Construction method of building waterproofing roll material combined with thermal insulation layer

Abstract

This invention relates to the field of building construction technology and discloses a construction method for combining waterproof membrane and thermal insulation layer. The method includes: retrieving historical data to construct a preliminary environmental stress distribution model; delineating boundaries for interface stability-related data points, marking areas exceeding preset thresholds as high-risk areas; extracting specific index data on weak adhesion to simulate the propagation path of delamination cracking; retrieving construction process adjustment schemes from a process library to select core parameters that determine durability performance optimization; performing stress distribution simulation calculations to generate a leakage risk reduction vector; comparing each component against preset thresholds, integrating energy-saving assurance index data, and establishing an interface stability enhancement configuration; extracting environmental stress adaptation adjustment values, performing time-series modeling on the adjustment values, and predicting long-term durability performance trends; and if the continuous downward trend exceeds a preset period, reverting to the high-risk area marking stage to initiate an iterative optimization process.

Description

Technical Field

[0001] This invention relates to the field of building construction technology, specifically to a construction method that combines waterproof building membrane with thermal insulation layer. Background Technology

[0002] In the long-term use of the interface between building waterproofing and insulation materials, there is a core and complex technical problem: how to ensure the stability and durability of the interface under multiple environmental stresses, and prevent material degradation, delamination and cracking, and leakage risks caused by factors such as temperature cycling and moisture erosion.

[0003] This issue is directly related to the overall safety and energy efficiency of building structures. Especially under extreme climatic conditions, the interface between the waterproof membrane and the insulation layer often becomes a weak point, prone to failure due to insufficient adhesion or material aging. For example, in cold northern regions, the temperature cycle formed by the alternation of low winter temperatures and high summer temperatures may lead to concentrated thermal stress at the interface, while in rainy southern regions, there is a risk of decreased adhesion performance due to long-term moisture erosion.

[0004] Around this core issue, several related sub-issues have emerged, such as how to accurately identify the specific location of high-risk areas and quantify their degradation trend, how to predict the propagation path of delamination cracking in complex environmental stress distribution, and how to assess the impact of adjusted process parameters on long-term durability.

[0005] These seemingly minor issues all serve the core problem: to reveal the failure mechanism of materials under specific environments through a comprehensive analysis and prediction of interface stability.

[0006] Especially in actual engineering scenarios, different building types and usage environments have significantly different requirements for material performance. For example, the stress distribution and failure modes of the joint between the exterior walls of high-rise buildings and the waterproofing and insulation interface of the basement are significantly different, which further increases the complexity of problem solving.

[0007] How to specifically analyze the coupling effect between environmental stress and material properties in these diverse business scenarios has become a technical challenge that urgently needs in-depth research. Summary of the Invention

[0008] The purpose of this invention is to provide a construction method for combining building waterproof membrane and thermal insulation layer to solve the problems mentioned in the background art.

[0009] To achieve the above objectives, the present invention provides a construction method for combining building waterproof membrane and thermal insulation layer, comprising: Historical environmental data and material property data of the joint between the waterproof membrane and the insulation layer are retrieved from the preset database. Temperature cycle records and moisture erosion records are separated from the historical environmental data. The temperature cycle records and moisture erosion records are systematically classified and organized to construct a preliminary environmental stress distribution model. Based on the preliminary environmental stress distribution model, the support vector machine algorithm is used to divide the boundary of the interface stability-related data points. When the boundary value in the division result exceeds the preset threshold, it is marked as a high-risk area, and the potential location of material degradation is accurately located. Specific index data of weak adhesion are extracted from high-risk areas, and the propagation path of layered cracking is simulated by gradient boosting decision tree algorithm, and finally a quantitative vector representation of the path is formed. Based on vector representation, if the peak data exceeds the preset threshold, a targeted construction process adjustment plan is retrieved from the preset process library, and the core parameters that determine the optimization of durability performance are selected from it. Using optimized parameters as input for building durability simulation, stress distribution simulation calculations are performed using the finite element analysis method to generate a leakage risk reduction vector. Each component of the leakage risk reduction vector is compared with a preset threshold. When all components are below the threshold, the energy-saving guarantee index data is integrated to establish the final interface stability enhancement configuration. Environmental stress adaptation adjustment values ​​are extracted from the final interface stability enhancement configuration. A recurrent neural network algorithm is applied to perform time series modeling on the adjustment values ​​to predict long-term durability performance trends and generate trend prediction sequences. Based on the trend prediction sequence, if a continuous downward trend is found to exceed the preset period, the process will be backtracked to the high-risk area marking stage to start the iterative optimization process and generate an updated version of the construction technology.

[0010] Preferably, the systematic classification and organization of temperature cycling records and moisture erosion records to construct a preliminary environmental stress distribution model includes: Temperature cycling records are divided into high-temperature and low-temperature stages, and moisture erosion records are divided into high-humidity and low-humidity stages. Temperature cycling records are divided into high-temperature and low-temperature stages, and moisture erosion records are divided into high-humidity and low-humidity stages. The environmental stress changes at the junction are calculated at different stages based on the classified temperature cycling records and moisture erosion records. A linear regression model was used to establish the correspondence between environmental stress variation values ​​and material property data to obtain the environmental stress influence coefficient. Based on the environmental stress influence coefficient, the stress parameters in the preliminary environmental stress distribution model were corrected, and then the comprehensive stress distribution of the joint under the combined action of temperature cycling and moisture erosion was calculated.

[0011] Preferably, the step of using a support vector machine algorithm to perform boundary division on interface stability-related data points, and marking high-risk areas when the boundary values ​​in the division result exceed a preset threshold, and accurately locating potential material degradation locations, includes: When delineating boundaries, it is necessary to select a set of data points related to interface stability from a pre-set database and construct the dataset required for boundary delineation based on the key parameters in the environmental stress distribution model. After executing the support vector machine algorithm, the boundary values ​​are compared with the threshold. If the limit is exceeded, the data points in the high-risk area are re-labeled to lock the range of material degradation. Local environmental stress characteristics are extracted to determine potential location distribution. A correspondence table between material degradation and environmental stress is established by classification and sorting, and priority ranking is implemented. Based on the sorting results, a layered interface stability analysis layer is created, integrating multi-dimensional data to identify key monitoring areas.

[0012] Preferably, the step of extracting specific index data of weak adhesion from high-risk areas, simulating the propagation path of layered cracking using a gradient boosting decision tree algorithm, and finally forming a quantized vector representation of the path includes: Material degradation-related indicator data were extracted from high-risk areas. A preliminary screening of indicator data with weak adhesion was conducted to form an indicator set. The gradient boosting decision tree algorithm was used to simulate the propagation path of layered cracking and generate an initial quantitative expression. In-depth analysis of key node data reveals the cracking distribution pattern, and nodes exceeding the threshold are assigned priority labels; Layered analysis focuses on obtaining the cracking path details of each layer within the interval, classifying and storing the path quantification results, and constructing a correspondence table between layered cracking and weak adhesion.

[0013] Preferably, the step of retrieving targeted construction process adjustment schemes from a preset process library and selecting the core parameters that determine durability performance optimization includes: Extract the peak data from the quantization vector and compare it with a preset threshold. If the peak data exceeds the preset threshold, retrieve the corresponding adjustment scheme from the process library, match the durability performance optimization parameter set, and then select the key subset. The random forest algorithm is used to predict the change vector of the impact of parameter adjustment on durability performance, and to determine the priority application sequence of adjustment schemes.

[0014] Preferably, the step of using optimized parameters as input for building durability simulation, combined with finite element analysis to perform stress distribution simulation calculations, and generating a leakage risk reduction vector includes: The set of durability performance optimization parameters is obtained from the adjustment plan record. The building durability simulation input data is determined. The stress distribution simulation calculation is performed on the building durability simulation input data using the finite element analysis method to obtain the leakage risk reduction vector. For the leakage risk reduction vector, extract the risk distribution peak data and verify whether it is lower than the preset threshold. If the risk distribution peak data is lower than the preset threshold, confirm that it is controllable, lock the priority solution and obtain the material strength improvement parameters.

[0015] Preferably, each component of the leakage risk reduction vector is compared with a preset threshold. When all components are below the threshold, the energy-saving assurance index data is integrated to establish the final interface stability enhancement configuration, including: The data of each component is extracted from the leakage risk reduction vector and compared one by one. If each component is lower than the preset threshold, the leakage risk is judged to be within the controllable range. Obtain data on indicators related to energy conservation assurance, process the data, and determine the degree of compliance with energy conservation assurance requirements; Analyze potential issues related to interface stability and filter indicator data to identify key factors influencing interface stability; Obtain the corresponding enhanced configuration parameters, perform logical matching on the parameters, and if the matching result meets the preset conditions, determine the preliminary enhanced configuration scheme. Obtain resource allocation data related to configuration, verify the resource allocation data, determine whether the resource allocation meets the stability judgment requirements, obtain the final configuration determination basis, integrate the data based on the basis, and obtain the final solution for interface stability enhancement configuration.

[0016] Preferably, the step of extracting environmental stress adaptation adjustment values ​​from the final interface stability enhancement configuration, applying a recurrent neural network algorithm to perform time series modeling on the adjustment values, predicting long-term durability performance trends, and generating trend prediction sequences includes: Environmental stress adaptation adjustment values ​​are obtained from the final interface stability enhancement configuration, time series data is constructed, and a recurrent neural network algorithm is used to train the time series data to generate a durability performance trend prediction sequence. The trend prediction values ​​are extracted from the durability performance trend prediction sequence and the threshold range is verified. If the trend prediction values ​​are all within the preset range, the performance trend data is determined to be stable, the adaptive adjustment parameters are obtained, and the durability prediction output is generated.

[0017] Preferably, if the trend prediction sequence shows a continuous downward trend exceeding a preset period, the process is backtracked to the high-risk area marking stage to initiate an iterative optimization process, producing an updated version of the construction technology, including: The trend prediction sequence is initially processed, and the data is scanned segment by segment using sequence analysis methods to determine whether there is a continuous downward pattern. If a continuous downward pattern is detected that exceeds the preset period, the relevant data segments are risk-assessed to determine the specific location of the high-risk area. The specific locations of high-risk areas are marked, and an iterative processing flow is initiated. The construction processes within the marked areas are analyzed layer by layer to obtain the direction for process optimization. The construction processes are then updated to generate updated process configuration schemes. For the updated process configuration scheme, the trend prediction data is reloaded through the sequence analysis module to determine whether the downward trend has been alleviated and to determine the final process adjustment result.

[0018] Compared with the prior art, the beneficial effects of the present invention are: 1. This application provides a construction method for combining waterproof membrane and thermal insulation layer. By constructing an environmental stress distribution model, using a support vector machine algorithm to identify high-risk areas, and combining a gradient boosting decision tree to simulate crack propagation paths, this method achieves precise quantitative analysis of the stress distribution and degradation trend at the interface between the waterproof membrane and the insulation layer under temperature and humidity cycles. This method can identify weak points at the material interface in advance and predict cracking risks, making adjustments to the construction process more targeted. This significantly reduces the incidence of leaks, delamination, and other defects caused by accumulated environmental stress, extending the overall service life of the roofing system.

[0019] 2. This application provides a construction method combining waterproof membrane and thermal insulation layer. Based on leakage risk reduction vector and finite element stress simulation results, it can automatically compare thresholds and retrieve optimized solutions from a preset process library. It uses a random forest algorithm to predict durability changes after parameter adjustments, forming a priority application sequence. This method can not only dynamically adjust construction parameters (such as material ratio and adhesive layer thickness) based on real-time monitoring data or predicted trends, but also trigger an iterative optimization process when continuous performance degradation is detected, generating an updated process version. This avoids "one-size-fits-all" construction, improves adaptability to different climatic conditions and engineering scenarios, and reduces resource waste caused by over-design or maintenance.

[0020] 3. This application provides a construction method combining waterproof membrane and thermal insulation layer. Based on the confirmed controllable leakage risk, it integrates energy-saving indicators (such as thermal conductivity and insulation layer thickness), outputs a stable and enhanced configuration through data matching and logical judgment, and combines a recurrent neural network for long-term durability trend prediction. This method organically combines waterproof reliability, interface stability, and building energy-saving requirements, achieving full-chain digital management from "environmental stress analysis—risk positioning—process adjustment—performance simulation—long-term prediction." It provides a continuously optimized technical closed loop for building envelope systems, ultimately extending service life while ensuring long-term stable energy-saving effects and reducing operation and maintenance costs. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the construction method of the present invention; Figure 2 A flowchart illustrating the working principle of step 2 of the present invention, which forms a quantized vector representation of the path. Figure 3 The flowchart illustrates the working principle of step 6 of this invention, which establishes the final interface stability enhancement configuration. Detailed Implementation

[0022] 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.

[0023] like Figure 1 As shown, the present application provides a construction method for combining a building waterproof membrane with a thermal insulation layer, which specifically includes the following steps: Step S1: Retrieve historical environmental data and material property data of the joint between the waterproof membrane and the insulation layer from the preset database. Separate temperature cycle records and moisture erosion records from the historical environmental data. Systematically classify and organize the temperature cycle records and moisture erosion records to construct a preliminary environmental stress distribution model.

[0024] In practice, historical environmental data and material property data of the bonding area between the waterproof membrane and the insulation layer are obtained from a pre-set database. Temperature cycling records and moisture erosion records are separated based on the historical environmental data. Temperature cycling records are divided into high-temperature and low-temperature stages, and moisture erosion records are divided into high-humidity and low-humidity stages. The environmental stress variation values ​​of the bonding area at different stages are determined using the categorized temperature cycling and moisture erosion records. A linear regression model is used to establish a correspondence between the environmental stress variation values ​​and the material property data, obtaining the environmental stress influence coefficient. The stress parameters in the preliminary environmental stress distribution model are adjusted based on the environmental stress influence coefficient. The comprehensive stress distribution of the bonding area under the combined effects of temperature cycling and moisture erosion is calculated using the adjusted environmental stress distribution model.

[0025] When analyzing environmental stress changes at the interface between waterproof membrane and insulation layer, historical environmental data can be used as a starting point. Let's assume historical data for the interface of a building's roof records temperature and humidity changes over the past five years. Temperature cycle records show that the highest summer temperature reaches 40 degrees Celsius, and the lowest winter temperature drops to -10 degrees Celsius each year. Moisture erosion records show that humidity reaches as high as 90% during the rainy season, while humidity is only 30% during the dry season. Using this data, temperature cycle records and moisture erosion records can be separated, allowing for the differentiation of high-temperature phases (e.g., summer days exceeding 35 degrees Celsius), low-temperature phases (e.g., winter days below 0 degrees Celsius), high-humidity phases (e.g., rainy season periods with humidity exceeding 80%), and low-humidity phases (e.g., periods with humidity below 40%). This differentiation helps to accurately identify the different environmental impacts on the interface.

[0026] To determine the environmental stress variation, the stress change can be quantified by the temperature difference between high-temperature and low-temperature stages, and the humidity difference between high-humidity and low-humidity stages. Assuming the stress change in the high-temperature stage is positive tensile stress, and in the low-temperature stage it is negative compressive stress, while the high-humidity stage may cause the material to absorb moisture and expand, and the low-humidity stage may induce shrinkage stress, this method clearly characterizes the stress state of the joint under different environmental conditions, laying the foundation for subsequent analysis. For example, when using a linear regression model to establish the correspondence between environmental stress variation and material property data, the tensile strength of the waterproof membrane and the coefficient of thermal expansion of the insulation layer can be considered as key attribute data. Assuming the tensile strength of the membrane is 5 MPa and the coefficient of thermal expansion of the insulation layer is 0.00002 degrees Celsius, regression analysis shows that for every 10-degree Celsius increase in temperature, the stress change increases by approximately 0.5 MPa, while for every 10% increase in humidity, the stress change increases by approximately 0.2 MPa. Determining this environmental stress influence coefficient helps quantify the specific impact of environmental factors on material properties, thereby improving the model's predictive accuracy.

[0027] When adjusting the preliminary environmental stress distribution model, the stress parameters in the model can be corrected based on the aforementioned influence coefficients. Assuming the stress parameter in the high-temperature stage of the preliminary model is 1.0 MPa, it may increase to 1.5 MPa after adjusting the influence coefficients to reflect the greater stress concentration in the actual environment. This adjustment makes the model closer to actual working conditions, avoiding underestimation of stress in the design and improving the durability design effect of the joint. For example, when calculating the comprehensive stress distribution, the adjusted model can simulate the superimposed effects of temperature cycling and moisture erosion. Assuming the comprehensive stress at the joint reaches 2.0 MPa in the high-temperature and high-humidity stage, while decreasing to 0.8 MPa in the low-temperature and low-humidity stage, this distribution result can intuitively reflect the synergistic effect of environmental factors, providing data support for subsequent material selection and structural optimization, ultimately effectively reducing the risk of cracking or delamination at the joint and extending the service life of the building roof.

[0028] Step S2: Based on the preliminary environmental stress distribution model, the support vector machine algorithm is used to divide the interface stability-related data points into boundaries. When the boundary value in the division result exceeds the preset threshold, it is marked as a high-risk area, and the potential location of material degradation is accurately located.

[0029] In practice, a set of data points related to interface stability is obtained from a pre-set database. Data is filtered based on key parameters in the environmental stress distribution model to obtain a basic dataset for boundary delineation. A support vector machine algorithm is used to perform boundary delineation on the filtered basic dataset. The delineation results are compared with a pre-set threshold to determine the distribution range of high-risk areas. If the comparison results show that the boundary values ​​exceed the pre-set threshold, the data points within the high-risk area are re-annotated to determine the potential location range of material degradation. Based on the re-annotated high-risk area data, local environmental stress features related to material degradation are extracted to obtain the specific distribution information of potential locations. By classifying and organizing the potential location distribution information, a correspondence table between material degradation and environmental stress is constructed to determine the priority ranking of each location within the high-risk area. Based on the priority ranking results, the potential locations within the high-risk area are layered and labeled to obtain a detailed distribution layer for interface stability analysis. By integrating the data from the detailed distribution layer, a multi-dimensional analysis record of the high-risk area is generated to identify key monitoring areas for material degradation. For example, when analyzing the interfacial stability of the bonding area between waterproof membrane and insulation layer, a data set related to interfacial stability can be obtained from a pre-set database. This data set typically includes historical environmental stress records, material interfacial bond strength, and local deformation. By filtering data based on key parameters in the environmental stress distribution model, such as selecting data points with stress concentration factors greater than 1.2 and records with bond strengths below 3 MPa, a basic dataset for boundary delineation can be obtained. This filtering effectively removes noisy data and improves the focus of subsequent analysis.

[0030] In one embodiment, a support vector machine (SVM) algorithm is used to perform boundary segmentation on the filtered basic dataset. The SVM identifies the maximum margin hyperplane, dividing data points into stable and potentially unstable regions. The segmentation results visually display the risk level of different areas of the interface. The segmentation results are compared with a preset threshold, for example, a bond strength degradation threshold of 2.5 MPa. If the comparison shows that the boundary value exceeds this threshold, it indicates the existence of a high-risk area requiring further processing. This comparison helps to quickly identify weak points in interface stability. For example, if the comparison results show that the boundary value exceeds the preset threshold, the data points in the high-risk area are secondary-labeled to determine the potential location range of material degradation. Secondary labeling can combine local stress peaks and moisture penetration depth; for example, areas with stress peaks exceeding 1.8 MPa and moisture penetration greater than 5 mm are labeled as potential degradation locations. This labeling can accurately pinpoint the starting point of material aging, avoiding the waste of resources from comprehensive inspections.

[0031] Based on the high-risk area data after secondary annotation, local environmental stress features related to material degradation are extracted, such as shear stress caused by temperature gradients and expansion stress caused by moisture, to obtain specific distribution information of potential locations. These feature extractions can reveal the mechanisms that induce degradation, providing a basis for targeted protection. For example, by classifying and organizing the distribution information of potential locations, a correspondence table between material degradation and environmental stress is constructed. The table records the correspondence between a 20% increase in degradation rate under high-temperature cycling and a 15% decrease in adhesion under high humidity, and uses this to determine the priority ranking of locations within the high-risk area. High-priority locations are usually the edges of the joints where stress and humidity are most severely superimposed. This correspondence table helps quantify risk levels and improve assessment efficiency.

[0032] Based on the priority ranking, potential locations within high-risk areas are stratified and marked, such as a high-priority red layer, a medium-priority yellow layer, and a low-priority green layer, resulting in a detailed distribution layer for interface stability analysis. This layer visually displays the spatial distribution of risks, facilitating focused attention by designers. By integrating data from this detailed distribution layer, multi-dimensional analysis records of high-risk areas are generated, including indicators such as stress, humidity, and degradation rate, ultimately identifying key monitoring areas for material degradation. For example, the corner of the roof edge is identified as a key monitoring area due to its highest comprehensive indicators. This multi-dimensional recording significantly improves the accuracy of interface durability prediction, reduces cracking risk, and extends the overall lifespan of the roof system.

[0033] Step S3: Extract specific index data of weak adhesion from high-risk areas, simulate the propagation path of layered cracking using gradient boosting decision tree algorithm, and finally form a quantitative vector representation of the path.

[0034] For specific implementation, please refer to Figure 2As shown, by acquiring data indicators related to material degradation from high-risk areas, a preliminary screening of potential locations with weak adhesion is conducted to obtain a set of indicators for subsequent simulation analysis. Based on the selected indicator set, a gradient boosting decision tree algorithm is used to simulate the propagation path of delamination cracking, obtaining an initial quantitative expression of the path. For the initial quantitative expression, key node data in the propagation path is extracted to determine the distribution pattern of delamination cracking in potential locations. If the distribution pattern shows that key nodes exceed a preset threshold range, these nodes are prioritized to obtain the key focus intervals of the propagation path within the high-risk area. By performing layered processing on the data in the key focus intervals, detailed information on the cracking path at each level is obtained to determine the specific distribution of weak adhesion in potential locations. The detailed information after layered processing is obtained, and the path data at each level is classified and stored to determine the path quantification results of material degradation within the high-risk area. Based on the classified and stored path quantification results, a correspondence table between delamination cracking and weak adhesion is constructed to obtain structured data records for subsequent analysis.

[0035] When analyzing high-risk areas at the interface between waterproof membrane and insulation layer, one can start with data indicators related to material degradation to initially screen potential locations with weak adhesion. High-risk areas typically refer to regions where historical data shows the bonding interface is easily affected by temperature cycling and moisture erosion. By collecting indicators such as the adhesion strength decay rate and interfacial peel energy in these areas, threshold comparisons can be performed to obtain a set of indicators for subsequent simulations. This screening helps focus on key degradation points and avoids wasting resources on comprehensive analysis. For example, suppose the initial adhesion strength of a certain interface in a high-risk area is 8 MPa, which decays to 4.5 MPa after five years of environmental effects, and the interfacial peel energy decreases from an initial 500 joules per square meter to 250 joules per square meter. By setting a threshold of decay rate exceeding 30%, these locations can be screened to form an indicator set, including the magnitude of adhesion strength decay and peel energy loss rate, thus providing targeted data support for simulations.

[0036] Based on the selected set of indicators, a gradient boosting decision tree algorithm is used to simulate the propagation path of delamination cracking. The gradient boosting decision tree iteratively constructs a weak learner, gradually optimizing the fit to the nonlinear decay relationship to obtain an initial quantitative expression of the path. For example, by inputting temperature cycle number, humidity fluctuation amplitude, and material degradation index into the simulation, the algorithm can output the probability distribution of crack path expansion from the edge to the center, initially expressed as a correlation curve between path length and decay rate. This method effectively captures path evolution under the interaction of multiple factors, improving the accuracy of the simulation. For the initial quantitative expression, key node data in the propagation path, such as path branch points or stress concentration points, are extracted to determine the distribution pattern of delamination cracking in potential locations.

[0037] In one possible implementation, if the path shows multiple branch nodes in the middle of the interface, the distribution pattern shows that the crack density decreases with increasing depth. Identifying this pattern helps reveal the hierarchical differences in decay. For example, if the distribution pattern shows that the local adhesion at key nodes is below 2 MPa and exceeds a preset threshold of 1.5 MPa, these nodes are prioritized, resulting in key areas of concern for the propagation path within high-risk zones. This prioritization allows for the allocation of analytical resources, quickly identifying high-risk segments and improving the efficiency of risk management. By stratifying the data in these key areas of concern, detailed information about the crack path at each level can be obtained, such as rapid surface expansion and slow deep penetration, allowing for the determination of the specific distribution of weak adhesion at potential locations. This stratification contributes to a more refined understanding of the decay mechanism.

[0038] After obtaining the detailed information after layering, the path data for each layer is categorized and stored according to expansion speed and direction, determining the path quantification results of material degradation in high-risk areas. For example, the quantified path length for the surface layer is 15 cm, and for the deeper layer, it is 5 cm, forming a clear quantification record. Based on the categorized and stored path quantification results, a correspondence table between layered cracking and weak adhesion is constructed. For example, high expansion speed corresponds to areas with adhesion below 3 MPa, resulting in structured data records for subsequent analysis. This table can intuitively guide material strengthening strategies, significantly reduce cracking risk, and extend the durability of the bonding area.

[0039] Step S4: Based on vector representation, if the peak data exceeds the preset threshold, a targeted construction process adjustment plan is retrieved from the preset process library, and the core parameters that determine the durability performance optimization are selected from it.

[0040] In practice, peak data is extracted based on the propagation path quantization vector representation. The peak data is then compared with a preset threshold. If the peak data exceeds the preset threshold, a corresponding construction process adjustment scheme is extracted from the preset process library. A set of durability performance optimization parameters is obtained by matching the adjustment schemes. A subset of key optimization parameters is selected based on this set. A random forest algorithm is used to predict the durability performance change vector after parameter adjustment for this subset of key optimization parameters. The priority application sequence of the adjustment schemes is determined based on the durability performance change vector. For example, when analyzing the durability performance optimization of the joint between waterproofing membrane and insulation layer on a building roof, the perspective of propagation path quantization vectors can be used to explore how environmental stress affects material performance. Specifically, data collected by sensors from factors such as temperature cycling and moisture erosion can be converted into vector form, and then peak data can be extracted. For example, in monitoring the joint of a building roof, peak temperature data may show a summer temperature of 42 degrees Celsius, far exceeding the conventional design value. Extracting this peak data provides a basis for subsequent judgments. For example, for comparing peak data with preset thresholds, a temperature threshold of 38 degrees Celsius and a humidity threshold of 85% can be set. If a peak temperature of 42 degrees Celsius is detected, significantly exceeding the threshold, the system will trigger an alert, indicating that the bonding area may face excessive thermal stress. Similarly, a humidity peak of 90% also warrants attention regarding the possibility of material moisture absorption and expansion. This comparative assessment method is intuitive and efficient, quickly identifying potential problems. When peak data exceeds the limit, when retrieving construction process adjustment plans from the preset process library, historical data can be used to select process plans specifically for high-temperature and high-humidity environments. For example, if the process library contains a roll material formulation with added heat-resistant aging agents and a construction method to enhance the bonding strength of the insulation layer, these plans can be considered as alternatives. By matching a set of durability performance optimization parameters, such as a 2% addition ratio of heat-resistant aging agents and an increase in bonding strength to 3 MPa, these parameter sets provide data support for subsequent optimization.

[0041] When selecting a subset of key optimization parameters, priority should be given to the proportion of heat-resistant aging agent and bond strength, as these directly affect the durability of the bonded area. During implementation, historical data analysis can be used to confirm that these two parameters perform most critically under high-temperature and high-humidity environments, thus forming a subset. This selection method ensures targeted optimization.

[0042] When using the random forest algorithm to predict the durability performance change vector, a subset of key parameters can be input to simulate the adjusted performance. Assuming the prediction results show that increasing the proportion of heat-resistant aging agent to 2.5% can extend durability by approximately 3 years, the output of this change vector provides a reference for subsequent decision-making.

[0043] When determining the priority sequence for adjustment schemes, they can be ordered according to the magnitude of the durability performance change vector. If the adjustment of the heat aging resistant agent results in a greater improvement in durability, this scheme should be implemented first, while the bond strength adjustment should be the second choice. This sequence ensures efficient resource utilization while maximizing the durability performance of the joint. Through these multi-faceted implementation methods, the environmental stress resistance of the building roof joints can be systematically improved, providing a guarantee for long-term use.

[0044] Step S5: Using the optimized parameters as input for building durability simulation, stress distribution simulation calculation is performed using the finite element analysis method to generate a leakage risk reduction vector.

[0045] In practice, a set of durability performance optimization parameters is obtained from the adjustment scheme record. The building durability simulation input data is determined based on this set of parameters. The finite element method is used to simulate and calculate the stress distribution of the input data. The leakage risk reduction vector is obtained through this simulation. Peak risk distribution data is extracted from the leakage risk reduction vector. If the peak risk distribution data is below a preset threshold, the current adjustment scheme record is designated as the priority scheme. Corresponding material strength enhancement parameters are then obtained from the priority scheme record.

[0046] When extracting the set of durability performance optimization parameters from adjustment scheme records, one can focus on historical data of the junction between waterproofing and insulation on building roofs. Assuming the records contain multiple parameters, such as the tensile strength of the waterproofing membrane, the compressive strength of the insulation layer, and the adhesion strength of the junction, this set of parameters provides the foundational data for subsequent simulations. A specific implementation method could be to filter adjustment scheme records from the past 5 years in the database, extracting parameter values ​​related to high-temperature and high-humidity environments, such as a tensile strength requirement of 2.5 MPa and an adhesion strength requirement of no less than 1.8 MPa.

[0047] To determine the input data for building durability simulation based on the set of durability performance optimization parameters, the aforementioned parameter values ​​can be used as input conditions, combined with environmental factors such as temperature fluctuations and humidity changes, to form a comprehensive dataset. In implementation, the simulation scenario can be set as a high-temperature and high-humidity summer environment, with an input temperature range of 30 to 45 degrees Celsius and a humidity range of 80% to 95%, thereby simulating the durability performance of the joint components.

[0048] When using the finite element method to simulate stress distribution, software tools can be used to mesh the input data to simulate the stress distribution at the joint under different environmental conditions. Specifically, the joint is divided into multiple micro-elements, and the stress change trend of each element is analyzed as temperature or humidity increases. For example, when the temperature reaches 42 degrees Celsius, localized stress concentration may occur in certain areas.

[0049] The leakage risk reduction vector obtained through stress distribution simulation can be transformed into a quantitative indicator reflecting the degree to which the adjustment scheme improves the leakage risk. The implementation method compares the changes in peak stress at the joint before and after the adjustment. Assuming the peak stress decreases from 3.2 MPa to 2.1 MPa after adjustment, a risk reduction vector can be generated, indicating a decrease in the likelihood of leakage.

[0050] When extracting peak stress data for risk distribution, the highest stress distribution value can be selected from the simulation results as a key indicator. Specifically, this involves analyzing stress data from different areas of the joint. Assuming a peak stress of 2.3 MPa in one area, while other areas are below 2.0 MPa, this peak value is used as the basis for subsequent judgment. If the peak stress data for risk distribution is below a preset threshold, the current adjustment plan is recorded as the priority plan. In practice, a stress threshold of 2.5 MPa can be set. If the peak stress is 2.3 MPa, below the threshold, the plan is considered to perform better in reducing leakage risk and can be prioritized. This method helps to quickly identify effective solutions.

[0051] When obtaining the corresponding material strength improvement parameters based on the priority scheme records, specific material improvement data can be extracted from the scheme. The implementation method is to check the material parameter adjustment records for waterproof membrane and insulation layer in the priority scheme, such as increasing the tensile strength of the membrane to 2.8 MPa and the adhesion to 2.0 MPa. These parameters provide clear guidance for subsequent construction.

[0052] Step S6: Each component of the leakage risk reduction vector is compared with a preset threshold. When all components are below the threshold, the energy-saving guarantee index data is integrated to establish the final interface stability enhancement configuration.

[0053] For specific implementation, please refer to Figure 3As shown, data for each component is extracted from the leakage risk reduction vector. The detection results for each component are compared one by one. If each component is below a preset threshold, the leakage risk is considered controllable, leading to a risk analysis conclusion. Based on the risk analysis conclusion, energy-saving assurance-related indicator data is obtained. A pre-established evaluation model is used to process the indicator data to determine the degree of compliance with energy-saving assurance requirements. Through the degree of compliance with energy-saving assurance requirements, potential problems with interface stability are analyzed. Data comparison tools are used to filter the indicator data, identifying key influencing factors for interface stability. Based on these key influencing factors, corresponding enhanced configuration parameters are obtained. Logical matching of these parameters is performed. If the matching result meets preset conditions, a preliminary enhanced configuration plan is determined. Using the preliminary enhanced configuration plan, resource allocation data related to the configuration is obtained. A logical judgment tool is used to verify the resource allocation data, determining whether the resource allocation meets the stability judgment requirements. Based on the stability judgment requirements, the final configuration determination basis is obtained. Data integration is performed on this basis to obtain the final enhanced configuration plan for interface stability.

[0054] When extracting data from the leakage risk reduction vector, the vector can be decomposed into multiple independent components. For example, if the vector contains three components, namely stress reduction of 0.8 MPa, adhesion increase of 0.6 MPa, and permeability decrease of 15%, extracting these components will provide a quantitative basis for subsequent comparisons.

[0055] The results of component testing are compared one by one. If each component is below a preset threshold, the leakage risk is considered to be within a controllable range, leading to a risk analysis conclusion. The thresholds can be set as a stress reduction of at least 0.5 MPa, an increase in adhesion of at least 0.4 MPa, and a decrease in permeability of at least 10%. In one possible implementation, if the actual component values ​​are 0.9 MPa, 0.7 MPa, and 18%, respectively, all exceeding the threshold requirements, the conclusion is that the risk is controllable. This comparison helps to quickly confirm the reliability of the solution and avoid increased long-term maintenance costs due to potential leakage hazards.

[0056] When obtaining energy-saving data based on risk analysis conclusions, if the risk is controllable, indicators such as thermal conductivity and insulation layer thickness are extracted from the project records. These indicators directly affect the control of energy loss in buildings under high temperature and high humidity environments.

[0057] A pre-established assessment model is used to process the indicator data to determine the degree of compliance with energy-saving guarantees. The model can be based on comparisons with standards and specifications. For example, if the thermal conductivity coefficient needs to be below 0.035 W / m Kelvin, and an actual value of 0.028 W / m Kelvin is input during implementation, the degree of compliance is high. This process ensures durability optimization while supporting energy-saving targets, thereby improving overall building performance.

[0058] By analyzing the compliance rate with energy-saving standards, potential problems with interface stability are identified. If the compliance rate exceeds 90%, the main potential issues are interface separation risks caused by temperature cycling. Data comparison tools are used to screen the indicator data, revealing key influencing factors on interface stability, such as the difference in thermal expansion coefficients between the insulation and waterproofing layers. Screening results show that a difference exceeding 5% easily leads to interface instability. This identification provides a basis for targeted enhancements, effectively reducing the probability of interface failure.

[0059] Based on the key factors affecting interface stability, corresponding reinforcement configuration parameters are obtained, such as increasing the thickness of the interfacial adhesive or selecting low-expansion materials. Logical matching is performed on these parameters. If the matching results meet preset conditions, a preliminary reinforcement configuration is determined. These conditions may include controlling the difference in the coefficient of thermal expansion within 3%. If the parameters are met during matching, the preliminary solution is to add a 0.5 mm thick adhesive reinforcement layer. This approach helps improve interface durability and reduce the risk of separation under high temperature and high humidity conditions.

[0060] The initial configuration enhancement plan obtains relevant resource allocation data, such as material usage and construction hours. Logical judgment tools are used to verify the resource allocation data to determine if it meets the stability requirements. If material usage does not exceed 10% of the budget and construction hours are within the standard range, the requirements are met. This verification ensures the feasibility of the plan and avoids resource waste. Based on the stability requirements, the final configuration determination criteria are obtained, such as overall cost and effectiveness ratio. Data is then integrated based on these criteria to obtain the final interface stability enhancement configuration plan.

[0061] The final integrated solution employs a specific adhesive combined with thickness adjustments, resulting in an overall improvement in interface stability of over 20%. This final solution not only enhances durability but also ensures long-term stability of energy-saving performance, leading to significant building maintenance benefits.

[0062] Step S7: Extract environmental stress adaptation adjustment values ​​from the final interface stability enhancement configuration, apply a recurrent neural network algorithm to perform time series modeling on the adjustment values, predict the long-term durability performance trend, and generate a trend prediction sequence.

[0063] In practice, environmental stress adaptation adjustment values ​​are obtained from the final interface stability enhancement configuration. Time series data is constructed based on these environmental stress adaptation adjustment values. A recurrent neural network algorithm is used to train the time series data to obtain a durability performance trend prediction sequence. Trend prediction values ​​are extracted from this sequence. Threshold comparisons are applied to the trend prediction values; if all trend prediction values ​​are within a preset range, the performance trend data is considered stable. Adaptation adjustment parameters are obtained based on the performance trend data. Durability prediction output is generated using these adaptation adjustment parameters.

[0064] When obtaining environmental stress adaptation adjustment values ​​from the final interface stability enhancement configuration, the compensation parameters for high-temperature and high-humidity cyclic loading in the configuration scheme can be directly extracted. These adjustment values ​​typically reflect the deformation compatibility of material layers under temperature and humidity fluctuations.

[0065] In one embodiment, assuming the enhanced configuration includes adding a 0.4 mm thick flexible transition layer and using an adhesive with a 12% reduced coefficient of thermal expansion, the corresponding environmental stress adaptation adjustment values ​​are a stress buffer rate of 28% and a deformation compatibility coefficient increased to 1.15. These values ​​are derived from the quantitative mapping of configuration parameters, providing basic data for subsequent predictions.

[0066] By constructing time-series data for environmental stress adaptation adjustment values, these values ​​can be combined with historical environmental load data to form a sequence arranged by time step. The sequence can cover the temperature and humidity cycles within a typical year, for example, by sampling once per quarter to record the response changes of the corresponding adjustment values ​​under simulated aging conditions.

[0067] In one possible implementation, the sequence data includes an initial adjustment of 28%, 26% after the third month, 25% after the sixth month, and 24% after the twelfth month, demonstrating a slight decay trend. This time series construction helps to capture the dynamic evolution under long-term environmental influences.

[0068] When training time series data using a recurrent neural network algorithm, the network's memory properties are utilized to capture the dependencies between sequences, resulting in a durability performance trend prediction sequence. After training, the prediction sequence shows that the durability retention rate is expected to be 92% in month 18 and 90% in month 24, forming a gently declining trend curve. This prediction sequence provides a basis for early identification of potential degradation points.

[0069] By extracting trend forecast values ​​from the durability performance trend forecast sequence, retention rates at key time points can be selected as representative values. For example, 90% at month 24 and 88% at month 36 can be extracted; these values ​​directly quantify the remaining performance of the interface under continuous environmental stress. In one possible implementation, the trend slope can also be calculated, such as keeping the average annual degradation rate below 1.2%, reflecting the long-term effectiveness of the configuration.

[0070] A threshold comparison is applied to the trend prediction values. If all trend prediction values ​​are within a preset range, the performance trend data is considered stable. The preset range can be set to a durability retention rate of no less than 85% and a degradation rate of no more than 2%.

[0071] In one embodiment, when the predicted durability retention rates extracted each time are 90%, 88%, and 86%, all exceeding the threshold, the conclusion is that the performance trend is stable. This comparison quickly verifies the configuration's adaptability to environmental stresses, reducing the risk of long-term failure.

[0072] When obtaining adaptive adjustment parameters based on performance trend data, if the trend is stable, the optimization parameters are deduced from the prediction model, such as further increasing the flexible layer thickness to 0.5 mm or adjusting the adhesive ratio. These parameters specifically strengthen or weaken the weakened components.

[0073] In one possible implementation, the acquired parameters include a buffer rate improvement target of 32%, directly guiding configuration iterations. For example, by adaptively adjusting parameters to generate durability prediction outputs, these parameters can be integrated into the original model, resulting in a more accurate long-term durability report. The output includes an expected service life extension of over 25 years and a reduction in interface separation probability to below 3%. This predictive output provides a reliable basis for building maintenance decisions, ensuring the continued stability of energy-saving and waterproofing performance under high-temperature and high-humidity environments, significantly reducing maintenance frequency and costs.

[0074] Step S8: Based on the trend prediction sequence, if a continuous downward trend is found to exceed the preset period, the process is backtracked to the high-risk area marking stage to start the iterative optimization process and produce an updated version of the construction technology.

[0075] In practice, the trend prediction sequence undergoes initial processing. Sequence analysis methods are used to scan the data segment by segment to determine if a continuous downward trend exists, yielding a decline detection result. Based on the decline detection result, if a continuous downward trend is detected exceeding a preset period, a risk assessment is performed on the relevant data segments to determine the specific location of high-risk areas. For the location information of high-risk areas, automated tools are used to mark the areas, generating marked data records. Using these marked data records, an iterative processing flow is initiated to analyze the construction processes within the marked areas layer by layer, identifying areas for process optimization. Based on these optimization directions, the construction processes are updated, generating an updated process configuration scheme. For the updated process configuration scheme, the trend prediction data is reloaded using the sequence analysis module to determine if the downward trend has been mitigated, thus determining the final process adjustment result.

[0076] When initially processing a trend prediction sequence, the data can be scanned segment by segment using sequence analysis methods to determine if a continuous decline pattern exists. In principle, this scan aims to capture potential adverse changes in the data, particularly signs of performance degradation under long-term environmental stress. For example, assuming the trend prediction sequence covers 36 months of durability performance data, and the scan reveals a continuous decline in retention rate from month 12 to month 18, from 95% to 90%, with the decline widening month by month, this pattern may indicate insufficient adaptability of the material interface under specific environmental conditions. The scanning method can be based on a time window, such as every 6 months as an analysis unit, comparing data changes segment by segment to obtain degradation detection results, providing a basis for subsequent risk assessment.

[0077] When conducting risk assessments on degradation detection results, if a continuous degradation pattern is detected exceeding a preset period, such as a degradation lasting six consecutive months, in-depth analysis of the relevant data segments is required to identify high-risk areas. For example, assuming a preset period of five months, and actual data showing a continuous degradation for six months from month 12 to month 18, this interval is marked as a high-risk area. Risk assessments can be conducted from two dimensions: environmental load intensity and material response characteristics. By combining historical temperature and humidity cycle data, it can be determined whether the degradation is related to extreme conditions, pinpointing the specific time points of high-risk areas, and providing precise targets for subsequent processing.

[0078] When marking high-risk areas, automated tools can be used to generate marked data records. For example, for high-risk areas from month 12 to month 18, the tool will automatically mark this interval in the data sequence and add environmental condition labels, such as a high proportion of high-temperature and high-humidity periods. This marking method facilitates rapid location of problem areas during subsequent iterative processing and provides an intuitive reference for construction process analysis.

[0079] When initiating an iterative processing flow for marked data records, the construction process within the marked area can be analyzed layer by layer to identify optimization directions. For example, if the analysis reveals that the thickness of the flexible transition layer in a high-risk area is insufficient to cope with stress concentration caused by humidity fluctuations, the optimization direction might be to increase the layer thickness or improve the material mix ratio. The analysis process can be carried out in layers, starting with process parameters and then combining them with environmental response data to gradually clarify the key areas for improvement.

[0080] When updating the construction process version based on process optimization, an updated process configuration scheme can be generated. For example, if the original configuration has a flexible layer thickness of 0.4 mm, the updated configuration increases it to 0.5 mm, and the thermal expansion coefficient of the adhesive is adjusted to further reduce the risk of stress concentration. This update scheme needs to be combined with actual construction feasibility to ensure that the adjusted process can be effectively implemented on site.

[0081] When reloading trend forecast data to determine if the downward trend has eased, the effectiveness of process adjustments can be verified through the sequence analysis module. For example, after updating the process, forecast data shows that the retention rate decline from month 12 to month 18 has narrowed from 5% to 2%, indicating that the trend has eased. The final process adjustment results will be based on this verification to determine whether further optimization is needed. This verification mechanism helps ensure the targetedness and reliability of process improvements, extending the service life of material interfaces.

[0082] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0083] 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 equivalents.

Claims

1. A construction method for combining building waterproof membrane and thermal insulation layer, characterized in that, include: Historical environmental data and material property data of the joint between the waterproof membrane and the insulation layer are retrieved from the preset database. Temperature cycle records and moisture erosion records are separated from the historical environmental data. The temperature cycle records and moisture erosion records are systematically classified and organized to construct a preliminary environmental stress distribution model. Based on the preliminary environmental stress distribution model, the support vector machine algorithm is used to divide the boundary of the interface stability-related data points. When the boundary value in the division result exceeds the preset threshold, it is marked as a high-risk area, and the potential location of material degradation is accurately located. Specific index data of weak adhesion are extracted from high-risk areas, and the propagation path of layered cracking is simulated by gradient boosting decision tree algorithm, and finally a quantitative vector representation of the path is formed. Based on vector representation, if the peak data exceeds the preset threshold, a targeted construction process adjustment plan is retrieved from the preset process library, and the core parameters that determine the optimization of durability performance are selected from it. Using optimized parameters as input for building durability simulation, stress distribution simulation calculations are performed using the finite element analysis method to generate a leakage risk reduction vector. Each component of the leakage risk reduction vector is compared with a preset threshold. When all components are below the threshold, the energy-saving guarantee index data is integrated to establish the final interface stability enhancement configuration. Environmental stress adaptation adjustment values ​​are extracted from the final interface stability enhancement configuration. A recurrent neural network algorithm is applied to perform time series modeling on the adjustment values ​​to predict long-term durability performance trends and generate trend prediction sequences. Based on the trend prediction sequence, if a continuous downward trend is found to exceed the preset period, the process will be backtracked to the high-risk area marking stage to start the iterative optimization process and generate an updated version of the construction technology.

2. The construction method for combining building waterproof membrane and thermal insulation layer according to claim 1, characterized in that, The systematic classification and organization of temperature cycling records and moisture erosion records, and the construction of a preliminary environmental stress distribution model, include: Temperature cycling records are divided into high-temperature and low-temperature stages, and moisture erosion records are divided into high-humidity and low-humidity stages. Temperature cycling records are divided into high-temperature and low-temperature stages, and moisture erosion records are divided into high-humidity and low-humidity stages. The environmental stress changes at the junction are calculated at different stages based on the classified temperature cycling records and moisture erosion records. A linear regression model was used to establish the correspondence between environmental stress variation values ​​and material property data to obtain the environmental stress influence coefficient. Based on the environmental stress influence coefficient, the stress parameters in the preliminary environmental stress distribution model were corrected, and then the comprehensive stress distribution of the joint under the combined action of temperature cycling and moisture erosion was calculated.

3. The construction method for combining building waterproof membrane and thermal insulation layer according to claim 1, characterized in that, The method involves using a support vector machine algorithm to delineate boundaries for interface stability-related data points. When the boundary values ​​in the delineation result exceed a preset threshold, they are marked as high-risk areas, and the potential locations of material degradation are accurately located. This includes: When delineating boundaries, it is necessary to select a set of data points related to interface stability from a pre-set database and construct the dataset required for boundary delineation based on the key parameters in the environmental stress distribution model. After executing the support vector machine algorithm, the boundary values ​​are compared with the threshold. If the limit is exceeded, the data points in the high-risk area are re-labeled to lock the range of material degradation. Local environmental stress characteristics are extracted to determine potential location distribution. A correspondence table between material degradation and environmental stress is established by classification and sorting, and priority ranking is implemented. Based on the sorting results, a layered interface stability analysis layer is created, integrating multi-dimensional data to identify key monitoring areas.

4. The construction method for combining building waterproof membrane and thermal insulation layer according to claim 1, characterized in that, The process involves extracting specific index data on weak adhesion from high-risk areas, simulating the propagation path of layered cracking using a gradient boosting decision tree algorithm, and ultimately forming a quantized vector representation of the path, including: Material degradation-related indicator data were extracted from high-risk areas. A preliminary screening of indicator data with weak adhesion was conducted to form an indicator set. The gradient boosting decision tree algorithm was used to simulate the propagation path of layered cracking and generate an initial quantitative expression. In-depth analysis of key node data reveals the crack distribution pattern, and nodes exceeding the threshold are assigned priority labels; Layered analysis focuses on obtaining the cracking path details of each layer within the interval, classifying and storing the path quantification results, and constructing a correspondence table between layered cracking and weak adhesion.

5. The construction method for combining building waterproof membrane and thermal insulation layer according to claim 1, characterized in that, The process of retrieving targeted construction process adjustment plans from a preset process library and selecting the core parameters that determine durability performance optimization includes: Extract the peak data from the quantization vector and compare it with a preset threshold. If the peak data exceeds the preset threshold, retrieve the corresponding adjustment scheme from the process library, match the durability performance optimization parameter set, and then select the key subset. The random forest algorithm is used to predict the change vector of the impact of parameter adjustment on durability performance, and to determine the priority application sequence of adjustment schemes.

6. The construction method for combining building waterproof membrane and thermal insulation layer according to claim 1, characterized in that, The process of using optimized parameters as input for building durability simulation, combined with finite element analysis to perform stress distribution simulation calculations, and generating a leakage risk reduction vector includes: The set of durability performance optimization parameters is obtained from the adjustment plan record. The building durability simulation input data is determined. The stress distribution simulation calculation is performed on the building durability simulation input data using the finite element analysis method to obtain the leakage risk reduction vector. For the leakage risk reduction vector, extract the risk distribution peak data and verify whether it is lower than the preset threshold. If the risk distribution peak data is lower than the preset threshold, confirm that it is controllable, lock the priority solution and obtain the material strength improvement parameters.

7. The construction method for combining building waterproof membrane and thermal insulation layer according to claim 1, characterized in that, Each component of the leakage risk reduction vector is compared to a preset threshold. When all components are below the threshold, the energy-saving assurance index data is integrated to establish the final interface stability enhancement configuration, including: The data of each component is extracted from the leakage risk reduction vector and compared one by one. If each component is lower than the preset threshold, the leakage risk is judged to be within the controllable range. Obtain data on indicators related to energy conservation assurance, process the data, and determine the degree of compliance with energy conservation assurance requirements; Analyze potential issues related to interface stability and filter indicator data to identify key factors influencing interface stability; Obtain the corresponding enhanced configuration parameters, perform logical matching on the parameters, and if the matching result meets the preset conditions, determine the preliminary enhanced configuration scheme. Obtain resource allocation data related to configuration, verify the resource allocation data, determine whether the resource allocation meets the stability judgment requirements, obtain the final configuration determination basis, integrate the data based on the basis, and obtain the final solution for interface stability enhancement configuration.

8. The construction method for combining building waterproof membrane and thermal insulation layer according to claim 1, characterized in that, The process of extracting environmental stress adaptation adjustment values ​​from the final interface stability enhancement configuration, applying a recurrent neural network algorithm to perform time series modeling on the adjustment values, predicting long-term durability performance trends, and generating trend prediction sequences includes: Environmental stress adaptation adjustment values ​​are obtained from the final interface stability enhancement configuration, time series data is constructed, and a recurrent neural network algorithm is used to train the time series data to generate a durability performance trend prediction sequence. The trend prediction values ​​are extracted from the durability performance trend prediction sequence and the threshold range is verified. If the trend prediction values ​​are all within the preset range, the performance trend data is determined to be stable, the adaptive adjustment parameters are obtained, and the durability prediction output is generated.

9. The construction method for combining building waterproof membrane and thermal insulation layer according to claim 1, characterized in that, If, based on the trend prediction sequence, a continuous downward trend is found to exceed a preset period, the process is regressed to the high-risk area marking stage to initiate an iterative optimization process, producing an updated version of the construction technology, including: The trend prediction sequence is initially processed, and the data is scanned segment by segment using sequence analysis methods to determine whether there is a continuous downward pattern. If a continuous downward pattern is detected that exceeds the preset period, the relevant data segments are risk-assessed to determine the specific location of the high-risk area. The specific locations of high-risk areas are marked, and an iterative processing flow is initiated. The construction processes within the marked areas are analyzed layer by layer to obtain the direction of process optimization. The construction processes are then updated to generate updated process configuration schemes. For the updated process configuration scheme, the trend prediction data is reloaded through the sequence analysis module to determine whether the downward trend has been alleviated and to determine the final process adjustment result.

Images