Method and system for detecting performance of bin plate type thermal insulation material

Abstract

The invention discloses a bin plate type thermal insulation material performance detection method and system, and relates to the technical field of material detection, and the method comprises the following steps: extracting performance characteristic data; predicting the protective performance change trend of the warehouse plate type thermal insulation material under different environmental conditions based on the performance characteristic data; establishing an analysis space used for protection performance change trend abnormity identification, sampling and correcting performance data in the analysis space, and analyzing a relation between the performance data and a target threshold value and a historical sequence; and through abnormal path identification and curve fitting, key factors influencing the protection performance of the warehouse plate type thermal insulation material are identified, and a production process and a maintenance strategy are adjusted based on the key factors. Through anomaly recognition and curve fitting, performance anomaly and key factors thereof can be found in time, and data support is provided for optimizing a production process and a maintenance strategy, so that the stability and reliability of the material are improved, and the long-term and efficient thermal insulation effect of the material is ensured.

Description

A method and system for testing the performance of silo-type thermal insulation materials Technical Field

[0001] This invention relates to the field of materials testing technology, and more specifically, to a method and system for testing the performance of silo-type thermal insulation materials. Background Technology

[0002] With the increasing demand for high-performance insulation materials in the construction industry, performance testing of silo-type insulation materials has become a crucial step. These materials typically employ a sandwich panel structure, composed of a metal panel and an insulation core made of materials such as polyurethane or rock wool, offering excellent thermal insulation performance. To ensure their long-term effectiveness in buildings, comprehensive performance evaluation is essential, primarily focusing on indicators such as thermal conductivity, insulation performance, and humidity stability. The application of intelligent testing technology makes these performance evaluations more accurate and efficient. Through intelligent sensors and data analysis methods, intelligent testing technology can monitor the material's performance under different environmental conditions in real time, providing high-precision test results. This not only ensures the stability and reliability of the material in various environments but also verifies its compliance with green building standards.

[0003] However, current technologies for testing the performance of silo-type thermal insulation materials often rely on traditional static testing methods, lacking dynamic monitoring and data analysis of environmental changes. Therefore, it is impossible to predict changes in the material's protective performance under different conditions in real time, and it is also difficult to identify performance anomalies and key factors in a timely manner. This results in an inability to quickly adjust production processes or maintenance strategies, easily affecting the long-term stability and reliability of the material, and failing to meet the stringent requirements of green buildings for high-performance thermal insulation materials.

[0004] No effective solutions have yet been proposed to address the problems in the relevant technologies. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention proposes a method and system for testing the performance of silo-type thermal insulation materials. This solves the problem mentioned in the background section, where existing performance testing methods for silo-type thermal insulation materials often rely on traditional static testing methods, lacking dynamic monitoring and data analysis of environmental changes. Therefore, it is impossible to predict changes in the material's protective performance under different conditions in real time, and it is also difficult to identify performance anomalies and key factors in a timely manner. This leads to an inability to quickly adjust production processes or maintenance strategies, easily affecting the long-term stability and reliability of the material, and failing to meet the stringent requirements of green buildings for high-performance thermal insulation materials.

[0006] To achieve the above objectives, the present invention provides the following technical solution: According to one aspect of the present invention, a method for testing the performance of silo-type thermal insulation materials is provided, comprising: S1, acquiring performance data of silo-type thermal insulation materials under different environmental conditions, and preprocessing the performance data to extract performance characteristic data; S2, predicting the trend of protective performance changes of silo-type thermal insulation materials under different environmental conditions based on the performance characteristic data; S3, establishing an analysis space for identifying anomalies in the trend of protective performance changes, sampling and correcting performance data in the analysis space, analyzing the relationship between performance data and target thresholds and historical sequences; and identifying key factors affecting the protective performance of silo-type thermal insulation materials through anomaly path identification and curve fitting, and adjusting production processes and maintenance strategies based on the key factors.

[0007] Furthermore, performance data of the silo-type thermal insulation material under different environmental conditions are acquired, and the performance data is preprocessed to extract performance feature data, including: data filtering of the performance data of the silo-type thermal insulation material under different environmental conditions to eliminate interference information; segmentation and analysis of the performance data using threshold segmentation to obtain preliminary segmentation results, and calculation of the total energy index of the material performance to form an initial total energy index; by updating the performance segmentation threshold and re-segmenting the performance data, the total energy index of the new segmented region is calculated to optimize the data processing process; the updated total energy index is compared with the initial total energy index, and if the new energy index value is less than the initial total energy index, the performance segmentation threshold is updated and the performance data is re-segmented until the total energy index no longer changes; data smoothing and denoising methods are used to optimize the segmented performance data and extract performance feature data.

[0008] Furthermore, the prediction of the protective performance change trend of silo insulation materials under different environmental conditions based on performance characteristic data includes: randomly generating weight matrices and bias matrices of input environmental factors based on the performance characteristic data of silo insulation materials to provide initial values ​​for constructing a performance prediction model; setting the number of hidden layer nodes as the initial value, calculating the hidden layer output matrix, and calculating the preliminary learning accuracy based on the protective performance predicted by the performance prediction model and the actual data to form a preliminary prediction effect; gradually increasing the number of hidden layer nodes, recalculating the hidden layer output matrix and the weights of the performance prediction model, optimizing the performance prediction model, and obtaining the optimal number of hidden layer nodes; using a decomposition algorithm to extract the initial equivalent weight matrix for the performance prediction of silo insulation materials, providing initial data input for the optimized performance prediction model; combining the optimal number of hidden layer nodes and the initial equivalent weight matrix, using the model formula to calculate the protective performance weights, and optimizing the performance prediction results; iteratively optimizing the weight parameters of the performance prediction model until the maximum number of iterations is reached, and finally outputting the protective performance change trend of silo insulation materials under different environmental conditions.

[0009] Furthermore, the initial equivalent weight matrix for predicting the performance of silo-type thermal insulation materials is extracted using a decomposition algorithm. This provides initial data input for the optimized performance prediction model. The process includes: randomly selecting a feature vector from the performance feature data matrix of the silo-type thermal insulation materials as the initial feature vector and recording its position index; defining the set of positions of unselected feature vectors as the remaining feature set; for each feature vector in the remaining feature set, calculating the projection of each feature vector onto the currently selected feature vector, and updating the correlation value between each feature vector and the selected feature vector; selecting the feature vector corresponding to the maximum projection value based on the projection value, updating the current feature vector set, and adding this feature vector to the selected feature vector set; iteratively updating the feature vector set until the maximum number of iterations is reached; using a filtering algorithm to verify the selected feature vector combination, and selecting the optimal feature vector combination as the initial data input for the performance prediction model.

[0010] Furthermore, the selected feature vector combinations are validated using a screening algorithm to select the optimal feature vector combination as the initial data input for the performance prediction model. This includes: initializing the number of feature vector combinations, setting the parameters of the screening algorithm, and determining the maximum number of iterations; in the main loop, if the random value is less than the preset dynamic transformation probability, different feature vector combinations are screened; otherwise, the same feature vector combination is screened, and the position and relevance of the feature vector combination are updated; the screened feature vector combinations are optimized using the golden sine guide operation, and their fitness values ​​are recalculated; based on the new fitness value, it is determined whether to update the feature vector combination. If the fitness of the new feature vector combination is the highest, the feature vector combination is updated, and the fitness of the new feature vector combination is compared with the current optimal feature vector combination. The feature vector combination with the highest fitness is selected as the optimal feature vector combination; it is checked whether the maximum number of iterations has been reached. If it has, the optimal feature vector combination is output as the initial data input for the performance prediction model; otherwise, iterative optimization continues.

[0011] Furthermore, an analysis space is established for identifying anomalies in the protective performance trend. Performance data is sampled and corrected within this analysis space, and the relationship between performance data and target thresholds and historical sequences is analyzed. Through anomaly path identification and curve fitting, key factors affecting the protective performance of silo-type insulation materials are identified. Based on these key factors, production processes and maintenance strategies are adjusted, including: setting a baseline state and target performance threshold for protective performance; establishing an analysis space for identifying anomalies in the protective performance trend by combining the initial analysis step size and neighborhood range; randomly sampling performance data within the analysis space; finding reference points that meet the target performance threshold; and adjusting the production process and maintenance strategies based on these reference points and the target performance threshold. The system checks the deviation between performance thresholds to determine if the current analysis step size needs adjustment. If the deviation exceeds the set range, an adjustment algorithm is introduced to correct the performance data and generate new performance points to enhance the accuracy of trend fitting. The system checks the consistency between the newly generated performance points and historical data. If there are no conflicts, the data is included in the analysis sequence; otherwise, correction points are generated again through weight adjustment. The system detects the distance between the trend sequence and the target threshold. If the preset requirements are met, anomaly paths are constructed and key nodes are traced back. Anomaly trajectories are extracted through pruning and curve fitting to identify key factors affecting the protective performance of silo-type thermal insulation materials. Production processes and maintenance strategies are then adjusted based on these key factors.

[0012] Furthermore, if the deviation exceeds the set range, an adjustment algorithm is introduced to correct the performance data and generate new performance points to enhance the accuracy of trend fitting. This includes: setting the parameters and maximum number of iterations of the adjustment algorithm; randomly generating an initial set of performance data points, where the initial position of each performance data point represents a performance value, and the speed reflects the magnitude of the correction, serving as the starting point of the optimization process; calculating the fitness of the initial set of performance data points, evaluating the deviation between each performance data point and the target threshold, and simultaneously initializing the individual optimal value and the global optimal value to prepare for the next round of optimization; determining whether to enter the crossover and mutation process of performance data points based on the set probability; if the random value is greater than the threshold, performing crossover and mutation operations; otherwise, updating the position of the performance data points; grouping the performance data points according to fitness, placing the best-performing performance data points into the optimal solution group for crossover operations, and placing the best-performing performance data points into the inferior solution group for mutation; merging the offspring generated by the optimal and inferior solution groups, establishing an expanded dataset with the current set of performance data points, sorting them by fitness, selecting the performance data points with the best fitness as the elite group, and continuing the next round of optimization to generate new performance points to enhance the accuracy of trend fitting.

[0013] Furthermore, the offspring generated from the optimal and unoptimized solution groups are merged to create an expanded dataset with the current performance data point set. This expanded dataset is then sorted by fitness, and the performance data points with the best fitness are selected as the elite group for the next round of optimization, generating new performance points to enhance the accuracy of trend fitting. This process includes: merging the offspring generated from the optimal and unoptimized solution groups to form a new dataset containing the current performance data point set and newly generated performance data points; sorting the merged dataset by fitness, evaluating the deviation of each performance data point from the target threshold, and identifying the best-performing performance data point; selecting the best-fit performance data point from the sorted dataset to form the elite group; and then, based on the elite group, continuing the next round of optimization to generate new performance points to enhance the accuracy of trend fitting.

[0014] Furthermore, the formula for evaluating the deviation of each performance data point from the target threshold is as follows: In the formula, M a ω represents the weighted absolute error between the a-th performance data point and the target threshold T; m represents the total number of performance data points; a α represents the weighting coefficient for the a-th performance data point; α represents the adjustment coefficient for the mean squared error; β represents the adjustment coefficient for the absolute error; h a represents the observed value of the a-th performance data point; T represents the target threshold.

[0015] According to another aspect of the present invention, a performance testing system for silo-type thermal insulation materials is also provided. The system includes: a data acquisition module for acquiring performance data of the silo-type thermal insulation material under different environmental conditions, preprocessing the performance data, and extracting performance characteristic data; a trend prediction module for predicting the trend of protective performance changes of the silo-type thermal insulation material under different environmental conditions based on the performance characteristic data; and a factor analysis and optimization module for establishing an analysis space for identifying anomalies in the protective performance change trend, sampling and correcting performance data in the analysis space, analyzing the relationship between performance data and target thresholds and historical sequences, and identifying key factors affecting the protective performance of the silo-type thermal insulation material through anomaly path identification and curve fitting, and adjusting production processes and maintenance strategies based on these key factors.

[0016] The beneficial effects of this invention are as follows: 1. This invention can achieve accurate monitoring and prediction of the performance of silo-type thermal insulation materials under different environmental conditions. Preprocessing and feature extraction of performance data improve data reliability, and the established prediction model can accurately predict the changing trend of protective performance, providing a basis for production and maintenance. Through anomaly identification and curve fitting, performance anomalies and their key factors can be discovered in a timely manner, providing data support for optimizing production processes and maintenance strategies, thereby improving the stability and reliability of the material and ensuring its long-term, efficient thermal insulation effect.

[0017] 2. This invention, by constructing a performance prediction model, can dynamically assess the changing trends of material protective performance under different environmental conditions and continuously improve the model's accuracy through iterative optimization. Furthermore, the screening algorithm and feature vector optimization improve the quality of the model's input data, ensuring more reliable prediction results. This effectively guides material production and maintenance, ensuring the stability and reliability of materials in practical applications.

[0018] 3. This invention introduces anomaly identification and adjustment algorithms to enable real-time monitoring of the protective performance trends of silo-type thermal insulation materials, identifying and correcting deviations to ensure accurate predictions. By optimizing performance data points and improving trend fitting, key factors can be accurately extracted, providing effective basis for production processes and maintenance strategies. Furthermore, by utilizing optimization operations such as cross-validation and mutation, the data fitting degree is continuously improved, enhancing the robustness and accuracy of the model, thereby ensuring the stability and reliability of the material in practical applications. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 is a flowchart of a method for testing the performance of silo-type thermal insulation materials according to an embodiment of the present invention; Figure 2 is a schematic block diagram of a system for testing the performance of silo-type thermal insulation materials according to an embodiment of the present invention.

[0021] In the diagram: 1. Data acquisition module; 2. Trend prediction module; 3. Factor analysis and optimization module. Detailed Implementation

[0022] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0023] In the description of this invention, unless otherwise stated, "a plurality of" means two or more. Furthermore, the terms "first," "second," "third," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0024] According to an embodiment of the present invention, a method and system for testing the performance of silo-type thermal insulation materials are provided.

[0025] The present invention will now be further described with reference to the accompanying drawings and specific embodiments. As shown in Figure 1, the performance testing method for silo-type thermal insulation materials according to an embodiment of the present invention includes: S1, acquiring performance data of silo-type thermal insulation materials under different environmental conditions, and preprocessing the performance data to extract performance characteristic data; specifically, various sensors can be used to acquire performance data of silo-type thermal insulation materials under different environmental conditions.

[0026] Specifically, different environmental conditions include temperature conditions (high temperature, low temperature, temperature fluctuation, continuous constant temperature environment, etc.), humidity conditions (high humidity environment, dry environment, environment with frequent changes in relative humidity, etc.), air flow and wind speed (still air, weak wind flow, strong wind flow, etc.), radiation and light (direct sunlight, ultraviolet radiation intensity, long-term lack of light, etc.), load and pressure (normal atmospheric pressure, high pressure / low pressure storage environment, stress state, etc.), chemically corrosive environment (acid and alkali vapors, salt spray environment, ordinary neutral environment, etc.), aging conditions (performance evolution under long-term high temperature / humidity / ultraviolet exposure, short-term use / long-term use, etc.), etc.

[0027] Specifically, performance data includes thermal performance data, humidity-related data, mechanical performance data, durability data, and environmental response data.

[0028] Specifically, performance characteristic data includes thermal characteristic parameters, humidity sensitivity index, mechanical characteristic parameters, time-varying characteristic parameters, and environmental adaptability index.

[0029] S2. Based on performance characteristic data, predict the trend of protective performance changes of silo-type thermal insulation materials under different environmental conditions; specifically, the trend of protective performance changes refers to the performance and trend of the thermal insulation and other protective functions of silo-type thermal insulation materials under different environmental conditions as time or environmental conditions change.

[0030] Specifically, the trends in protective performance include changes in thermal conductivity, thermal insulation performance, humidity stability, mechanical properties, aging performance, thermal expansion and dimensional changes, stress and deformation, and durability.

[0031] S3. Establish an analysis space for identifying abnormal trends in protective performance. Sample and correct performance data in the analysis space, and analyze the relationship between performance data and target thresholds and historical sequences. Through abnormal path identification and curve fitting, identify key factors affecting the protective performance of silo-type thermal insulation materials, and adjust production processes and maintenance strategies based on these key factors.

[0032] Specifically, key factors include ambient temperature, humidity, material aging degree, mechanical stress, ultraviolet radiation, rate of temperature and humidity change, initial properties of the material, chemical reactions and corrosion, thermal expansion and contraction, and material manufacturing process.

[0033] In this optional embodiment, acquiring performance data of the silo-type thermal insulation material under different environmental conditions and preprocessing the performance data to extract performance feature data includes: filtering the performance data of the silo-type thermal insulation material acquired under different environmental conditions to eliminate interference information; segmenting and analyzing the performance data using threshold segmentation to obtain preliminary segmentation results and calculating the total energy index of the material performance to form an initial total energy index; optimizing the data processing process by updating the performance segmentation threshold and re-segmenting the performance data; comparing the updated total energy index with the initial total energy index, and if the new energy index value is less than the initial total energy index, continuing to update the performance segmentation threshold and re-segment the performance data until the total energy index no longer changes; and optimizing the segmented performance data using data smoothing and denoising methods to extract performance feature data.

[0034] In this optional embodiment, predicting the trend of protective performance changes of silo-type thermal insulation materials under different environmental conditions based on performance characteristic data includes: randomly generating a weight matrix and bias matrix of input environmental factors based on the performance characteristic data of silo-type thermal insulation materials to provide initial values ​​for constructing a performance prediction model; setting the number of hidden layer nodes as the initial value, calculating the hidden layer output matrix, and calculating the preliminary learning accuracy based on the protective performance predicted by the performance prediction model and the actual data to form a preliminary prediction effect; gradually increasing the number of hidden layer nodes, recalculating the hidden layer output matrix and the weights of the performance prediction model, optimizing the performance prediction model, and obtaining the optimal number of hidden layer nodes; using a decomposition algorithm to extract the initial equivalent weight matrix for the performance prediction of silo-type thermal insulation materials to provide initial data input for the optimized performance prediction model; combining the optimal number of hidden layer nodes and the initial equivalent weight matrix, using the model formula to calculate the protective performance weights and optimize the performance prediction results; iteratively optimizing the weight parameters of the performance prediction model until the maximum number of iterations is reached, and finally outputting the trend of protective performance changes of silo-type thermal insulation materials under different environmental conditions.

[0035] Specifically, a performance prediction model is constructed by randomly generating weight and bias matrices of environmental factors as initial input data. The number of hidden layer nodes is set as the initial value, the hidden layer output matrix is ​​calculated, and the model's predicted protective performance is compared with actual data to evaluate the initial learning accuracy and obtain a preliminary assessment of the prediction effect. The number of hidden layer nodes is gradually increased, and the hidden layer output matrix and weights are recalculated to optimize the performance prediction model until the optimal number of hidden layer nodes is found. An initial equivalent weight matrix is ​​extracted using a decomposition algorithm to provide data input for the optimized model. Combined with the optimal number of hidden layer nodes, the protective performance weights are calculated using the model formula to further optimize the prediction results. Based on this, the model's weight parameters are iteratively optimized until the maximum number of iterations is reached. The model can accurately predict the performance of silo-type thermal insulation materials under different environmental conditions, thereby achieving dynamic monitoring and optimization of material performance.

[0036] In this optional embodiment, the initial equivalent weight matrix for predicting the performance of silo-type thermal insulation materials is extracted using a decomposition algorithm. This provides initial data input for the optimized performance prediction model. The process includes: randomly selecting a feature vector from the performance feature data matrix of the silo-type thermal insulation materials as the initial feature vector and recording its position index; defining the set of positions of unselected feature vectors as the remaining feature set; for each feature vector in the remaining feature set, calculating the projection of each feature vector onto the currently selected feature vector, and updating the correlation value between each feature vector and the selected feature vector; selecting the feature vector corresponding to the maximum projection value based on the projection value, updating the current feature vector set, and adding this feature vector to the selected feature vector set; iteratively updating the feature vector set until the maximum number of iterations is reached; using a filtering algorithm to verify the selected feature vector combination; and selecting the optimal feature vector combination as the initial data input for the performance prediction model.

[0037] Specifically, the decomposition algorithm is a continuous projection algorithm, a dimensionality reduction technique based on projection. In this invention, it is used to extract an initial equivalent weight matrix from the performance characteristic data matrix of the silo insulation material. By randomly selecting an initial eigenvector, its projection onto other eigenvectors is calculated, and the correlation between them is updated. Based on the projection values, the most correlated eigenvector is selected, and the eigenvector set is updated progressively. Finally, the optimal eigenvector combination is selected as the initial data input for the performance prediction model, thereby optimizing the model's performance.

[0038] In this optional embodiment, the selection of the optimal feature vector combination as the initial data input for the performance prediction model is verified using a screening algorithm. This includes: initializing the number of feature vector combinations, setting the parameters of the screening algorithm, and determining the maximum number of iterations; in the main loop, if the random value is less than the preset dynamic transformation probability, different feature vector combinations are screened; otherwise, the same feature vector combination is screened, and the position and correlation of the feature vector combination are updated; the screened feature vector combinations are optimized using a golden sine guide operation, and their fitness values ​​are recalculated; based on the new fitness value, it is determined whether to update the feature vector combination. If the fitness of the new feature vector combination is the highest, the feature vector combination is updated, and the fitness of the new feature vector combination is compared with the current optimal feature vector combination. The feature vector combination with the highest fitness is selected as the optimal feature vector combination; it is checked whether the maximum number of iterations has been reached. If it has, the optimal feature vector combination is output as the initial data input for the performance prediction model; otherwise, iterative optimization continues.

[0039] Specifically, the screening algorithm is a hybrid pollen algorithm, which is an optimization algorithm that simulates pollen dispersal and search behavior. In this invention, this algorithm is used to screen feature vector combinations. By randomly selecting and updating feature vectors in the feature space, the pollen dispersal process is simulated. The selected feature vector combinations are optimized using a golden sine guide operation, the fitness is calculated, and the combinations are updated step by step until the feature vector combination with the best fitness is found as the initial data input for the performance prediction model, ensuring that the selection of feature vectors is optimal.

[0040] In this optional embodiment, an analysis space is established for identifying anomalies in the trend of protective performance changes. Performance data is sampled and corrected within this analysis space, and the relationship between the performance data and the target threshold and historical sequences is analyzed. Through anomaly path identification and curve fitting, key factors affecting the protective performance of the silo-type insulation material are identified. Adjustments to the production process and maintenance strategies based on these key factors include: setting a baseline state and target performance threshold for protective performance; establishing an analysis space for identifying anomalies in the trend of protective performance changes by combining an initial analysis step size and neighborhood range; randomly sampling performance data within the analysis space; finding reference points that meet the target performance threshold; and adjusting the production process and maintenance strategies based on these reference points and the target performance threshold. The deviation between the target performance threshold and the target threshold is used to determine whether the current analysis step size needs to be adjusted. If the deviation exceeds the set range, an adjustment algorithm is introduced to correct the performance data and generate new performance points to enhance the accuracy of trend fitting. The consistency between the newly generated performance points and historical data is checked. If there is no conflict, the data is included in the analysis sequence; otherwise, correction points are generated again through weight adjustment. The distance between the trend sequence and the target threshold is detected. If the preset requirements are met, anomaly paths are constructed and key nodes are traced back. Anomaly trajectories are extracted through pruning and curve fitting to identify key factors affecting the protective performance of silo insulation materials. Production processes and maintenance strategies are adjusted based on these key factors.

[0041] Specifically, a baseline state and target performance threshold for protective performance are set, and an analysis space for identifying anomalies in protective performance trends is established by combining the initial analysis step size and neighborhood range. Performance data is randomly sampled within this analysis space to find reference points that meet the target performance threshold. The deviation between the reference point and the target threshold is used to determine whether the analysis step size needs adjustment. If the deviation exceeds a set range, an adjustment algorithm is used to correct the performance data, generating new performance points to enhance the accuracy of trend fitting. The consistency between the newly generated performance points and historical data is checked. If there are no conflicts, they are included in the analysis sequence; otherwise, correction points are generated again through weight adjustment. The distance between the trend sequence and the target threshold is further examined. If the preset requirements are met, an anomaly path is constructed, and key nodes are traced back. Anomaly trajectories are extracted through pruning and curve fitting to identify key factors affecting protective performance. Based on the identified key factors, production processes and maintenance strategies are adjusted to ensure that the material's protective performance remains stable under different environments.

[0042] Specifically, the algorithm used to identify key factors affecting the protective performance of silo-type thermal insulation materials is the Rapid Expanding Random Tree Algorithm (RRT algorithm), a tree-based path planning algorithm. In this invention, the RRT algorithm is used to identify key factors affecting the protective performance of silo-type thermal insulation materials. By establishing a random tree in the analysis space, the algorithm expands from an initial point, gradually exploring the relationship between performance data, target thresholds, and historical sequences. During each expansion, the RRT algorithm quickly finds key factors and backtracks abnormal trajectories through pruning and path optimization. This helps identify factors affecting the material's protective performance and provides data support for adjusting production processes and maintenance strategies.

[0043] In this optional embodiment, if the deviation exceeds a set range, an adjustment algorithm is introduced to correct the performance data and generate new performance points to enhance the accuracy of trend fitting. This includes: setting the parameters and maximum number of iterations of the adjustment algorithm; randomly generating an initial set of performance data points, where the initial position of each performance data point represents a performance value, and the speed reflects the magnitude of the correction, serving as the starting point of the optimization process; calculating the fitness of the initial set of performance data points, evaluating the deviation between each performance data point and the target threshold, and simultaneously initializing the individual optimal value and the global optimal value to prepare for the next round of optimization; determining whether to enter the crossover and mutation process of performance data points based on a set probability; if the random value is greater than the threshold, performing crossover and mutation operations; otherwise, updating the position of the performance data points; grouping the performance data points according to fitness, placing the best-performing performance data points into the optimal solution group for crossover operations, and placing the best-performing performance data points into the inferior solution group for mutation operations; merging the offspring generated by the optimal and inferior solution groups, establishing an expanded dataset with the current set of performance data points, sorting them by fitness, selecting the performance data points with the best fitness as the elite group, and continuing the next round of optimization to generate new performance points to enhance the accuracy of trend fitting.

[0044] Specifically, the algorithm is adjusted to a gravitational search algorithm, an optimization algorithm that simulates the gravitational forces and motion of celestial bodies. In this invention, this algorithm is used to correct performance data points by guiding them towards the global optimum through gravity, thereby optimizing the performance data and enhancing the accuracy of trend fitting. By adjusting the position and fitness of the data points, the algorithm gradually approaches the optimal performance point, improving the accuracy of the prediction model.

[0045] In this optional embodiment, merging the offspring generated from the optimal and unoptimized solution groups to create an expanded dataset with the current performance data point set, sorting by fitness, and selecting the performance data point with the best fitness as the elite group, continues the next round of optimization to generate new performance points to enhance the accuracy of trend fitting. This includes: merging the offspring generated from the optimal and unoptimized solution groups to form a new dataset containing the current performance data point set and newly generated performance data points; sorting the merged dataset by fitness, evaluating the deviation of each performance data point from the target threshold, and identifying the best-performing performance data point; selecting the performance data point with the best fitness from the sorted dataset to form the elite group; and based on the elite group, continuing the next round of optimization to generate new performance points to enhance the accuracy of trend fitting.

[0046] In this optional embodiment, the formula for evaluating the deviation of each performance data point from the target threshold is: In the formula, M a ω represents the weighted absolute error between the a-th performance data point and the target threshold T; m represents the total number of performance data points; a α represents the weighting coefficient for the a-th performance data point; α represents the adjustment coefficient for the mean squared error; β represents the adjustment coefficient for the absolute error; h a represents the observed value of the a-th performance data point; T represents the target threshold.

[0047] According to another embodiment of the present invention, as shown in FIG2, a performance testing system for silo-type thermal insulation materials is also provided. The system includes: a data acquisition module 1, used to acquire performance data of silo-type thermal insulation materials under different environmental conditions, and preprocess the performance data to extract performance characteristic data; a trend prediction module 2, used to predict the trend of protective performance change of silo-type thermal insulation materials under different environmental conditions based on the performance characteristic data; and a factor analysis and optimization module 3, used to establish an analysis space for identifying anomalies in the trend of protective performance change, sample and correct performance data in the analysis space, analyze the relationship between performance data and target thresholds and historical sequences, and identify key factors affecting the protective performance of silo-type thermal insulation materials through anomaly path identification and curve fitting, and adjust production processes and maintenance strategies based on key factors.

[0048] The data acquisition module 1 is connected to the trend prediction module 2 and the factor analysis and optimization module 3.

[0049] To facilitate understanding of the above technical solutions of the present invention, the following provides a detailed description of the performance testing of the silo-type thermal insulation material in actual practice.

[0050] I. Obtaining performance data and extracting performance characteristic data of silo-type thermal insulation materials: 1) Using multiple sensors to obtain performance data of silo-type thermal insulation materials under different environmental conditions.

[0051] Thermal conductivity is obtained in real time using a heat flow meter (e.g., a hot-wire sensor). For example, the thermal conductivity measured in the experimental environment is shown in Table 1: Table 1 Thermal Conductivity The moisture absorption rate of the material under different humidity conditions is obtained using a humidity sensor. For example, the moisture absorption rate data is shown in Table 2: Table 2 Moisture Absorption Rate Data The compressive strength of materials under different loading conditions was tested using a universal testing machine. For example, the compressive strength data of the materials is shown in Table 3: Table 3 Compressive Strength Data of Materials The aging degree of materials is measured through long-term exposure tests (e.g., thermo-oxidative aging). For example, the test cycle is 1000 hours, and the changes in compressive strength before and after aging are recorded as shown in Table 4: Table 4 Changes in Compressive Strength The material's response under different environments was recorded using temperature and humidity sensors and ultraviolet sensors.

[0052] 2) Extract feature data from the collected data: thermal characteristic parameters: such as the thermal conductivity and insulation performance coefficient of the material at different temperatures.

[0053] Humidity sensitivity indicators: the curve of moisture absorption rate changing with humidity, and the effect of humidity on thermal conductivity.

[0054] Mechanical characteristic parameter: the trend of compressive strength changing with temperature.

[0055] Time-varying characteristic parameters: the performance changes of materials at different test time points.

[0056] Environmental adaptability index: The effect of ultraviolet radiation on the mechanical properties of materials.

[0057] II. Establish a performance prediction model and predict the trend of protective performance changes: 1) Use the extracted performance feature data to establish a regression model or machine learning model (such as support vector regression model, neural network model or extreme learning machine ELM, etc.) to predict the trend of protective performance changes of silo insulation materials under different environmental conditions.

[0058] For example, if the trend of thermal conductivity change is chosen as the prediction target, the relationship between thermal conductivity and temperature can be obtained by model fitting as follows: thermal conductivity = 0.0008P + 0.022, where P represents temperature.

[0059] 2) For example, the model predicts the trend of thermal conductivity changes at different temperatures as shown in Table 5.

[0060] Table 5. Model predictions of thermal conductivity at different temperatures. 3) Other protective performance trends include: thermal insulation performance trends: for example, model calculations show that thermal insulation performance gradually decreases as temperature increases.

[0061] Humidity stability trend: The model may predict that in high humidity environments, the humidity sensitivity of insulation materials will be significantly enhanced, leading to a gradual decline in thermal insulation performance.

[0062] Mechanical property change trend: It is predicted that high temperature will lead to a decrease in compressive strength.

[0063] III. Identification of abnormal trends in protective performance and analysis of key factors: 1) Establishment and sampling of analysis space: By establishing an analysis space for identifying abnormal trends in protective performance (with thermal conductivity change and temperature and humidity as two dimensions), sampling is performed in the analysis space to obtain new performance data.

[0064] Target threshold: For example, the specified target thermal conductivity threshold is 0.045 W / m·K. In actual testing, if the thermal conductivity exceeds this threshold, it indicates that the thermal insulation performance of the material may not meet the requirements.

[0065] Analysis of the samples revealed that the thermal conductivity was 0.050 W / m·K at 50°C, exceeding the set target threshold, indicating that the protective performance of the material decreased at this temperature.

[0066] 2) Abnormal path identification and correction: Abnormal path identification identifies thermal conductivity data that exceeds the threshold, marks them as abnormal data points, and corrects them.

[0067] Adjustment algorithm: The model weights are optimized using an adjustment algorithm (such as genetic algorithm, particle swarm optimization algorithm or gravitational search algorithm) to calculate the corrected thermal conductivity.

[0068] For example, after correction, at a temperature of 50°C, the thermal conductivity data was adjusted to 0.0455 W / m·K, which is close to the target threshold.

[0069] 3) Key Factor Analysis and Adjustment: Based on the sampling and correction results, the following factors were found to have a significant impact on protective performance: Temperature: Increased temperature leads to a significant increase in thermal conductivity.

[0070] Humidity: When the humidity is too high, the moisture absorption rate of the material increases, which leads to an increase in thermal conductivity.

[0071] Aging: Material aging leads to a gradual increase in thermal conductivity and a gradual decrease in compressive strength.

[0072] Based on these key factors, adjust production processes and maintenance strategies: optimize material density and structure during production to reduce the impact of temperature on thermal conductivity.

[0073] Enhance the waterproof performance of materials to reduce the impact of humidity on performance.

[0074] Conduct aging tests in advance and adjust the material's service and maintenance cycles based on the test results.

[0075] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for testing the performance of silo-type thermal insulation materials, characterized in that, include: S1. Obtain performance data of silo-type thermal insulation material under different environmental conditions, and preprocess the performance data to extract performance characteristic data; S2. Predict the changing trend of protective performance of silo-type thermal insulation materials under different environmental conditions based on performance characteristic data; S3. Establish an analysis space for identifying abnormal trends in protection performance, sample and correct performance data in the analysis space, and analyze the relationship between performance data and target thresholds and historical sequences. By identifying abnormal paths and curve fitting, key factors affecting the protective performance of silo-type thermal insulation materials are identified, and production processes and maintenance strategies are adjusted based on these key factors.

2. The method for testing the performance of silo-type thermal insulation materials according to claim 1, characterized in that, The process of acquiring performance data of silo-type thermal insulation materials under different environmental conditions and preprocessing the performance data to extract performance feature data includes: filtering the performance data of silo-type thermal insulation materials acquired under different environmental conditions to eliminate interference information; segmenting and analyzing the performance data using threshold segmentation to obtain preliminary segmentation results and calculating the total energy index of the material performance to form an initial total energy index; optimizing the data processing process by updating the performance segmentation threshold and re-segmenting the performance data; comparing the updated total energy index with the initial total energy index, and if the new energy index value is less than the initial total energy index, continuing to update the performance segmentation threshold and re-segment the performance data until the total energy index no longer changes; and optimizing the segmented performance data using data smoothing and denoising methods to extract performance feature data.

3. The method for testing the performance of silo-type thermal insulation materials according to claim 1, characterized in that, The method for predicting the protective performance variation trend of silo-type thermal insulation materials under different environmental conditions based on performance characteristic data includes: randomly generating a weight matrix and bias matrix of input environmental factors based on the performance characteristic data of silo-type thermal insulation materials to provide initial values ​​for constructing a performance prediction model; setting the number of hidden layer nodes as the initial value, calculating the hidden layer output matrix, and calculating the preliminary learning accuracy based on the protective performance predicted by the performance prediction model and the actual data to form a preliminary prediction effect; gradually increasing the number of hidden layer nodes, recalculating the hidden layer output matrix and the weights of the performance prediction model, optimizing the performance prediction model, and obtaining the optimal number of hidden layer nodes; using a decomposition algorithm to extract the initial equivalent weight matrix for the performance prediction of silo-type thermal insulation materials to provide initial data input for the optimized performance prediction model; combining the optimal number of hidden layer nodes and the initial equivalent weight matrix, using the model formula to calculate the protective performance weights and optimize the performance prediction results; iteratively optimizing the weight parameters of the performance prediction model until the maximum number of iterations is reached, and finally outputting the protective performance variation trend of silo-type thermal insulation materials under different environmental conditions.

4. The method for testing the performance of silo-type thermal insulation materials according to claim 3, characterized in that, The process of extracting the initial equivalent weight matrix for predicting the performance of silo-type thermal insulation materials using a decomposition algorithm, and providing initial data input for the optimized performance prediction model, includes: randomly selecting a feature vector from the performance feature data matrix of the silo-type thermal insulation materials as the initial feature vector, and recording the position index of the initial feature vector; defining the set of positions of the unselected feature vectors as the remaining feature set; for each feature vector in the remaining feature set, calculating the projection of each feature vector onto the currently selected feature vector, and updating the correlation value between each feature vector and the selected feature vector; selecting the feature vector corresponding to the maximum projection value based on the magnitude of the projection value, updating the current feature vector set, and adding the feature vector to the selected feature vectors; iteratively updating the feature vector set until the maximum number of iterations is reached; using a filtering algorithm to verify the selected feature vector combination, and selecting the optimal feature vector combination as the initial data input for the performance prediction model.

5. The method for testing the performance of silo-type thermal insulation materials according to claim 4, characterized in that, The process of using a filtering algorithm to verify selected feature vector combinations and selecting the optimal feature vector combination as the initial data input for the performance prediction model includes: initializing the number of feature vector combinations, setting the parameters of the filtering algorithm, and determining the maximum number of iterations; in the main loop, if the random value is less than the preset dynamic transformation probability, different feature vector combinations are filtered; otherwise, the same feature vector combination is filtered, and the position and correlation of the feature vector combination are updated; the filtered feature vector combinations are optimized using a golden sine guide operation, and their fitness values ​​are recalculated; based on the new fitness value, it is determined whether to update the feature vector combination; if the fitness of the new feature vector combination is the highest, the feature vector combination is updated, and the fitness of the new feature vector combination is compared with the current optimal feature vector combination, and the feature vector combination with the highest fitness is selected as the optimal feature vector combination; it is checked whether the maximum number of iterations has been reached; if it has, the optimal feature vector combination is output as the initial data input for the performance prediction model; otherwise, iterative optimization continues.

6. The method for testing the performance of silo-type thermal insulation materials according to claim 1, characterized in that, The analysis space is established for identifying abnormal trends in protection performance. Performance data is sampled and corrected in the analysis space, and the relationship between performance data and target thresholds and historical sequences is analyzed. By identifying abnormal paths and curve fitting, key factors affecting the protective performance of silo-type thermal insulation materials are identified. Adjustments to production processes and maintenance strategies based on these key factors include: setting a baseline state and target performance threshold for protective performance; establishing an analysis space for identifying abnormal trends in protective performance by combining the initial analysis step size and neighborhood range; randomly sampling performance data within the analysis space to find reference points that meet the target performance threshold; determining whether to adjust the current analysis step size based on the deviation between the reference point and the target performance threshold; if the deviation exceeds a set range, introducing an adjustment algorithm to correct the performance data and generate new performance points to enhance the accuracy of trend fitting; verifying the consistency between the newly generated performance points and historical data; if there are no conflicts, they are included in the analysis sequence; otherwise, correction points are generated again through weight adjustment; detecting the distance between the trend sequence and the target threshold; if the preset requirements are met, an abnormal path is constructed and key nodes are traced back; and abnormal trajectories are extracted through pruning and curve fitting to identify key factors affecting the protective performance of silo-type thermal insulation materials, and production processes and maintenance strategies are adjusted based on these key factors.

7. The method for testing the performance of silo-type thermal insulation materials according to claim 6, characterized in that, If the deviation exceeds the set range, an adjustment algorithm is introduced to correct the performance data and generate new performance points to enhance the accuracy of trend fitting. This includes: setting the parameters and maximum number of iterations of the adjustment algorithm; randomly generating an initial set of performance data points, where the initial position of each performance data point represents a performance value, and the speed reflects the magnitude of the correction, serving as the starting point of the optimization process; calculating the fitness of the initial set of performance data points, evaluating the deviation between each performance data point and the target threshold, and simultaneously initializing the individual optimal value and the global optimal value to prepare for the next round of optimization; determining whether to enter the crossover and mutation process of performance data points based on the set probability; if the random value is greater than the threshold, performing crossover and mutation operations; otherwise, updating the position of the performance data points; grouping the performance data points according to fitness, placing the best-performing performance data points into the optimal solution group for crossover operations, and placing the best-performing performance data points into the inferior solution group for mutation operations; merging the offspring generated by the optimal and inferior solution groups, establishing an expanded dataset with the current set of performance data points, sorting them by fitness, selecting the performance data points with the best fitness as the elite group, and continuing the next round of optimization to generate new performance points to enhance the accuracy of trend fitting.

8. The method for testing the performance of silo-type thermal insulation materials according to claim 7, characterized in that, The process of merging the offspring generated from the optimal and inefficient solution groups to create an expanded dataset with the current performance data point set, sorting the dataset by fitness, and selecting the performance data points with the best fitness as the elite group for the next round of optimization to generate new performance points and enhance the accuracy of trend fitting includes: merging the offspring generated from the optimal and inefficient solution groups to form a new dataset containing the current performance data point set and newly generated performance data points; sorting the merged dataset by fitness, evaluating the deviation of each performance data point from the target threshold, and identifying the best-performing performance data point; selecting the performance data points with the best fitness from the sorted dataset to form the elite group; and based on the elite group, continuing the next round of optimization to generate new performance points and enhance the accuracy of trend fitting.

9. The method for testing the performance of silo-type thermal insulation materials according to claim 8, characterized in that, The formula for evaluating the deviation of each performance data point from the target threshold is: In the formula, M a ω represents the weighted absolute error between the a-th performance data point and the target threshold T; m represents the total number of performance data points; a α represents the weighting coefficient for the a-th performance data point; α represents the adjustment coefficient for the mean squared error; β represents the adjustment coefficient for the absolute error; h a represents the observed value of the a-th performance data point; T represents the target threshold.

10. A performance testing system for silo-type thermal insulation materials, used to implement the performance testing method for silo-type thermal insulation materials according to any one of claims 1-9, characterized in that, The system includes: a data acquisition module for acquiring performance data of silo-type thermal insulation materials under different environmental conditions, preprocessing the performance data, and extracting performance characteristic data; a trend prediction module for predicting the trend of protective performance changes of silo-type thermal insulation materials under different environmental conditions based on performance characteristic data; and a factor analysis and optimization module for establishing an analysis space for identifying anomalies in protective performance change trends, sampling and correcting performance data in the analysis space, analyzing the relationship between performance data and target thresholds and historical sequences, and identifying key factors affecting the protective performance of silo-type thermal insulation materials through anomaly path identification and curve fitting, and adjusting production processes and maintenance strategies based on these key factors.

Images