Modularized prefabricated hot water circulation warehouse surface heat preservation and maintenance system

Abstract

The invention discloses a modular prefabricated hot water circulation warehouse surface heat preservation and maintenance system, and particularly relates to the technical field of building construction and maintenance, regional modular processing is performed on a concrete structure, and a historical temperature sequence, environmental parameters and concrete maturity data are analyzed in real time in combination with an LSTM neural network prediction model and a reinforcement learning algorithm; predicting a future temperature trend, dynamically generating a regulation and control instruction through an edge computing gateway, preferentially heating a low-temperature clustering module group by adopting a greedy scheduling algorithm when the environment temperature is lower than 0 DEG C, adaptively adjusting a heating threshold value of an adjacent module 5 in combination with a fuzzy PI D controller, and preferentially heating a maximum temperature difference region through the greedy scheduling algorithm in a seam region; when the environment temperature is higher than 0 DEG C, full-area synchronous heating is adopted, and the joint secondary heating efficiency is optimized; hot water flow speed is adjusted in a stepped mode to meet temperature difference control standards.

Description

Technical Field

[0001] This invention relates to the field of building construction and maintenance technology, specifically to a modular prefabricated hot water circulation tank surface insulation and maintenance system. Background Technology

[0002] With the development of building construction and infrastructure development, winter curing technology for concrete has become an important means of ensuring the strength development of concrete under cold conditions, and is widely used in the construction of large hydraulic structures, bridges, and high-rise buildings. The main purpose of winter curing of concrete is to maintain the temperature of the concrete through heating or insulation measures, so as to avoid the slowing down or even stopping of the cement hydration reaction due to low temperature, thereby affecting the structural performance.

[0003] Currently, the main methods for winter curing of concrete include heat storage, steam heating, electric heating, and heated enclosures. Heat storage utilizes the heat generated from preheating raw materials and cement hydration, combined with insulation materials to extend the cooling time, allowing the concrete to reach its critical strength before freezing. Common measures include aggregate preheating, hot water mixing, insulation during transportation, and embedding foam boards in the formwork. Steam heating achieves curing by introducing steam into the external panels, plastic film covers, or internal channels of the formwork, and is suitable for various cast-in-place and precast components. Electric heating converts electrical energy into heat energy for localized heating by inserting electrodes or using resistance wires or electromagnetic induction heaters. Heated enclosures create a microclimate by constructing enclosed spaces and using stoves or electric heating equipment to provide suitable hardening conditions for large-volume concrete.

[0004] However, existing technologies have certain limitations. For example, the heat storage method is highly dependent on insulation materials and has limited effectiveness in extreme low-temperature environments; the steam heating method requires additional piping and has high energy consumption, while improper humidity control may cause cracking of the concrete surface; the electric heating method, although heating up rapidly, consumes a lot of power and poses safety hazards; although the greenhouse method can effectively improve the construction environment, the traditional tied-up greenhouse is complex to install and dismantle, and although the prefabricated greenhouse has improved, it still requires a lot of manpower and resources for support and lifting operations.

[0005] Therefore, this invention provides a modular prefabricated hot water circulation sump surface insulation and curing system, which aims to achieve efficient, energy-saving and safe concrete curing by integrating heating units, water pipes, insulation covering materials, temperature measuring units, data transmission units and remote control systems, thereby meeting the needs of modern building construction for high-quality and low-cost curing systems. Summary of the Invention

[0006] This invention aims to solve the problems of high heating energy consumption, uneven local temperature difference, and large control delay in the existing winter curing process of concrete. This invention provides a modular prefabricated hot water circulating slab surface insulation curing system. This system integrates technologies such as multi-source sensor data acquisition, machine learning prediction, fuzzy control, topological clustering and greedy scheduling to construct a data-driven, adaptive, and closed-loop intelligent curing control process, which significantly improves the energy efficiency and quality stability of the concrete construction process, thereby solving the problems mentioned in the background technology.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a modular prefabricated hot water circulation sump surface insulation and curing system, which divides the concrete structure sump surface into multiple regional modules for processing according to the geometric parameters of the target concrete structure sump surface, the system comprising:

[0008] The data acquisition module includes water pipes laid on the surface of each area module and temperature measuring units installed inside each area module. The operating power of the heating unit is adjusted by the remote control system to make hot water circulate in the water pipes to achieve uniform heating and heat preservation of the concrete. During the hot water circulation process, the temperature information collected by the temperature measuring units is transmitted to the remote control system through the data transmission unit, and the hot water flow rate and temperature are dynamically adjusted.

[0009] The input module combines the current maturity value of each regional module, historical temperature sequence, ambient temperature and humidity, hot water flow rate, and concrete maturity to form the input features of the LSTM model. The LSTM model is used to predict the temperature change trend in a specified future period in real time. Based on the dual constraints of the prediction curve and the preset maturity threshold, the remaining heating time of each regional module is calculated to achieve on-demand heating and precise shutdown. The data processing unit generates control commands, which are sent to the hot water valves and heating units of the corresponding regional modules through the signal transmission unit to dynamically adjust the hot water flow rate and heating power.

[0010] The distributed control module allows each regional module to independently execute intelligent temperature control algorithms. The generated control commands are sent to the field execution devices via the edge computing gateway, and real-time temperature measurement and adjustment data are uploaded to the remote control center via the communication unit. Combined with the global scheduling commands issued by the control center, the distributed collaborative optimization of each regional module is carried out. The distributed control module also determines whether to perform zoned cyclic heating or synchronous heating of the entire area based on the ambient temperature.

[0011] Preferably, the external insulation material covering the water pipe is a composite layer of polyurethane foam and waterproof membrane.

[0012] Preferably, each regional module is equipped with a fiber optic temperature sensor network to collect the surface and internal temperature of the regional module in real time. A maturity-driven LSTM predictive control algorithm model is introduced. Based on the prediction results of this model and the maturity growth trend, the predicted temperature value at each future time is iteratively calculated and the predicted maturity is updated accordingly to adjust the hot water flow rate and heating power, as follows:

[0013] First, collect the current concrete temperature T of each module area. i (t), hot water flow rate Q i (t), ambient temperature and humidity T env (t), RH(t), and the accumulated maturity value M i (t); maturity is calculated using the following formula:

[0014] M i (t+Δt)=M i (t)+max(T i (t)-T0,0)·Δt;

[0015] Where T0 is the reference temperature, Δt is the control time step in hours; if the current maturity M i (t) When the threshold corresponding to the concrete strength grade is reached (e.g., 400℃·h for C30 grade), the module automatically terminates the heating process; based on the prediction results and maturity growth trend, the system iteratively calculates the predicted temperature value for each future moment and updates the predicted maturity accordingly:

[0016]

[0017] when Find the minimum time step τ * This allows us to calculate the remaining heating time Δt required for the current module. rem =τ * • Δt; The system is based on the remaining heating time Δt rem Calculate the current hot water flow rate or heating power control value for the current area module:

[0018] u i (t)=μ i (t)·T base ;

[0019] μ i (t) = clip(u min ,u max ,K p (Δt rem -Δt0)) ψ ;

[0020] Where u i(t) represents the control command (such as the opening degree of the hot water valve, the pump speed, or the auxiliary heating power), μ i (t) is the adjustment coefficient, T base The baseline heating cycle (preset based on concrete grade and construction environment, e.g., 10 min, 15 min); u min ,u max : Upper and lower limits of the adjustment coefficient; K p : Proportional gain coefficient; Δt rem : Predicted remaining heating time; Δt0: Allowable switching lead; ψ: Nonlinear shaping exponent; clip(·): Upper and lower limit constraint function to prevent overheating or abnormal flow rate.

[0021] Preferably, when the ambient temperature is below 0℃, a zoned cyclic heating method is adopted. The real-time temperature distribution and prediction data are analyzed by a greedy scheduling algorithm, and the low-temperature clustering module group (the original interval module heating area 4) is heated to the preset temperature first. Then, the fuzzy PID controller is used to automatically adjust the heating overlap area and switching threshold of adjacent modules according to the temperature gradient between the modules.

[0022] Preferably, when the ambient temperature is above 0℃, a full-area synchronous heating method is adopted, that is, hot water circulation heating is carried out on each area module at the same time. After the temperature of all modules stabilizes and drops to the preset temperature, the insulation material is removed, and then the joint area between each area module is heated by hot water circulation in an intermittent manner.

[0023] Preferably, the fuzzy PID controller automatically adjusts the heating overlap area and switching threshold of adjacent modules based on the temperature gradient between regional modules. Specifically, the fuzzy PID controller dynamically analyzes the temperature gradient, rate of change, and historical error between adjacent regional modules. After the module heating task is completed, a greedy scheduling algorithm is introduced to address the problem of temperature difference accumulation in the joint areas between regional modules. All joint areas are sorted according to temperature difference priority, and local heating resources are allocated in sequence until the temperature of all regional modules is balanced.

[0024] Preferably, the fuzzy PID controller control process is as follows:

[0025] First, real-time temperature values ​​T of each module are collected using a high-precision fiber optic temperature sensor. i and its rate of change A weighted adjacency graph is constructed to provide a quantitative basis for the strength of thermal connections between modules in density clustering algorithms, enabling the clustering process to accurately identify low-temperature regions and heat conduction paths. The edge weights reflect the degree of thermal diffusion coupling between modules. Through density clustering algorithms (such as DBSCAN), regions with current temperatures below a set threshold (such as 10℃) and spatial proximity are identified to form a low-temperature cluster module group C. lowIf the average temperature drop rate of the low-temperature clustering module group exceeds 2℃ / h, the system will treat it as a high-priority heating target; simultaneously, the fuzzy PID controller receives the temperature difference ΔT = T between the modules in the receiving area. i -T i+1 Rate of temperature change With the error accumulation as input and referring to the preset fuzzy rule set, the water flow rate, start-up delay, and switching threshold adjustment value of the heating overlap area can be output.

[0026] The preferred greedy scheduling algorithm process is as follows:

[0027] First, record the average temperature of the module edges on both sides of each joint. Calculate the temperature difference at the seam A priority queue is established in descending order. Starting from the joint with the largest temperature difference, the system sequentially schedules local heating units to perform hot water circulation until... Alternatively, heat the equipment once it reaches the set upper limit time (e.g., 15 minutes), then update the temperature difference data and reorder it until the temperature difference in all seam areas meets the control standard. The specific steps are as follows:

[0028] Step 1. Temperature difference calculation in the seam area: The system constructs a set of seams S = {s1, s2, ..., s} for all module boundaries. n-1 The average temperature of the boundary area between adjacent modules is measured in real time using a high-precision fiber optic temperature sensor, and the temperature difference of each joint is calculated.

[0029] Step 2. Priority queue construction: ΔT i Arranged in descending order, forming a greedy scheduling queue Q = sorted(ΔT) i The system prioritizes the seam with the largest temperature difference and then performs heating sequentially.

[0030] Step 3. Localized Heating Strategy: For each joint area, the system controls the opening time and flow rate of the solenoid valve in the local water supply branch, sets a heating time window (e.g., 10-15 minutes), and continuously collects temperature data. If the temperature difference ΔT i If the temperature is ≤0.5℃, heating of the joint area should be terminated.

[0031] Step 4. Dynamic feedback update: After each round of heating, the temperature difference of all seams is recalculated and the priority queue is updated until the temperature difference of all seam areas meets the control standard. This method ensures that heat energy is preferentially supplied to the most unbalanced parts and avoids repeated heating, thereby improving heating efficiency.

[0032] The present invention has the following advantages:

[0033] This invention modularizes concrete structures by region. It combines an LSTM neural network prediction model with reinforcement learning algorithms to analyze historical temperature sequences, environmental parameters, and concrete maturity data in real time, predicting future temperature trends. An edge computing gateway dynamically generates control commands. When the ambient temperature is below 0°C, a greedy scheduling algorithm prioritizes heating low-temperature clustering modules, while a fuzzy PID controller adaptively adjusts the heating thresholds of adjacent modules. For joint areas, the greedy scheduling algorithm prioritizes heating the region with the largest temperature difference. When the ambient temperature is above 0°C, simultaneous heating of the entire region is employed, and the efficiency of secondary heating at joints is optimized. The hot water flow rate is adjusted in steps to meet temperature difference control standards. Compared with existing technologies, this invention can adaptively adjust the heating cycle of each regional module, avoiding overheating or lag, effectively reducing energy consumption and improving the uniformity of concrete curing. It is particularly suitable for complex construction environments with severe low-temperature fluctuations or large diurnal temperature differences. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the partitioned circulating heating method in the embodiment;

[0035] Figure 2 This is a diagram showing the connection relationships between the temperature measuring unit, remote control system, heating unit, data transmission unit, water pipes, and distributed heating unit in the embodiment.

[0036] Figure 3 The flowchart for the execution of the intelligent temperature control algorithm in the implementation plan.

[0037] In the diagram: 1. Concrete structure; 2. Module boundary; 3. Flexible water pipe; 4. Heating area of ​​the interval module; 5. Adjacent modules; 6. Joint area; 7. Independent solenoid valve. Detailed Implementation

[0038] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. 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.

[0039] like Figures 1-3As shown, this embodiment provides a modular prefabricated hot water circulation chamber surface insulation and curing system. Based on the geometric parameters of the target concrete structure 1 chamber surface, it is divided into multiple regional modules along the long side of the concrete structure 1 chamber surface. Each regional module has a length of not less than 3m and not more than 1 / 8 of the length of the long side of the concrete structure 1. Each regional module is equipped with an independent solenoid valve, a distributed heating unit, and a high-precision fiber optic temperature sensor network. The high-precision fiber optic temperature sensor network is laid out in a 5cm×5cm grid, with a sampling frequency increased to 1 time / minute. The high-precision fiber optic temperature sensor model is Micron Optics OS4100 series, with a measurement range of -20℃ to +120℃ (a wider range can be customized), accuracy of ±0.1℃, resolution of 0.01℃, wavelength range of 1510–1590nm, grating spacing customizable to meet the 5cm×5cm grid layout, and interface for use with a fiber optic demodulator (such as SI255), with a sampling frequency ≥1Hz, which can be reduced to 1 time / minute via software.

[0040] The system includes:

[0041] The data acquisition module includes flexible water pipes 3 laid on the surface of each area module (the water pipes are covered with insulation material; in this embodiment, a composite layer of polyurethane foam and waterproof membrane is used) and temperature measuring units installed inside each area module. The operating power of the heating unit is adjusted by a remote control system to circulate hot water within the water pipes, achieving uniform heating and insulation of the concrete structure 1. During the hot water circulation process, the temperature information collected by the temperature measuring units is transmitted to the remote control system via a data transmission unit. Figure 2 (The control box shown) dynamically adjusts the hot water flow rate and temperature;

[0042] The input module combines the current maturity value of each regional module, historical temperature sequence, ambient temperature and humidity, hot water flow rate, and concrete maturity to form the input features of the LSTM model. The LSTM model is used to predict the temperature change trend in a specified future period in real time. Based on the dual constraints of the prediction curve and the preset maturity threshold, the remaining heating time of each regional module is calculated to achieve on-demand heating and precise shutdown. The data processing unit generates control commands, which are sent to the hot water valves and heating units of the corresponding regional modules through the signal transmission unit to dynamically adjust the hot water flow rate and heating power.

[0043] A fiber optic temperature sensor network collects real-time data on the surface and internal temperatures of the regional modules. A maturity-driven LSTM predictive control algorithm is introduced. Based on the model's prediction results and maturity growth trends, the predicted temperature value for each future time point is iteratively calculated, and the predicted maturity is updated accordingly to adjust the hot water flow rate and heating power, as detailed below:

[0044] First, collect the current concrete temperature T of each module area.i (t), hot water flow rate Q i (t), ambient temperature and humidity T env (t), ambient humidity RH(t), and accumulated maturity value M i (t);

[0045] Among them, the collected hot water flow rate Q i (t) is used to reflect the instantaneous heating capacity of each module and serves as the input variable for the temperature prediction model and pump speed control; ambient temperature T env (t) is used to evaluate the heat exchange rate between concrete and the environment, and directly participates in the judgment in low temperature clustering scheduling and temperature difference compensation strategy; the ambient humidity RH(t) is used to correct the evaporative heat dissipation factor in the prediction model, and indirectly affects the calculation accuracy of temperature prediction and maturity growth.

[0046] Maturity is calculated using the following formula:

[0047] M i (t+Δt)=M i (t)+max(T i (t)-T0,0)·Δt;

[0048] Where T0 is the reference temperature; T i This refers to the current temperature value, not the complete historical temperature series; Δt is the control time step in hours; if the current maturity M... i (t) If the threshold corresponding to the concrete strength grade is reached (400℃·h for C30 grade), the module automatically terminates the heating process; based on the prediction results and maturity growth trend, the system iteratively calculates the predicted temperature value for each future moment and updates the predicted maturity accordingly:

[0049]

[0050] when Find the minimum time step τ * This allows us to calculate the remaining heating time Δt required for the current module. rem =τ * • Δt; The system is based on the remaining heating time Δt rem Calculate the current hot water flow rate or heating power control value for the current area module:

[0051] u i (t)=μ i (t)·T base ;

[0052] μ i (t) = clip(u min ,u max ,K p(Δt rem -Δt0)) ψ ;

[0053] Where u i (t) represents the control command (such as the opening degree of the hot water valve, the pump speed, or the auxiliary heating power), μ i (t) is the adjustment coefficient, T base The baseline heating cycle (preset based on concrete grade and construction environment, e.g., 10 min, 15 min); u min ,u max : Upper and lower limits of the adjustment coefficient; K p : Proportional gain coefficient; Δt rem : Predicted remaining heating time; Δt0: Allowable switching lead time; ψ: Nonlinear shaping exponent; clip(·): Upper and lower limit constraint function to prevent overheating or abnormal flow rate; Enables millisecond-level independent control of heating power, hot water flow rate, and temperature at the module level, predicts temperature change trends for the next 1-3 hours (error ≤ ±1.5℃), and the system completes data preprocessing and model inference locally through an edge computing gateway to generate control commands (flow rate ±0.2m). 3 The system operates at a speed of ±5℃, ensuring a response delay of <500ms. When the concrete surface temperature of a module falls below a preset value, the system automatically increases the power of the heating unit and the hot water flow rate, and can also trigger the distributed heating unit to provide auxiliary electric heating.

[0054] The distributed control module allows each regional module to independently execute intelligent temperature control algorithms, including fuzzy PID controllers, to dynamically adjust heating power and switching timing during execution. The generated control commands are sent to the field execution devices via an edge computing gateway, and real-time temperature measurement and adjustment data are uploaded to the remote control center via a communication unit. Combined with global scheduling commands issued by the control center, distributed collaborative optimization is performed on each regional module. The distributed control module also determines whether to perform zoned cyclic heating or synchronous heating across the entire area based on the ambient temperature.

[0055] When the ambient temperature is below 0℃, a zoned circulating heating method is adopted. A greedy scheduling algorithm analyzes real-time temperature distribution and prediction data, prioritizing the heating of the low-temperature clustering module group (originally interval module 4) to the preset temperature. When the predicted temperature drop rate of a certain module group is greater than 2℃ / h, the system automatically executes the "flow rate increase to 1.8m" command. 3 The system executes the command " / min+start auxiliary electric heating". Then, the fuzzy PID controller automatically adjusts the heating overlap area and switching threshold of adjacent modules 5 according to the temperature gradient between the modules. When the temperature difference between adjacent modules 5 is greater than 15℃, the system will extend the heating time of the current module by 10 minutes and reduce the initial flow rate of the next module by 20%.

[0056] When the ambient temperature is above 0℃, a full-area synchronous heating method is adopted, that is, hot water circulation heating is applied to all modules simultaneously. Once the temperature of all modules stabilizes and drops to the preset temperature, the insulation material is removed. Then, a second round of hot water circulation heating is applied to the joint areas 6 between the modules, following an interval-sealed approach. At a certain construction site, when the ambient temperature was 5℃, all modules were first heated simultaneously to 25℃. After the temperature stabilized, the insulation material was removed, and the joint areas 6 were heated a second time. This method significantly improves heating efficiency and reduces construction time.

[0057] To further optimize system performance during the hot water circulation process, the remote control system also integrates a fuzzy PID controller, which dynamically adjusts parameters based on temperature difference, temperature gradient, and predicted data. For example, the fuzzy set of temperature difference is defined as {small, medium, large}, and the fuzzy set of temperature is defined as {low, medium, high}. The hot water flow rate and heating power are automatically adjusted through fuzzy inference rules. This algorithm can effectively cope with complex changes in the on-site environment and improve the system's adaptability.

[0058] The fuzzy PID controller automatically adjusts the heating overlap area and switching threshold of adjacent modules 5 according to the temperature gradient between regional modules. Specifically, the fuzzy PID controller dynamically analyzes the temperature gradient, rate of change and historical error of adjacent modules 5. After the module heating task is completed, in view of the problem that temperature difference accumulation is easy to occur in the seam area 6 between the modules, a greedy scheduling algorithm is introduced to sort all seam areas 6 according to the priority of temperature difference and allocate local heating resources in sequence until the temperature of all regional modules is balanced.

[0059] The control process of the fuzzy PID controller is as follows:

[0060] First, real-time temperature values ​​T of each module are collected using a high-precision fiber optic temperature sensor. i and its rate of change A weighted adjacency graph is constructed to provide a quantitative basis for the strength of thermal connections between modules in density clustering algorithms, enabling the clustering process to accurately identify low-temperature regions and heat conduction paths. The edge weights reflect the degree of thermal diffusion coupling between modules. Through density clustering algorithms (such as DBSCAN), regions with current temperatures below a set threshold (such as 10℃) and spatial proximity are identified to form a low-temperature cluster module group C. low If the average temperature drop rate of the low-temperature clustering module group exceeds 2℃ / h, the system will treat it as a high-priority heating target; simultaneously, the fuzzy PID controller receives the temperature difference ΔT = T between the modules in the receiving area. i -T i+1 Rate of temperature change With the error accumulation as input and referring to the preset fuzzy rule set, the water flow rate, start-up delay, and switching threshold adjustment value of the heating overlap area can be output.

[0061] The greedy scheduling algorithm process is as follows:

[0062] First, record the average temperature at the edges of the modules on both sides of each seam (i.e., seam area 6). Calculate the temperature difference at the seam A priority queue is established in descending order. Starting from the joint with the largest temperature difference, the system sequentially schedules local heating units to perform hot water circulation until... Alternatively, heat the material once it reaches the set upper limit time (e.g., 15 minutes), then update the temperature difference data and reorder it until the temperature difference in all seam areas meets the control standard. The specific steps are as follows:

[0063] Step 1. Temperature difference calculation in the seam area 6: The system constructs a seam set S = {s1, s2, ..., s} for all module boundaries. n-1 The average temperature of adjacent modules 5 is measured in real time using a high-precision fiber optic temperature sensor, and the temperature difference of each joint area 6 is calculated.

[0064] Step 2. Priority queue construction: ΔT i Arranged in descending order, forming a greedy scheduling queue Q = sorted(ΔT) i The system prioritizes the seam with the largest temperature difference and then performs heating sequentially.

[0065] Step 3. Local heating strategy: For each joint area 6, the system controls the opening time and flow rate of the solenoid valve of the local water pipe branch, sets a heating time window (e.g., 10-15 minutes), and continuously collects temperature data. If the temperature difference ΔT i If the temperature is ≤0.5℃, heating of the joint area 6 will be terminated;

[0066] Step 4. Dynamic feedback update: After each round of heating, the temperature difference of all seam areas 6 is recalculated and the priority queue is updated until the temperature difference of all seam areas 6 meets the control standard. This method ensures that heat energy is preferentially supplied to the most unbalanced parts and avoids repeated heating, thereby improving heating efficiency.

[0067] X1: Control group 1 (Baseline-LSTM): No maturity level specified;

[0068] X (1) ={T i (tk:t),Q i (t),T env (t),RH(t)}

[0069] Experimental group 2 (LSTM+Maturity): Maturity was added;

[0070] X (2) ={T i(tk:t),Q i (t),T env (t),RH(t),M i (t)}

[0071] Optional Group 3 (LSTM+Maturity+Interaction): Adds M to Group 2. i (t)×Q i (t), M i (t)×T env (t) Interaction characteristics (examining the effects of coupling).

[0072] Optional Group 4 (Uncertainty Output): Based on Group 2, use quantile loss / MC Dropout to output... The upper and lower quantiles (serving risk-sensitive fuzzy PID).

[0073] Fairness constraints: Except for different input features, the network structure, training strategy, and closed-loop control parameters (fuzzy PID, dynamic topology threshold, greedy seam strategy) remain consistent.

[0074] X2: Model and training configuration (unified), two-layer LSTM (hidden=64), fully connected readout layer; Adam (lr=1e-3), MSE loss; early stopping (patience=10); normalization by module / channel; 5-fold time block cross-validation.

[0075] The algorithm of this invention can adaptively adjust the heating cycle of each regional module to avoid overheating or lag, effectively reduce energy consumption and improve the curing uniformity of concrete structures, and is particularly suitable for complex construction environments with drastic low temperature fluctuations or large day-night temperature differences.

[0076] In practical applications (large-scale bridge projects), the modular prefabricated hot water circulating chamber surface insulation and curing system of this invention successfully solved the curing problems during winter construction, significantly improving the concrete strength development speed and reducing energy consumption. Furthermore, in the construction of a high-rise building, the heating path was optimized through a greedy scheduling algorithm, and combined with an edge computing gateway to achieve millisecond-level response, greatly shortening the construction cycle and improving project quality.

[0077] Although the present invention has been described in detail above with general descriptions and specific embodiments, modifications or improvements can be made to it, which will be obvious to those skilled in the art. Therefore, all such modifications or improvements made without departing from the spirit of the present invention fall within the scope of protection claimed by the present invention.

Claims

1. A modular prefabricated hot water circulation storage tank surface insulation and curing system, which divides the concrete structure storage tank surface into multiple regional modules for processing according to the geometric parameters of the target concrete structure storage tank surface, characterized in that: The system includes: The data acquisition module includes water pipes laid on the surface of each area module and temperature measuring units installed inside each area module. The operating power of the heating unit is adjusted by the remote control system to make the hot water circulate in the water pipes. During the hot water circulation process, the temperature information collected by the temperature measuring units is transmitted to the remote control system through the data transmission unit, and the hot water flow rate and temperature are dynamically adjusted. The input module combines the current maturity value of each regional module, historical temperature sequence, ambient temperature and humidity, hot water flow rate, and concrete maturity to form the input features of the LSTM model. The LSTM model is used to predict the temperature change trend in a specified future period in real time. Based on the dual constraints of the prediction curve and the preset maturity threshold, the remaining heating time of each regional module is calculated. The data processing unit generates control commands, which are sent to the hot water valves and heating units of the corresponding regional modules through the signal transmission unit to dynamically adjust the hot water flow rate and heating power. The distributed control module allows each regional module to independently execute intelligent temperature control algorithms. The generated control commands are sent to the field execution devices via the edge computing gateway, and real-time temperature measurement and adjustment data are uploaded to the remote control center via the communication unit. Combined with the global scheduling commands issued by the control center, the distributed collaborative optimization of each regional module is carried out. The distributed control module also determines whether to perform zoned cyclic heating or synchronous heating of the entire area based on the ambient temperature.

2. The modular prefabricated hot water circulation tank surface insulation and curing system according to claim 1, characterized in that: The water pipes are covered with thermal insulation material.

3. The modular prefabricated hot water circulation tank surface insulation and curing system according to claim 1, characterized in that: Each regional module is equipped with a fiber optic temperature sensor network to collect real-time surface and internal temperatures. A maturity-driven LSTM predictive control algorithm model is introduced. Based on the model's prediction results and maturity growth trends, the predicted temperature value for each future time point is iteratively calculated, and the predicted maturity is updated accordingly to adjust the hot water flow rate and heating power, as detailed below: First, collect the current concrete temperature T of each module area. i (t), hot water flow velocity Q i (t), ambient temperature and humidity T env (t), RH(t), and the accumulated maturity value M i (t); maturity is calculated using the following formula: M i (t+Δt)=M i (t)+max(T i (t)-T0,0)·Δt; Where T0 is the reference temperature, Δt is the control time step in hours; if the current maturity M i (t) If the threshold corresponding to the concrete strength grade is reached, the module automatically terminates the heating process; based on the prediction results and maturity growth trend, the system iteratively calculates the predicted temperature value for each future moment and updates the predicted maturity accordingly: when Find the minimum time step τ * This allows us to calculate the remaining heating time Δt required for the current module. rem =τ * • Δt; The system is based on the remaining heating time Δt rem Calculate the current hot water flow rate or heating power control value for the current area module: u i (t)=μ i (t)·T base ; μ i (t)=clip(u min ,u max ,K p (Δt rem -Δt0)) ψ ; Where u i (t) represents the control command (such as the opening degree of the hot water valve, the pump speed, or the auxiliary heating power), μ i (t) is the adjustment coefficient, T base As the reference heating cycle; u min ,u max : Upper and lower limits of the adjustment coefficient; K p : Proportional gain coefficient; Δt rem : Predicted remaining heating time; Δt0: Allowable switching lead; ψ: Nonlinear shaping exponent; clip(·): Upper and lower limit constraint function to prevent overheating or abnormal flow rate.

4. The modular prefabricated hot water circulation tank surface insulation and curing system according to claim 1, characterized in that: When the ambient temperature is below 0℃, a zoned cyclic heating method is adopted. The real-time temperature distribution and prediction data are analyzed by a greedy scheduling algorithm. The low-temperature clustering module group is heated to the preset temperature first. Then, the fuzzy PID controller is used to automatically adjust the heating overlap area and switching threshold of adjacent modules according to the temperature gradient between regional modules.

5. The modular prefabricated hot water circulation tank surface insulation and curing system according to claim 2, characterized in that: When the ambient temperature is above 0℃, a full-area synchronous heating method is adopted, that is, hot water circulation heating is carried out on each area module at the same time. After the temperature of all modules stabilizes and drops to the preset temperature, the insulation material is removed, and then the joint area between each area module is heated by hot water circulation in an intermittent manner.

6. The modular prefabricated hot water circulation tank surface insulation and curing system according to claim 4, characterized in that: The fuzzy PID controller automatically adjusts the heating overlap area and switching threshold of adjacent modules based on the temperature gradient between regional modules. Specifically, the fuzzy PID controller dynamically analyzes the temperature gradient, rate of change, and historical error between adjacent modules. After the module heating task is completed, a greedy scheduling algorithm is introduced to address the problem of temperature difference accumulation in the joint areas between modules. All joint areas are sorted according to temperature difference priority, and local heating resources are allocated in sequence until the temperature of all regional modules is balanced.

7. The modular prefabricated hot water circulation tank surface insulation and curing system according to claim 6, characterized in that: The control process of the fuzzy PID controller is as follows: First, real-time temperature values ​​T of each module are collected using a high-precision fiber optic temperature sensor. i and its rate of change A weighted adjacency graph is constructed, where the edge weights reflect the degree of thermal diffusion coupling between modules. A density clustering algorithm is used to identify spatially adjacent regions whose current temperature is below a set threshold, forming a low-temperature cluster module group C. low If the average temperature drop rate of the low-temperature clustering module group exceeds 2℃ / h, the system will treat it as a high-priority heating target; simultaneously, the fuzzy PID controller receives the temperature difference ΔT = T between the modules in the receiving area. i -T i+1 Rate of temperature change With the error accumulation as input and referring to the preset fuzzy rule set, the water flow rate, start-up delay, and switching threshold adjustment value of the heating overlap area can be output.

8. The modular prefabricated hot water circulation tank surface insulation and curing system according to claim 6, characterized in that: The greedy scheduling algorithm process is as follows: First, record the average temperature of the module edges on both sides of each joint. Calculate the temperature difference in the joint area A priority queue is established in descending order. Starting from the joint with the largest temperature difference, the system sequentially schedules local heating units to perform hot water circulation until... Alternatively, heating can be performed once the set upper limit time is reached, followed by updating the temperature difference data and reordering it until the temperature difference in all seam areas meets the control standard. The specific steps are as follows: Step 1. Temperature difference calculation in the seam area: The system constructs a set of seams S = {s1, s2, ..., s} for all module boundaries. n-1 The average temperature of adjacent modules is measured in real time using a high-precision fiber optic temperature sensor, and the temperature difference in each joint area is calculated. Step 2. Priority queue construction: ΔT i Arranged in descending order, forming a greedy scheduling queue Q = sorted(ΔT) i The system prioritizes the seam with the largest temperature difference and then performs heating sequentially. Step 3. Localized Heating Strategy: For each joint area, the system controls the opening time and flow rate of the solenoid valve in the local water supply branch, sets a heating time window, and continuously collects temperature data. If the temperature difference ΔT i If the temperature is ≤0.5℃, heating of the joint area should be terminated; Step 4. Dynamic feedback update: After each round of heating, recalculate the temperature difference of all seam areas and update the priority queue until the temperature difference of all seam areas meets the control standard.

Images