A method for dynamic monitoring and adaptive control of the casting process using a moving formwork

Abstract

This invention relates to the field of civil engineering technology and discloses a dynamic monitoring and adaptive control method for the pouring process of a moving formwork. The method includes the following steps: synchronously collecting boundary measurement data from an array of acoustic and electrical sensors on the moving formwork; reconstructing the three-dimensional distribution of state parameters such as internal density, hydration degree, temperature, and free water content of the concrete in real time by solving a joint inversion objective function that includes acoustic and electrical forward models; comparing the reconstructed internal state field with a preset optimal state trajectory to obtain the state deviation; and then using a model predictive control algorithm combined with a discrete-time state transition model to generate and execute control commands that adjust construction parameters such as the power of the vibration system, the heat flux rate of the formwork temperature control, and the formwork travel speed. This invention achieves real-time and comprehensive monitoring of the internal physical state of concrete and proactively ensures the quality of concrete pouring through adaptive and predictive control.

Description

Technical Field

[0001] This invention relates to the field of civil engineering technology, specifically to a method for dynamic monitoring and adaptive control of the pouring process using a moving formwork. Background Technology

[0002] Large-volume concrete structures, such as bridges, dams, and nuclear power plant containment structures, are often constructed using sliding formwork (moving formwork) for continuous pouring. This construction method offers advantages such as high efficiency and good structural integrity. However, during the pouring, vibration, hydration, and hardening processes of concrete, key physical parameters such as internal density, degree of hydration, temperature, and free water content dynamically change over time and space. Loss of control over these internal states can easily lead to quality defects in the concrete, such as porosity, cracks, and insufficient strength, severely impacting the structural performance and safety.

[0003] Current technologies for monitoring the concrete pouring process using moving formwork primarily rely on point sensors (such as temperature sensors and strain gauges) for localized data acquisition, or on single non-destructive testing techniques (such as ultrasonic testing and resistivity methods) to obtain limited-dimensional information. However, these methods typically struggle to provide the three-dimensional spatial distribution and real-time evolution of multiple key state parameters within the concrete, such as internal density, hydration degree, temperature, and free water content. This makes it impossible to comprehensively and accurately characterize the overall quality and potential defects of the concrete. The limited sensitivity of single-sensor modes to the complex physical states within concrete results in insufficient spatial coverage and parameter characterization of the acquired monitoring data, making it difficult for construction personnel to accurately determine the overall density and hydration uniformity of the concrete.

[0004] Furthermore, existing technologies for controlling the concrete pouring process mostly employ open-loop control based on experience or preset parameters, or reactive adjustments based on local, surface measurement data. These methods typically lack accurate simulation of deep physical processes such as concrete hydration kinetics and thermodynamics, and fail to fully utilize real-time acquired internal state information for forward-looking prediction and optimization. Simple feedback regulation is insufficient to effectively address the complex nonlinear and time-varying characteristics of the concrete's internal state, resulting in insufficient robustness and accuracy of the control strategy. Therefore, it is difficult to achieve refined, adaptive adjustment of construction parameters such as vibration power, formwork temperature control, or travel speed to ensure that the concrete's internal state always evolves along the optimal trajectory, thereby proactively preventing and controlling quality problems. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a dynamic monitoring and adaptive control method for the pouring process of a moving formwork. This method solves the problem that existing technologies struggle to obtain real-time and comprehensive three-dimensional dynamic evolution of key state parameters such as concrete density, hydration degree, temperature, and free water content, and lack forward-looking adaptive control methods based on accurate physical models, making it difficult to proactively ensure pouring quality.

[0006] This invention provides a method for dynamic monitoring and adaptive control of the casting process using a moving formwork, the method comprising the following steps:

[0007] Step S1: Synchronously collect boundary measurement data.

[0008] This step utilizes acoustic and electrical sensor arrays arranged on a moving mold frame to simultaneously acquire acoustic and electrical boundary measurement data. The acoustic sensor array consists of piezoelectric ceramic transducers used to transmit and receive ultrasonic signals. The electrical sensor array consists of metal electrodes used to apply excitation current and measure boundary potentials.

[0009] Step S2: Real-time reconstruction of the internal state field of concrete.

[0010] This step, based on the boundary measurement data acquired in step S1, calculates the three-dimensional distribution of the internal state field of the concrete in real time by solving a joint inversion objective function. The internal state field is defined as the state field within a given spatial location. and time A state vector Including density hydration degree ,temperature and free water content At least one state parameter in it.

[0011] Preferably, the internal state field also includes density. hydration degree ,temperature and free water content The three-dimensional distribution.

[0012] Preferably, the joint inversion objective function It is set as the sum of the acoustic data fitting term, the electrical data fitting term, and the regularization term, and its form is as follows:

[0013] ;

[0014] in, The internal state field to be solved; These are the actual acoustic boundary measurement data collected. These are the actual electrical boundary measurement data collected. The acoustic data fitting term is:

[0015] ;

[0016] Acoustic boundary data used to measure predictions based on the internal state field. Acoustic boundary measurement data actually collected The differences between them Its weight matrix. The electrical data fitting term is:

[0017] ;

[0018] Electrical boundary data used to measure predictions based on the internal state field. Compared with the actual collected electrical boundary measurement data The differences between them Its weight matrix. The regularization term... This is used to provide prior constraints for solving the joint inversion objective function, so as to improve the stability and uniqueness of the inversion results.

[0019] In one specific embodiment, the predicted acoustic boundary data includes the sound wave propagation time calculated using an acoustic forward model. The predicted electrical boundary data includes the boundary potential distribution calculated using an electrical forward model.

[0020] Preferably, the acoustic forward model establishes the sound velocity field based on the state parameters in the internal state field. The functional relationship between the state parameters and the state parameters:

[0021] ;

[0022] speed of sound in a medium It is the internal state. The function is mainly affected by density and hydration. Based on the sound velocity field, the first... acoustic wave path sound wave propagation time The calculation formula is as follows:

[0023] ;

[0024] Preferably, the electrical forward model establishes the conductivity field based on the state parameters in the internal state field. The functional relationship between the state parameters is as follows:

[0025] ;

[0026] conductivity of the medium Both are internal states The boundary potential distribution is obtained by solving the following boundary value problem based on the conductivity field, which is particularly sensitive to free water content and temperature. :

[0027] ;

[0028] in, This is the area for concrete pouring. For gradient operators, This is the dot product operator.

[0029] Step S3: Compare the state field with the optimal state trajectory.

[0030] This step will reconstruct the internal state field of the concrete in real time. Compared with the preset optimal state trajectory A point-by-point comparison is performed to obtain the state deviation, which reflects the difference between the current state and the ideal state. The optimal state trajectory is a path that describes the ideal trajectory of the evolution of each state parameter in the internal state field over time, which is predetermined based on the concrete mix proportion, component design, and environmental conditions.

[0031] Step S4: Generate and execute control commands.

[0032] Based on the state deviation, this step generates control instructions for adjusting the construction parameters of the moving formwork through a model predictive control algorithm, and then executes the control instructions.

[0033] In one specific embodiment, the model predictive control algorithm in each control cycle Internally, by solving a time domain prediction algorithm for the future. To generate the optimal control commands for the range optimization problem.

[0034] The objective of solving the optimization problem is to make the future evolution process of the internal state field... Approaching the optimal state trajectory Furthermore, penalties are imposed on the amplitude or rate of change of the control command. The optimization problem can be expressed as:

[0035] ;

[0036] The constraints are:

[0037] ;

[0038] in, It is the control sequence to be optimized; and These are the weight matrices for the state trajectory deviation penalty term and the control input magnitude penalty term, respectively; The current moment; It is a pre-set future moment. For prediction in the time domain; It controls the length of the time domain, usually ; and These are the lower and upper limits of the control commands, respectively, to ensure that the control commands are within the physically feasible range; Is Time prediction The system state at any given moment; It is a pre-set future moment. The optimal state trajectory.

[0039] Preferably, to solve the optimization problem, a discrete-time state transition model that couples the hydration kinetics and thermodynamics of concrete is used. Used to determine the current internal state field and input control commands Predict the future evolution process of the internal state field. The predicted future evolution process is then used to solve the optimization problem.

[0040] In one specific embodiment, the control commands for adjusting the construction parameters of the movable formwork include commands for adjusting at least one of the following parameters: vibration system power. Heat flux of the mold frame temperature control system Travel speed of the moving formwork The control command vector It can be represented as:

[0041] ;

[0042] Preferably, within each control cycle, the solution to the optimization problem is an optimal control sequence composed of control commands at multiple future time points. The method executes only the first control command in the optimal control sequence. In the next control cycle... Based on the new real-time reconstructed internal state field Repeat the process of solving the optimization problem.

[0043] This invention provides a method for dynamic monitoring and adaptive control of the casting process using a moving formwork. It offers the following advantages:

[0044] 1. This invention synchronously collects boundary measurement data from acoustic and electrical sensor arrays arranged on a moving formwork, and reconstructs the internal state field of concrete in real time by solving a joint inversion objective function based on the boundary measurement data. It can obtain the three-dimensional distribution of various state parameters such as concrete density, hydration degree, temperature and free water content. By utilizing the complementarity of multimodal sensing information, it overcomes the limitations of reconstructing multi-physics field states in complex media by a single sensing mode, and improves the accuracy and comprehensiveness of internal state field reconstruction.

[0045] 2. This invention applies a joint inversion algorithm based on acoustic and electrical forward modeling, along with a discrete-time state transition model that couples concrete hydration kinetics and thermodynamic processes, to the reconstruction and prediction of the internal state field. This gives the calculation and future evolution prediction of the internal state field a clear physical basis, enhances the physical authenticity of the internal state field reconstruction results and the predictive reliability of the model predictive control algorithm, and provides a solid basis for subsequent control decisions.

[0046] 3. This invention obtains the state deviation by comparing the real-time reconstructed internal state field with the preset optimal state trajectory, and uses a model predictive control algorithm to dynamically generate and execute control commands for adjusting the construction parameters of the moving formwork based on the state deviation. This achieves adaptive and predictive control of the concrete pouring process, which can promptly correct the internal state deviation generated during construction, guide the internal state parameters of the concrete to evolve along the ideal path, and thus actively maintain and ensure the quality of concrete pouring. Attached Figure Description

[0047] Figure 1 This is a schematic diagram of the dynamic monitoring and adaptive control method for the casting process of a moving formwork according to an embodiment of the present invention;

[0048] Figure 2 This is a schematic diagram of the model predictive control algorithm in an embodiment of the present invention. Detailed Implementation

[0049] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0050] Please see the appendix Figure 1-2 This invention provides a method for dynamic monitoring and adaptive control of the casting process using a moving formwork, comprising the following steps:

[0051] First, the concrete during the pouring process is measured in real time using a sensor array. Specifically, acoustic and electrical sensor arrays, arranged on a moving formwork, synchronously acquire boundary measurement data of the concrete. The acoustic sensor array is used to acquire data related to ultrasonic wave propagation, while the electrical sensor array is used to acquire boundary potential data.

[0052] The acoustic sensor array acquires ultrasonic propagation-related data, including sound wave propagation time. The electrical sensor array acquires boundary potential data in response to the applied excitation current. This synchronous acquisition ensures that data from different physical modes correspond to the internal state of the concrete at the same point in time.

[0053] Based on boundary measurement data, the method reconstructs the internal state field of concrete in real time. The internal state field is defined as the state field at a given spatial location. and time Physical state vector at the location Its components can include density. hydration degree ,temperature and free water content At least one of the following. The reconstruction process involves solving a joint inversion objective function. accomplish.

[0054] In one embodiment, the internal state field The components include density. hydration degree ,temperature and free water content The three-dimensional distribution.

[0055] Joint inversion objective function The format is as follows:

[0056] ;

[0057] in, The internal state field to be solved; Acoustic boundary data predicted based on the internal state field; These are the actual acoustic boundary measurement data collected. Electric boundary data based on predictions from the internal state field; These are actual electrical boundary measurement data collected. This is the weight matrix for the fitting terms of the electrical data; This is a regularization term used to introduce prior constraints;

[0058] The sound velocity field is calculated using the acoustic forward model, which establishes the acoustic forward model. Functional relationship with internal state field And calculate the sound wave propagation time based on the sound speed field. ,in For the first A sound wave propagation path, For path infinitesimal elements;

[0059] The conductivity field is calculated using the forward electrical model, which establishes the conductivity field. Functional relationship with internal state field And solve the boundary value problem of electric potential based on the conductivity field. ,in For potential distribution, This is the area for concrete pouring. For gradient operators, This is the dot product operator.

[0060] Solving the joint inversion objective function is a nonlinear optimization problem, which is solved numerically using an iterative algorithm.

[0061] Next, the internal state field will be reconstructed in real time. The trajectory of the preset optimal state By comparing the results, the state deviations are obtained. The optimal state trajectory is a path that describes the ideal trajectory of the evolution of each state parameter in the internal state field over time, which is predetermined based on the concrete mix proportion, component design, and environmental conditions.

[0062] Finally, based on the state deviation, a model predictive control (MPC) algorithm is used to generate control commands for adjusting the construction parameters of the moving formwork, and these control commands are executed. The MPC algorithm operates in each control cycle. Within, solve a problem in the time domain that predicts the future. This is a range optimization problem to determine the optimal control sequence. The objective function of the optimization problem Set as:

[0063] ;

[0064] in, For at any time Predicted future moments The internal state field of the system For the future The optimal state trajectory. For the future The control input vector, The weight matrix is ​​the weight matrix for the state trajectory deviation penalty term. The weight matrix is ​​used to control the penalty term for the input size.

[0065] Model predictive control algorithms penalize the magnitude or rate of change of control commands.

[0066] The optimization problem is constrained by the discrete-time state transition model. and initial state .in This is a discrete-time state transition model that couples the hydration kinetics and thermodynamic processes of concrete.

[0067] Control input vector The components may include the power of the vibration system. Heat flux of the mold frame temperature control system and the speed of movement of the moving mold .

[0068] In each control cycle, the MPC algorithm executes only the first control instruction in the optimal control sequence, denoted as . The above process is repeated in the next control cycle based on the new real-time reconstructed internal state field.

[0069] In step S1, the method simultaneously acquires boundary measurement data from acoustic and electrical sensor arrays arranged on the moving formwork. This process aims to obtain non-invasive boundary information reflecting the internal physical state of the concrete.

[0070] Acoustic sensor arrays typically consist of multiple piezoelectric ceramic transducers (PZTs) embedded or attached to the inner wall of a moving formwork in contact with the concrete in a specific array pattern (e.g., a ring array or a rectangular array). During data acquisition, specific transducers in the acoustic sensor array are driven to emit ultrasonic pulses, while other transducers act as receivers to detect the ultrasonic signals after they have traveled through the concrete medium. For each transmitter-receiver pair, the sensor array records the propagation time of the ultrasonic signal, forming acoustic boundary measurement data.

[0071] The frequency range of ultrasonic pulses is typically between 50 kHz and 500 kHz. The propagation time of sound waves can be measured by detecting the first arrival time of the signal or by determining the phase delay between the two sensor signals using a cross-correlation algorithm.

[0072] The contact between the transducer and the concrete is typically optimized using a coupling agent (such as water or a specialized gel) to ensure efficient transfer of acoustic energy. A typical density of sensor arrays can be 4 to 16 transducers per square meter to ensure effective coverage of the target area.

[0073] The electrical sensor array consists of multiple conductive electrodes, which are also embedded or attached to the inner wall of the moving mold in a specific array pattern (e.g., a ring array). They are typically arranged spatially in an alternating or coordinated manner with the acoustic sensor array to cover the same monitoring area. During data acquisition, the electrical sensor array applies a preset excitation current through different electrode pairs, while simultaneously measuring the resulting boundary potential on the remaining electrode pairs. By changing the current injection mode and performing multiple sets of measurements, a series of boundary potential data can be obtained, forming electrical boundary measurement data.

[0074] The excitation current is typically a low-frequency alternating current or direct current to avoid polarization effects and frequency dependence. The amplitude and frequency of the excitation current are set according to the electrical characteristics of the concrete and the sensitivity of the measurement system. Commonly used current injection modes include adjacent injection measurements and skip injection measurements.

[0075] The electrodes can be made of corrosion-resistant and conductive materials such as stainless steel or titanium alloy to ensure long-term stable operation in concrete environments. Typical electrode sizes are circular sheets with a diameter of 10 mm to 30 mm or strips with a length of 50 mm to 100 mm.

[0076] Synchronous acquisition refers to the operation of acoustic and electrical sensor arrays coordinated by a master control unit, ensuring that boundary measurement data for two different modes are acquired within a very short time window. This tight time synchronization guarantees that the acquired acoustic and electrical data correspond to the instantaneous internal state of the concrete at a specific monitoring moment, thus providing consistent input for subsequent joint inversion. The frequency of synchronous acquisition can be set according to the hardening rate of the concrete and the control cycle requirements, for example, performing a complete boundary data scan every few minutes.

[0077] The main control unit ensures precise timing alignment between acoustic signal transmission and electrical current excitation through a shared clock signal or a precise triggering mechanism, and records their respective responses. For example, a common sampling trigger signal can be set, and acoustic data acquisition and electrical data acquisition can be started simultaneously on the rising edge of the trigger signal.

[0078] The main control unit also includes a data storage module, which stores the collected boundary measurement data along with timestamps for subsequent processing and analysis. The data storage medium can be a solid-state drive (SSD) or a network storage device.

[0079] In step S2, the method reconstructs the internal state field of the concrete in real time based on the boundary measurement data acquired in step S1. This process aims to infer the three-dimensional spatial distribution of the physical state parameters inside the concrete that cannot be directly measured from the externally measurable response by solving an inverse problem.

[0080] The internal state field is defined as the state of concrete at any spatial location. and time physical state vector The state vector must contain at least density. hydration degree ,temperature and free water content One or more parameters in the formula.

[0081] density It reflects the degree of compaction and porosity inside the concrete, and has a direct impact on its mechanical properties.

[0082] hydration degree This indicates the progress of cement hydration reaction, which is closely related to the strength development of concrete.

[0083] temperature It is a key factor affecting hydration rate and thermal stress.

[0084] Free water content This directly affects the electrical properties and workability of concrete. The goal of the reconstruction is to obtain the three-dimensional spatial distribution of these parameters within the pouring area.

[0085] The reconstruction process relies on the establishment of multiphysics forward models, which describe the state field within a known internal environment. Under these conditions, the sensor array will obtain the theoretical boundary measurement values.

[0086] For the acoustic forward model, the propagation speed of sound waves in the concrete medium is... It is the internal state field The function, especially affected by density and hydration degree The acoustic forward model establishes a functional relationship between the sound velocity field and the state parameters, which can be expressed as:

[0087] ;

[0088] in, For the nonlinear mapping function. An ultrasonic transmitter and receiver pair, the propagation time of the ultrasonic wave from the transmitter to the receiver. The sound speed field along the propagation path The line integral of is calculated using the following formula:

[0089] ;

[0090] in, Representing a path The differential length element is used. By calculating the propagation time of all transmit and receive pairs, the predicted acoustic boundary data can be obtained. It is a collection of all The vector.

[0091] For the forward model of electrical conductivity, the electrical conductivity of the concrete medium Similarly, internal state field A function of free water content and temperature Particularly sensitive. The forward model of electricity establishes a functional relationship between the conductivity field and the state parameters, which can be expressed as:

[0092] ;

[0093] in, This is a nonlinear mapping function. Given the conductivity distribution... and boundary current excitation Under these conditions, the field Electric potential distribution within It is described by the following partial differential equation of steady-state current field:

[0094] ;

[0095] in, It is the gradient operator. It is a dot product operator. This represents the concrete pouring area. Under specific boundary conditions (e.g., the electrode boundary where the current is applied is a Neumann boundary condition, and the electrode boundary where the potential is measured is a Dirichlet boundary condition), this equation can be solved numerically (e.g., the finite element method) to obtain the potential distribution on the boundary electrodes, i.e., the predicted electrical boundary data. .

[0096] The reconstruction of the internal state field is achieved by solving a joint inversion optimization problem. This problem aims to find an internal state field. The three-dimensional distribution minimizes the difference between the boundary data predicted by the forward model and the actual boundary measurement data collected in step S1. The three-dimensional distribution is obtained by discretizing the concrete pouring area, for example, using a finite element mesh or voxel mesh, transforming the continuous internal state field into discrete node or element variables. The inversion objective is to solve for the state parameter values ​​at these discrete points. The density and resolution of the finite element mesh or voxel mesh are balanced according to the required reconstruction accuracy and computational resources; typically, the mesh element size can be 10 cm to 30 cm.

[0097] Joint inversion objective function Set as:

[0098] ;

[0099] in, For acoustic data fitting, used to quantize the predicted acoustic boundary data. Acoustic boundary measurement data actually collected The differences between them It is a positive definite weight matrix, which can be determined based on the covariance of acoustic measurement errors;

[0100] For electrical data fitting, it is used to quantify the electrical boundary data for prediction. Compared with the actual collected electrical boundary measurement data The differences between them It is a positive definite weight matrix, which can be determined based on the covariance of electrical measurement errors;

[0101] This is a regularization term used to introduce prior information about the solution to overcome the ill-posedness of the inversion problem and ensure the stability and spatial smoothness of the solution.

[0102] The regularization term can use Tikhonov regularization, for example... ,in This is a regularization parameter that controls the strength of regularization.

[0103] This nonlinear optimization problem can be solved using iterative algorithms, such as gradient-based optimization methods like the Gauss-Newton method or the conjugate gradient method, which converge at each time step to obtain the actual internal state field at the current moment. .

[0104] Besides Tikhonov regularization, the regularization term can also employ total variational (TV) regularization to allow for discontinuous regions, such as cracks or defects that are prone to occur inside concrete, while maintaining the piecewise smoothness of the solution. In each iteration, the iterative algorithm updates the internal state field based on the gradient of the objective function and / or Hessian matrix information. The estimated value is obtained until the preset convergence criterion is met (e.g., the change in the objective function value is less than a threshold or the gradient norm is less than a threshold).

[0105] In step S3, the method uses the concrete internal state field reconstructed in real time in step S2. Compared with the preset optimal state trajectory A comparison is made to obtain the state deviation. This step aims to quantify the gap between the current actual internal state of the concrete and the desired ideal state.

[0106] Optimal state trajectory It is an ideal state distribution in four dimensions (three spatial dimensions plus time dimension), which predetermines the internal state parameters (including density) of concrete from pouring to initial setting and subsequent curing. hydration degree ,temperature and free water content (etc.) in a specific spatial location and time The ideal evolution path. The optimal state trajectory is set by calculation and planning based on specific engineering requirements, concrete mix proportions, component design drawings, and expected environmental conditions (e.g., ambient temperature and humidity).

[0107] For example, for a specific construction stage, there might be an ideal temperature gradient to avoid excessive thermal stress; or a target hydration development curve to ensure sufficient early strength before formwork removal; and an ideal density distribution to guarantee structural load-bearing capacity. These ideal parameter values, varying with time and space, collectively constitute the optimal state trajectory.

[0108] The calculation and planning of the optimal state trajectory can be performed offline based on a detailed simulation model of concrete hydration and coupled thermodynamics and mechanics to obtain an ideal parameter evolution path that meets structural performance and construction requirements. This planning result is stored as reference data for real-time comparison.

[0109] The comparison process at each discrete control time step Proceed. At time step The method will reconstruct the internal state field from the real-time data obtained in step S2. The optimal state trajectory at the same time step and spatial location as preset Perform point-by-point (or region-wide) comparisons. Through these comparisons, calculate the deviation of each state parameter at each spatial location. For example, for any discrete spatial point... In time The state deviation can be expressed as:

[0110] ;

[0111] in, It is the state deviation vector, density deviation Hydration deviation Temperature deviation Deviation in free water content .

[0112] These deviation data constitute a deviation field, which fully reflects the degree and direction of the current internal state of the concrete deviating from the ideal trajectory in space and time. The state deviation will serve as an important input to the model predictive control algorithm in the subsequent step S4, guiding the generation of control commands.

[0113] In step S4, based on the state deviation obtained in step S3, the method generates control instructions for adjusting the construction parameters of the moving formwork using a model predictive control (MPC) algorithm, and executes the control instructions. This step aims to proactively optimize the control sequence so that the internal state field of the concrete approaches the preset optimal state trajectory as closely as possible in the future time domain.

[0114] The core of the model predictive control algorithm lies in using a discrete-time prediction model to simulate the future evolution of the internal state field of concrete. This discrete-time prediction model... It couples the hydration kinetics and thermodynamics of concrete, in which... Indicates the current control cycle Predicting future moments based on the obtained internal state field information The internal state field of the concrete. Indicates at time The applied control input vector, It is a nonlinear state transition function.

[0115] This discrete-time prediction model is based on the internal state field at the current moment. and a series of hypothetical future control inputs Predicting the internal state field of concrete in the future The evolutionary path at each time step. The length of the prediction time domain is the number of future steps the model predicts forward.

[0116] Discrete-time prediction model The mass and energy conservation equations and corresponding constitutive relations in the concrete hydration process are discretized using numerical methods (such as the finite difference method or the finite element method). The model's inputs include the initial state of the concrete, environmental boundary conditions, and current control commands, while the output is the internal state field at the next time step.

[0117] The accuracy of the prediction model is crucial to the performance of the MPC algorithm; therefore, the model parameters can be calibrated and validated through offline experiments or historical data.

[0118] In each control cycle Within this framework, the Model Predictive Control (MMCC) algorithm solves a finite-time optimization problem to determine the optimal sequence of control commands. The objective function of the optimization problem is... The aim is to minimize the deviation between the internal state field and the optimal state trajectory in the prediction time domain, while limiting the amplitude or rate of change of the control command. Its mathematical expression is as follows:

[0119] ;

[0120] This optimization problem is subject to the following constraints:

[0121] ,for ;

[0122] ;

[0123] ,for ;

[0124] in, It is the sequence of control commands to be optimized in the prediction time domain; The current moment; It is a positive definite weight matrix used to measure the penalty intensity for deviations in the internal state field; It is a positive definite weight matrix used to penalize the magnitude of control commands in order to avoid overly aggressive control operations; For prediction in the time domain; It controls the length of the time domain, usually ; and These are the lower and upper limits of the control commands, respectively, to ensure that the control commands are within the physically feasible range; Is Time prediction The system state at any given moment; It is a pre-set future moment. The optimal state trajectory.

[0125] This optimization problem can be solved using numerical optimization algorithms (e.g., interior-point method, sequential quadratic programming) to obtain an optimal control command sequence, denoted as . .

[0126] Numerical optimization algorithms are typically executed repeatedly within each control cycle to adapt to real-time changes in the internal state and external environment. During the solution process, constraints imposed on the amplitude and rate of change of control commands (e.g., maximum vibration power, maximum heating / cooling rate, minimum / maximum travel speed of the formwork) ensure the physical feasibility and safety of the commands.

[0127] Control commands used to adjust the construction parameters of the moving formwork are usually in vector form. This indicates that its components may include the power of the vibratory system. Heat flux of the mold frame temperature control system and the speed of movement of the moving mold At least one of them. For example, control command vector. It can be represented as:

[0128] ;

[0129] in, For at any time The power applied by the vibration system affects the density of the concrete; For at any time The heat flow rate provided by the formwork temperature control system (which can be positive to indicate heating and negative to indicate cooling) is used to regulate the temperature evolution of the concrete. For at any time The moving speed of the formwork directly affects the residence time of concrete within the formwork, thus affecting the degree of hydration and density.

[0130] In each control cycle Internally, after solving the optimization problem, the model predictive control algorithm obtains the future... The optimal control sequence at each time point However, the method only executes the first control instruction in the sequence. .

[0131] After execution is complete, the system enters the next control cycle. In the new control cycle, the method will re-acquire boundary measurement data and reconstruct an updated real-time internal state field. Then, the process of setting up, solving, and executing the optimization problem is repeated. This rolling optimization strategy can effectively cope with unknown disturbances and model uncertainties that may occur in actual construction, and improve the adaptability and robustness of the control system.

Claims

1. A method for dynamic monitoring and adaptive control of a mobile formwork casting process, characterized in that, The method comprises the following steps: S1, synchronously collecting boundary measurement data of an acoustic sensor array and an electrical sensor array arranged on a mobile formwork; S2, based on the boundary measurement data, reconstructing the internal state field of the concrete in real time by solving a joint inversion objective function, the internal state field including a three-dimensional distribution of at least one state parameter of density, hydration degree, temperature and free water content; wherein the joint inversion objective function is set as: ； wherein is an internal state field to be solved; is acoustic boundary data predicted based on the internal state field; is actually acquired acoustic boundary measurement data; is a weight matrix for the acoustic data fitting terms; is electrical boundary data predicted based on the internal state field; is actually acquired electrical boundary measurement data; is a weight matrix for the electrical data fitting terms; is a regularization term for introducing a priori constraints; is an acoustic data fitting term; is an electrical data fitting term; The regularization term adopts a Tikhonov regularization, and an expression thereof is: ； wherein, is a regularization parameter; is a concrete placement area; is a gradient operator; is an internal state field at a spatial location ; and is a spatial location; The joint inversion objective function is solved by an iterative algorithm, which comprises a gradient-based optimization method, converging at each time step to the actual internal state field at the current time ; S3, comparing the real-time reconstructed internal state field with a preset optimal state trajectory to obtain a state deviation; S4, generating a control instruction for adjusting a mobile formwork construction parameter based on the state deviation through a model predictive control algorithm, and executing the control instruction.

2. The method of claim 1, wherein, In step S2, the joint inversion objective function is composed of an acoustic data fitting term, an electrical data fitting term, and a regularization term: The acoustic data fitting term is used to quantify the difference between the predicted acoustic boundary data and the actually acquired acoustic boundary measurement data ; The electrical data fitting term is used to quantify the difference between the predicted electrical boundary data and the actually acquired electrical boundary measurement data ; The regularization term is used to provide a priori constraint for solving the joint inversion objective function.

3. The method of claim 2, wherein the method further comprises: The predicted acoustic boundary data is a sound wave propagation time calculated through an acoustic forward model, and the predicted electrical boundary data is a boundary potential distribution calculated through an electrical forward model.

4. The method of claim 1, wherein, In step S4, in each control period, the model predictive control algorithm generates an optimal control instruction by solving an optimization problem with a future prediction time domain as a range, and a solving target of the optimization problem is to make a future evolution process of the internal state field close to the optimal state trajectory while constraining an amplitude or a change rate of the control instruction.

5. The method of claim 4, wherein, To solve the optimization problem, a discrete-time state transition model based on a mechanism of concrete hydration dynamics and thermodynamics is adopted to predict the future evolution process of the internal state field according to a current internal state field and an input control instruction, and the predicted future evolution process is used for solving calculation of the optimization problem.

6. The method of claim 1, wherein, In step S4, the control instruction for adjusting the mobile formwork construction parameter comprises an adjustment instruction for at least one of the following parameters: vibration system power, formwork temperature control system heat flow rate, and mobile formwork traveling speed.

7. The method of claim 4, wherein the method further comprises: In each control period, a solving result of the optimization problem is an optimal control sequence composed of the control instructions at multiple future time points, and only a first control instruction in the optimal control sequence is executed, and in the next control period, a solving process of the optimization problem is repeated based on a new real-time reconstructed internal state field.

8. The method of claim 1, wherein the method further comprises: In step S3, the optimal state trajectory is a path of an ideal trajectory describing evolutions of the state parameters in the internal state field over time, which is determined in advance according to concrete proportioning, component design, and environmental conditions.

9. The method of claim 3, wherein the method further comprises: The acoustic forward model establishes a functional relationship between a sound velocity field and the state parameters in the internal state field based on the state parameters, and calculates the sound wave propagation time according to the sound velocity field.

10. The method of claim 3, wherein the method further comprises: The electrical forward model establishes a functional relationship between an electrical conductivity field and the state parameters in the internal state field based on the state parameters, and solves a potential boundary value problem according to the electrical conductivity field to obtain the boundary potential distribution.

Images