow regulating valve controlled by an intelligent controller. The processor is also used to generate independent opening signals for each electronic flow control valve based on the spatial temperature field gradient output by the prediction model, so as to achieve differentiated cooling control in different regions.

10. The intelligent temperature-controlled curing system for large-volume concrete in bridge piers as described in claim 1, characterized in that, When the processor executes the online adaptive parameter identification program, it specifically adopts the recursive least squares method with a variable forgetting factor, and adjusts the size of the forgetting factor in real time according to the temperature change rate collected by the temperature sensor array. When the temperature change rate exceeds the preset threshold, the forgetting factor is reduced to enhance the tracking sensitivity of the identification algorithm to the current data. When the temperature change rate is lower than the preset threshold, the forgetting factor is increased to suppress the fluctuation of the identification results caused by measurement noise, thereby ensuring that the time-varying parameters of the prediction model can stably converge to the true physical characteristics throughout the entire maintenance cycle.

Images