Thermal insulation layer construction precision control method based on multi-sensor fusion

Abstract

The invention discloses a thermal insulation layer construction precision control method based on multi-sensor fusion, and belongs to the technical field of construction precision control. According to the method, physical field data are synchronously collected through multi-source heterogeneous sensors such as high-frequency micro-seismic sensors, infrared thermal imaging sensors and millimeter-wave radars, full-chain prediction and regulation from the construction process to the future performance are achieved in combination with a data fusion and reverse optimization algorithm, the construction quality defect recognition precision is improved to the centimeter level, and the construction quality defect recognition accuracy is improved to the centimeter level. And through dynamic risk assessment and self-adaptive compensation construction, the defect repair response time is shortened to the second level, finally, the service life of the thermal insulation layer is prolonged, and the maintenance of the whole life cycle of the building is reduced.

Description

Technical Field

[0001] This invention belongs to the field of construction precision control technology, specifically referring to a method for controlling the construction precision of thermal insulation layers based on multi-sensor fusion. Background Technology

[0002] In building energy conservation projects, the construction quality of the insulation layer directly affects the building's energy efficiency and service life. Traditional construction relies on manual experience or post-construction inspection, making it difficult to detect hidden defects such as insufficient density, excessive moisture content, and uneven bonding in real time, leading to frequent problems such as hollowing and detachment.

[0003] However, existing multi-sensor fusion-based methods for controlling the construction accuracy of thermal insulation layers still have certain shortcomings. Existing technologies only focus on quality inspection during the construction phase and have not established a correlation model from construction quality to long-term performance. They cannot simulate the performance degradation trajectory of thermal insulation layers under complex service environments, making it impossible for construction personnel to predict high-risk areas in advance and adjust process parameters accordingly. Ultimately, the accumulation of defects leads to a surge in later maintenance costs. Fixed thresholds are used for risk level classification without dynamically adjusting the thresholds in conjunction with real-time environmental parameters. Under special climatic conditions such as high humidity or extreme cold, fixed thresholds cannot accurately reflect actual risks, leading to misjudgments or omissions. After defects are discovered, process parameters are usually adjusted manually through trial and error. The repair process lacks quantitative basis, is inefficient, and has uncertain results. Therefore, a multi-sensor fusion-based method for controlling the construction accuracy of thermal insulation layers is proposed. Summary of the Invention

[0004] The purpose of this invention is to provide a method for controlling the construction accuracy of thermal insulation layers based on multi-sensor fusion, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for controlling the construction accuracy of thermal insulation layers based on multi-sensor fusion, comprising the following steps: S1. Collect physical field data during construction by synchronously acquiring data through a multi-source heterogeneous sensor array; S2. Based on the collected physical field data of the construction process, perform data fusion and calculation to obtain implicit construction quality indicators that reflect the internal state of the insulation layer. S3. Input the calculated implicit indicators together with the real-time environmental parameters to predict the future performance degradation of the insulation layer; S4. Based on the performance prediction results, conduct a dynamic assessment of the defect risk level of the current construction unit area; S5. Based on the assessed risk level, perform reverse calculations to optimize the set of construction process parameters that can eliminate the risk; S6. Send the optimized set of construction process parameters to the actuator for adaptive real-time control and compensation of construction parameters. S7. Based on the adjusted construction effect, data is collected again through the sensor array and verified and optimized.

[0006] Preferably, in step S1, a high-frequency micro-vibration sensor, a long-wave infrared thermal imager, a microwave millimeter-wave radar, and a hyperspectral imager sensor are configured. The sensors are installed on a unified platform of the intelligent construction equipment. Each sensor is equipped with a dedicated signal conditioning circuit, including signal amplification, filtering, and level conversion. The optimal acquisition frequency is determined based on the data characteristics of each sensor. The sensor data acquisition is synchronized with the motion control system of the construction equipment to trigger the construction actions synchronously. According to the grid division of the construction area, sensor data acquisition is initiated in each construction unit area.

[0007] Preferably, in step S2, the physical field data collected during the construction process is cleaned in real time, the hollow vibration frequency in the microseismic field is extracted from the microseismic spectrum, and the interface density index is calculated based on the vibration-thermal coupling physical model, thus achieving the following: , In the formula, D represents the density index. Represents the vibration frequency coefficient. This represents the peak frequency related to the hollowness in the microseismic spectrum. Represents the temperature gradient coefficient. It represents the standard deviation of the infrared temperature field.

[0008] Preferably, in step S2, the spatial standard deviation of the dielectric constant is extracted from millimeter-wave radar data, and the moisture content and bonding uniformity are calculated based on the dielectric-chemical coupling physical model, which is implemented as follows: , In the formula, W represents the moisture content. Represents the dielectric constant coefficient. This represents the standard deviation of the dielectric constant of millimeter-wave radar. Indicates density, moisture content conversion factor; The bonding uniformity is achieved as follows: , In the formula, U represents the uniformity of adhesion. Indicates the bond density coupling coefficient. This represents the standard deviation of the reflectance of hyperspectral adhesives. This represents the average reflectance of the hyperspectral adhesive. The continuous values ​​output by the fusion model are mapped to standardized indicators that are operable in the engineering, and indicator values ​​are generated independently for each construction unit area.

[0009] Preferably, in step S3, the output implicit construction quality indicators are precisely correlated with the collected real-time environmental parameters according to the construction unit area, and preprocessed. The preprocessed data is then used to predict future performance degradation, achieving the following: , In the formula, Indicates the future Annual performance degradation rate Indicates the weighting coefficient. Indicates ambient temperature. Indicates ambient humidity. Indicates wind speed. Indicates solar radiation intensity. This indicates the reference environmental parameters.

[0010] Preferably, in S3, based on the future The annual performance degradation rate, simulating the long-term performance changes of the insulation layer under real service conditions, is achieved as follows: , In the formula, This represents the percentage of performance degradation over service time t. This represents the decay rate constant.

[0011] Preferably, in S4, the performance degradation prediction data of each construction unit area is received in real time, and the data integrity is checked in real time: if the unit area is missing a prediction value or the degradation ratio exceeds the preset range, the system automatically triggers the re-prediction process of S3, and the collected real-time environmental parameters are combined with the climate characteristics of the current construction area. The environmental threshold is dynamically adjusted based on the correlation analysis between historical construction data and climate conditions; the predicted attenuation ratio of each unit area is compared step by step according to the dynamic threshold. By precisely linking risk levels with the coordinates of construction unit areas, a three-dimensional spatial heat map is dynamically generated.

[0012] Preferably, in step S5, the risk level and corresponding predicted attenuation ratio of the current construction unit area are received from step S4, and the target attenuation ratio required to eliminate the risk is determined in combination with the dynamic environmental threshold rule of step S4. Based on the technical specifications of construction equipment and historical construction database, the physical executable range is defined for each process parameter; the target attenuation ratio is input into the performance attenuation prediction model of S3 for reverse optimization process; The parameter combination that satisfies the objective and requires the fewest iterations is selected from the population, verified, and the optimization results are encapsulated into a structured instruction package.

[0013] Preferably, in step S6, the actuator receives the generated structured instruction packet through a real-time communication protocol. After receiving the instruction, the actuator immediately initiates a self-check of the equipment status: calibrating the position of the robotic arm and checking the working status of the sensors; when the actuator moves to the medium-risk construction unit area, the system automatically triggers a fine-tuning of the process parameters; when the actuator is positioned in a high-risk area, the system immediately executes a shutdown and repair process: continuously receiving real-time environmental parameters during the control process and dynamically optimizing the repair strategy.

[0014] Preferably, in step S7, after the compensation construction is completed, the system automatically triggers the precise re-acquisition of the sensor array. The sensor array starts synchronous acquisition within 10 seconds after the construction equipment completes the work in the unit area. The re-acquisitioned physical field data is input into the fusion calculation process of step S2 to generate a new implicit index after adjustment. This index is compared with the risk level threshold of step S4 in real time to dynamically determine whether the risk level has dropped to low risk. If the risk level does not meet the standard, the system automatically marks the area as needing secondary repair and triggers the repeated adjustment process of step S6 until the risk level stabilizes in the low risk range.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention solves the core problems of imperceptible hidden quality defects, poor environmental adaptability, and delayed defect repair in the construction of thermal insulation layers by using multi-sensor fusion and closed-loop control technology. It significantly improves construction accuracy and long-term service performance. By synchronously collecting physical field data from multiple heterogeneous sensors such as high-frequency micro-vibration, infrared thermal imaging, and millimeter-wave radar, and combining data fusion and reverse optimization algorithms, it realizes full-chain prediction and control from the construction process to future performance. It not only improves the identification accuracy of construction quality defects to the centimeter level, but also shortens the defect repair response time to the second level through dynamic risk assessment and adaptive compensation construction. Ultimately, it extends the service life of the thermal insulation layer and reduces the maintenance cost of the building throughout its entire life cycle. 2. This invention uses a pre-trained digital twin model to input density, moisture content, bonding uniformity, and real-time environmental parameters to simulate the future performance degradation trajectory. Furthermore, it introduces an exponential degradation model to simulate the performance degradation ratio over the service life, achieving full-cycle prediction from construction completion to the end of service. This allows construction personnel to predict high-risk areas in advance, optimize process parameters accordingly, avoid the surge in maintenance costs caused by defect accumulation, and significantly improve the long-term service reliability of the insulation layer. 3. This invention dynamically adjusts the risk assessment criteria by predicting the attenuation ratio and combining it with real-time environmental parameters, making the assessment results more consistent with actual working conditions. Through real-time projection of a three-dimensional spatial heat map and an AR interface, construction personnel can intuitively locate defective areas and prioritize the treatment of high-risk units. The system automatically records the judgment logic to ensure the transparency and traceability of risk assessment, providing accurate input for subsequent reverse optimization and significantly improving the efficiency of construction decision-making. 4. This invention drives a digital twin model to perform reverse calculations based on risk level and target attenuation ratio, generating the optimal combination of process parameters such as spraying air pressure, discharge speed, and robotic arm trajectory. In high-risk areas, a genetic algorithm is used to quickly optimize parameters such as air pressure and discharge speed, reducing the predicted attenuation ratio to the target value. During the optimization process, the feasibility of parameters is strictly verified to ensure the executability of instructions. Finally, the instructions are packaged into a structured instruction package and sent to the actuator through an industrial communication protocol, realizing closed-loop control from risk assessment to process adjustment, significantly improving the accuracy and automation level of defect repair. Attached Figure Description

[0016] Figure 1 The following is the operation flow of the insulation layer construction accuracy control method based on multi-sensor fusion of the present invention. Figure 1 ; Figure 2 The following is the operation flow of the insulation layer construction accuracy control method based on multi-sensor fusion of the present invention. Figure 2 ; Figure 3 The following is the operation flow of the insulation layer construction accuracy control method based on multi-sensor fusion of the present invention. Figure 3 ; Figure 4 The following is the operation flow of the insulation layer construction accuracy control method based on multi-sensor fusion of the present invention. Figure 4 . Detailed Implementation

[0017] The technical solutions of 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.

[0018] Example

[0019] Please see Figures 1-4 As shown, the present invention provides a technical solution comprising the following steps: S1. Collect physical field data during construction by synchronously acquiring data through a multi-source heterogeneous sensor array; S2. Based on the collected physical field data of the construction process, perform data fusion and calculation to obtain implicit construction quality indicators that reflect the internal state of the insulation layer. S3. Input the calculated implicit indicators together with the real-time environmental parameters to predict the future performance degradation of the insulation layer; S4. Based on the performance prediction results, conduct a dynamic assessment of the defect risk level of the current construction unit area; S5. Based on the assessed risk level, perform reverse calculations to optimize the set of construction process parameters that can eliminate the risk; S6. Send the optimized set of construction process parameters to the actuator for adaptive real-time control and compensation of construction parameters. S7. Based on the adjusted construction effect, data is collected again through the sensor array and verified and optimized.

[0020] In this embodiment, step S1 includes a high-frequency micro-vibration sensor, a long-wave infrared thermal imager, a microwave millimeter-wave radar, and a hyperspectral imager sensor. The high-frequency micro-vibration sensor performs spectral analysis on the collected vibration signals, and the frequency domain obtained by Fourier transform is represented as the micro-vibration spectrum. The long-wave infrared thermal imager receives the infrared radiation emitted from the surface of the insulation layer, and according to the Stefan-Boltzmann law, the thermal imager converts the received infrared radiation into an electrical signal.

[0021] The electrical signal is converted into a temperature value by total radiation thermometry, generating a temperature field distribution image; the spatial distribution data of the surface temperature of the insulation layer is obtained by infrared temperature field calculation.

[0022] Specifically, sensors are installed on a unified platform of intelligent construction equipment. Each sensor is equipped with a dedicated signal conditioning circuit, including signal amplification, filtering, and level conversion. Standardized interfaces are used for physical connection between the sensors and the acquisition system. Based on hardware-level timestamp synchronization technology, the time reference of each sensor is unified through PTP. The optimal acquisition frequency is determined according to the data characteristics of each sensor, such as 100Hz for micro-vibration sensors, 30Hz for thermal imagers, 60Hz for millimeter-wave radar, and 20Hz for hyperspectral imagers. This is linked with the motion control system of the construction equipment to synchronize sensor data acquisition with construction actions. Sensor data acquisition is initiated in each construction unit area according to the grid division of the construction area.

[0023] In this embodiment, in step S2, the physical field data collected during the construction process is cleaned in real time, the hollow vibration frequency in the microseismic spectrum is extracted from the microseismic spectrum, and the interface density index is calculated based on the vibration-thermal coupling physical model. This is achieved as follows: , In the formula, D represents the density index. Represents the vibration frequency coefficient. This represents the peak frequency related to the hollowness in the microseismic spectrum. Represents the temperature gradient coefficient. It represents the standard deviation of the infrared temperature field.

[0024] In this embodiment, in step S2, the spatial standard deviation of the dielectric constant is extracted from millimeter-wave radar data, and the moisture content and bonding uniformity are calculated based on the dielectric chemical coupling physical model, which is implemented as follows: , In the formula, W represents the moisture content. Represents the dielectric constant coefficient. This represents the standard deviation of the dielectric constant of millimeter-wave radar. Indicates density, moisture content conversion factor; The bonding uniformity is achieved as follows: , In the formula, U represents the uniformity of adhesion. Indicates the bond density coupling coefficient. This represents the standard deviation of the reflectance of hyperspectral adhesives. The value represents the average reflectance of the hyperspectral adhesive. The density D affects the bonding effect. The lower the D, the worse the bonding quality. Therefore, (1−D) is introduced as a correction factor to map the continuous value output by the fusion model into a standardized index that can be operated in the project, and to generate index values ​​independently for each construction unit area.

[0025] In this embodiment, in step S3, the output implicit construction quality indicators are precisely correlated with the collected real-time environmental parameters according to the construction unit area, and preprocessed. The preprocessed data is then used to predict future performance degradation, which is achieved as follows: , In the formula, Indicates the future Annual performance degradation rate Indicates the weighting coefficient. Indicates ambient temperature. Indicates ambient humidity. Indicates wind speed. Indicates solar radiation intensity. This indicates the reference environmental parameters.

[0026] Specifically, density D is negatively correlated with performance degradation, 1−D represents the degree of hollowness, and the larger 1−D is, the faster the performance degradation; weighting coefficient The moisture content W was determined through fitting of a large amount of experimental data; it is positively correlated with performance degradation, and the larger W is, the faster the performance degradation; the weighting coefficient... The bonding uniformity U was determined through fitting experimental data; it is negatively correlated with performance degradation, where 1−U represents the degree of bonding non-uniformity, and the larger 1−U is, the faster the performance degradation; the weighting coefficient... Determined through fitting experimental data.

[0027] In this embodiment, in step S3, based on the future The annual performance degradation rate, simulating the long-term performance changes of the insulation layer under real service conditions, is achieved as follows: , In the formula, This represents the percentage of performance degradation over service time t. This represents the decay rate constant.

[0028] In this embodiment, in step S4, performance degradation prediction data for each construction unit area is received in real time, including the predicted degradation ratio and degradation rate. Real-time integrity checks are performed on the data: if a unit area lacks a predicted value or the degradation ratio exceeds a preset range, the system automatically triggers the re-prediction process in step S3. Real-time environmental parameters are collected, including ambient temperature, humidity, wind speed, and solar radiation intensity, combined with the current climatic characteristics of the construction area. In high humidity environments, such as humidity ≥70%, the high-risk threshold will be automatically raised to 18%, and the medium-risk threshold will be adjusted to 15% simultaneously. In extremely cold environments, such as temperatures ≤-10°C: the high-risk threshold will be automatically lowered to 12%, and the medium-risk threshold will be adjusted to 10% simultaneously. Standard environment, such as humidity <70% and temperature >-10°C: maintain the baseline threshold; Environmental thresholds are dynamically adjusted based on the correlation analysis between historical construction data and climate conditions; the predicted attenuation ratio for each unit area is compared step by step according to the dynamic threshold. A value with an attenuation ratio less than or equal to the dynamic low-risk threshold is marked as low-risk. For example, an attenuation ratio of 8.5% and a humidity of 65% are considered low-risk. A dynamic low-risk threshold less than the attenuation ratio less than or equal to the dynamic medium-risk threshold is marked as medium-risk. For example, an attenuation ratio of 13.2% and a humidity of 65% are considered medium-risk. A decay rate greater than the dynamic risk threshold is marked as high risk; for example, a decay rate of 17.8% and a humidity of 75% are both high risk. The system automatically records the judgment logic, such as humidity 75% being a high-risk threshold of 18%, and attenuation ratio of 17.8% being less than 18% being a high-risk condition; By precisely linking risk levels with the coordinates of construction unit areas, a three-dimensional spatial heat map is dynamically generated. Low-risk areas: highlighted in green in the construction digital twin; Medium-risk areas: highlighted in yellow; High-risk areas: Displayed as bright red and flashing warning; Heat maps are projected in real time through the AR interface of intelligent construction equipment.

[0029] In this embodiment, in step S5, the risk level and corresponding predicted attenuation ratio of the current construction unit area are received from step S4. Combined with the dynamic environmental threshold rule in step S4, the target attenuation ratio required to eliminate the risk is determined. High-risk areas: The target attenuation rate is less than or equal to the low-risk dynamic threshold, such as ≤12% when the humidity is 75%; Medium-risk areas: The target attenuation ratio is less than or equal to the medium-risk dynamic threshold. For example, when the humidity is 65%, the target attenuation ratio is ≤10%. Based on the technical specifications of construction equipment and historical construction databases, the physically feasible range is defined for each process parameter, including: Spraying air pressure: 0.5MPa to 2.0MPa; Discharge rate: 0.1L / s to 0.5L / s; Robotic arm movement trajectory: 0.1m / s to 1.0m / s; Distance between the nozzle and the wall: 10cm to 30cm.

[0030] Verification with the equipment manufacturer ensures that boundary values ​​cover the safe operating range and excludes unexecutable parameters.

[0031] Input the target attenuation ratio into the S3 performance attenuation prediction model and perform the reverse optimization process: Generate initial parameter population: Randomly generate 100 sets of process parameter combinations within the feasible region; Fitness assessment: For each set of parameters, the predicted decay rate is calculated using the model, compared with the target value, and the difference value is calculated as: difference = predicted value - target value; Iterative optimization: Based on the difference value, selection, crossover and mutation are performed, and it converges quickly within 50 iterations.

[0032] The parameter combination that satisfies the objective and requires the fewest iterations is selected from the population, verified, and the optimization results are encapsulated into a structured instruction package.

[0033] In this embodiment, in step S6, the actuator receives a generated structured instruction packet via a real-time communication protocol, which includes the coordinates of the construction area, risk level, optimized parameter values, and operation type. Upon receiving the instruction, the actuator immediately initiates a self-check of its equipment status: calibrating the robotic arm position and checking the sensor's working status. When the actuator moves to a medium-risk construction unit area, the system automatically triggers a fine-tuning of the process parameters. Adaptive parameter adjustment: Dynamically modify spraying air pressure, material output speed, and nozzle-to-wall distance while maintaining construction continuity; Intelligent trajectory compensation: The robotic arm's movement speed is finely adjusted to the optimized value, and the nozzle angle is compensated in real time to ensure seamless connection between compensation construction and main construction; Self-healing closed-loop verification: The device's built-in sensors continuously monitor the interface status. If the deviation exceeds the threshold, it will automatically trigger millisecond-level fine-tuning. When the actuator is located in a high-risk area, the system immediately executes a shutdown and repair procedure: Safe positioning and protocol activation: The device quickly and accurately returns to the target coordinates and activates the safety protocol, such as turning off the power to adjacent devices and opening the dustproof protective cover; Multidimensional repair operation: Secondary spraying: If the spraying amount is increased by 10%, the nozzle distance is shortened to 12cm; Ultrasonic compaction: such as activating a high-frequency vibration module to eliminate internal voids; Humidity control: If the moisture content is >3.0%, local dehumidification will be activated simultaneously; Repair compliance verification: After repair, the density and moisture content are detected in real time by sensors. If the standard is met, construction is resumed. If the standard is not met, the repair process is automatically repeated. During the control process, real-time environmental parameters are continuously received, and the repair strategy is dynamically optimized.

[0034] In this embodiment, in step S7, after the compensation construction is completed, the system automatically triggers the precise re-acquisition of the sensor array. The sensor array starts synchronous acquisition within 10 seconds after the construction equipment completes the work in the unit area. The re-acquisitioned physical field data is input into the fusion calculation process of step S2 to generate a new implicit index after adjustment. This index is compared with the risk level threshold of step S4 in real time to dynamically determine whether the risk level has dropped to low risk. If the risk level does not meet the standard, the system automatically marks the area as needing secondary repair and triggers the repeated adjustment process of step S6 until the risk level stabilizes in the low risk range.

[0035] Working principle: By integrating high-frequency micro-vibration sensors, long-wave infrared thermal imagers, millimeter-wave radars, and hyperspectral imagers, vibration signals, temperature fields, dielectric constants, and adhesive distribution data are collected simultaneously during construction. The sensors are installed on a unified platform for intelligent construction equipment, and hardware-level time synchronization and spatial calibration ensure that the data are strictly aligned in timestamps and spatial coordinates. Dedicated signal conditioning circuits and standardized interface design are used. By collecting multi-source data, implicit indicators of construction quality are extracted through physical models and data fusion algorithms. For example, the interface density is calculated from the microseismic spectrum and infrared temperature field, and the moisture content and bonding uniformity are calculated by combining the dielectric constant of millimeter-wave radar and the characteristics of hyperspectral adhesives, quantifying defects such as hollowness and humidity accumulation inside the insulation layer. All indicators are mapped to standardized values ​​that can be operated in the project and are precisely bound to the construction unit area. The implicit indicators are combined with real-time environmental parameters and input into a pre-trained digital twin model to simulate the long-term performance degradation trajectory of the insulation layer under future service environment. The performance degradation trend is quantified by linearly combining construction quality indicators and normalized environmental parameters with weighted coefficients. Furthermore, an exponential decay model is introduced to simulate the performance degradation ratio under service time, realizing full-cycle prediction from construction completion to the end of service. Based on the degradation ratio predicted by S3, the risk threshold is dynamically adjusted in combination with real-time environmental parameters, and the predicted degradation ratio of each construction unit area is compared level by level to mark low, medium and high risk levels. Through real-time projection of three-dimensional spatial heat map and AR interface, construction personnel can intuitively locate defect areas and prioritize the treatment of high-risk units. The system automatically records the judgment logic. The system derives a set of process parameters that can eliminate defects by working backward from the risk level, and performs inverse optimization using a digital twin model. Based on the target attenuation ratio, an initial parameter population is generated within the technical specifications of the construction equipment, and iterative optimization using a genetic algorithm quickly converges to the optimal solution. The parameter combination that satisfies the target and requires the fewest iterations is selected and encapsulated as a structured instruction package. Upon receiving the structured instruction package, the actuator immediately initiates a self-check of the equipment status. Fine-tuning of process parameters is triggered in medium-risk areas to ensure the continuity of the main construction. Shutdown repairs are performed in high-risk areas, with operational safety ensured through safety protocols. Built-in sensors monitor the repair effect in real time; if the target is not met, the repair process is automatically repeated. Combined with dynamic optimization strategies for environmental parameters, the system achieves accuracy and efficiency in defect repair. After compensation construction is completed, the system automatically triggers sensor array re-acquisition, synchronously acquiring the adjusted physical field data, which is then input into the fusion calculation process to generate new implicit indicators. These indicators are compared in real time with the risk threshold to dynamically determine whether the risk level has decreased to low risk. If the target is not met, the area is marked as requiring secondary repair, triggering a repeated adjustment process until the risk level stabilizes. The prediction model and process parameters are continuously optimized through closed-loop verification and incremental model training.

[0036] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their likenesses.

[0037] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. A method for controlling the construction accuracy of thermal insulation layers based on multi-sensor fusion, characterized in that, Includes the following steps: S1. Collect physical field data during construction by synchronously acquiring data through a multi-source heterogeneous sensor array; S2. Based on the collected physical field data of the construction process, perform data fusion and calculation to obtain implicit construction quality indicators that reflect the internal state of the insulation layer. S3. Input the calculated implicit indicators together with the real-time environmental parameters to predict the future performance degradation of the insulation layer; S4. Based on the performance prediction results, conduct a dynamic assessment of the defect risk level of the current construction unit area; S5. Based on the assessed risk level, perform reverse calculations to optimize the set of construction process parameters that can eliminate the risk; S6. Send the optimized set of construction process parameters to the actuator for adaptive real-time control and compensation of construction parameters. S7. Based on the adjusted construction effect, data is collected again through the sensor array and verified and optimized.

2. The method for controlling the construction accuracy of thermal insulation layer based on multi-sensor fusion according to claim 1, characterized in that: In S1, a high-frequency micro-vibration sensor, a long-wave infrared thermal imager, a microwave millimeter-wave radar, and a hyperspectral imager are configured. The sensors are installed on a unified platform of the intelligent construction equipment. Each sensor is configured with a dedicated signal conditioning circuit, including signal amplification, filtering, and level conversion processing. The optimal acquisition frequency is determined based on the data characteristics of each sensor. The sensor data acquisition is synchronized with the motion control system of the construction equipment to trigger the construction actions synchronously. According to the grid division of the construction area, sensor data acquisition is started in each construction unit area.

3. The method for controlling the construction accuracy of thermal insulation layer based on multi-sensor fusion according to claim 1, characterized in that: In step S2, the physical field data collected during the construction process is cleaned in real time. The vibration frequency of hollow areas in microseismic events is extracted from the microseismic spectrum. Based on the vibration-thermal coupling physical model, the interface density index is calculated. ， In the formula, D represents the density index. Represents the vibration frequency coefficient. This represents the peak frequency related to the hollowness in the microseismic spectrum. Represents the temperature gradient coefficient. It represents the standard deviation of the infrared temperature field.

4. The method for controlling the construction accuracy of thermal insulation layer based on multi-sensor fusion according to claim 3, characterized in that: In step S2, the spatial standard deviation of the dielectric constant is extracted from millimeter-wave radar data, and the moisture content and bonding uniformity are calculated based on the dielectric chemical coupling physical model, as follows: ， In the formula, W represents the moisture content. Represents the dielectric constant coefficient. This represents the standard deviation of the dielectric constant of millimeter-wave radar. Indicates density, moisture content conversion factor; The uniformity of adhesion is achieved as follows: ， In the formula, U represents the uniformity of adhesion. Indicates the bond density coupling coefficient. This represents the standard deviation of the reflectance of hyperspectral adhesives. This represents the average reflectance of the hyperspectral adhesive. The continuous values ​​output by the fusion model are mapped to standardized indicators that are operable in the engineering, and indicator values ​​are generated independently for each construction unit area.

5. The method for controlling the construction accuracy of thermal insulation layer based on multi-sensor fusion according to claim 1, characterized in that: In step S3, the output implicit construction quality indicators are precisely correlated with the collected real-time environmental parameters according to the construction unit area, and preprocessed. The preprocessed data is then used to predict future performance degradation, achieving the following: ， In the formula, Indicates the future Annual performance degradation rate Indicates the weighting coefficient. Indicates ambient temperature. Indicates ambient humidity. Indicates wind speed. Indicates solar radiation intensity. This indicates the reference environmental parameters.

6. The method for controlling the construction accuracy of thermal insulation layer based on multi-sensor fusion according to claim 5, characterized in that: In S3, based on the future The annual performance degradation rate, simulating the long-term performance changes of the insulation layer under real service conditions, is achieved as follows: ， In the formula, This represents the percentage of performance degradation over service time t. This represents the decay rate constant.

7. The method for controlling the construction accuracy of thermal insulation layer based on multi-sensor fusion according to claim 1, characterized in that: In S4, the performance degradation prediction data of each construction unit area is received in real time, and the data integrity is checked in real time: if the unit area is missing a prediction value or the degradation ratio exceeds the preset range, the system automatically triggers the re-prediction process of S3, and collects real-time environmental parameters, combined with the climate characteristics of the current construction area. The environmental threshold is dynamically adjusted based on the correlation analysis between historical construction data and climate conditions; the predicted attenuation ratio of each unit area is compared step by step according to the dynamic threshold; the risk level is precisely bound to the coordinates of the construction unit area to dynamically generate a three-dimensional spatial heat map.

8. The method for controlling the construction accuracy of thermal insulation layer based on multi-sensor fusion according to claim 1, characterized in that: In step S5, the risk level and corresponding predicted attenuation ratio of the current construction unit area are received from step S4, and the target attenuation ratio required to eliminate the risk is determined in combination with the dynamic environmental threshold rule of step S4. Based on the technical specifications of construction equipment and historical construction database, the physical executable range is defined for each process parameter. The target attenuation ratio is input into the performance attenuation prediction model of S3 for reverse optimization. The parameter combination that satisfies the objective and requires the fewest iterations is selected from the population, verified, and the optimization results are encapsulated into a structured instruction package.

9. The method for controlling the construction accuracy of thermal insulation layer based on multi-sensor fusion according to claim 1, characterized in that: In step S6, the actuator receives the generated structured instruction packet through a real-time communication protocol. After receiving the instruction, the actuator immediately initiates a self-check of the equipment status: calibrating the position of the robotic arm and checking the working status of the sensors; when the actuator moves to the medium-risk construction unit area, the system automatically triggers a fine-tuning of the process parameters. When the actuator is located in a high-risk area, the system immediately executes a shutdown and repair procedure; during the control process, it continuously receives real-time environmental parameters and dynamically optimizes the repair strategy.

10. The method for controlling the construction accuracy of thermal insulation layer based on multi-sensor fusion according to claim 1, characterized in that: In S7, after the compensation construction is completed, the system automatically triggers the precise re-acquisition of the sensor array. The sensor array starts synchronous acquisition within 10 seconds after the construction equipment completes the operation in the unit area. The re-acquisitioned physical field data is input into the fusion calculation process of S2 to generate a new implicit index after adjustment. It is compared with the risk level threshold of S4 in real time to dynamically determine whether the risk level has dropped to low risk. If the risk level does not meet the standard, the system automatically marks the area as needing secondary repair and triggers the repeated adjustment process of S6 until the risk level stabilizes in the low risk range.

Images