Method for testing thermal performance of building wall material based on multi-modal sensing

Abstract

The invention relates to the technical field of building material testing, and discloses a method for testing thermal performance of a building wall material based on multi-modal sensing. The method comprises the following steps: arranging a multi-mode sensor on a building wall, acquiring thermal performance data of the wall, segmenting the building wall into a plurality of test sections, and acquiring historical thermal data of each test section; thermal performance parameters are calculated based on historical thermal data and wall thermal performance data collected in real time, the thermal performance grade of each test section is determined, and then a test plan is made according to the grade. In the testing process, environmental interference is detected, real-time thermal data are analyzed, the testing process is corrected based on an analysis result, and finally the thermal performance index of the building wall material is calculated. According to the method, data can be acquired in multiple dimensions, the thermal performance of different sections of the wall body can be accurately monitored, environmental interference can be effectively handled, the accuracy of a test result is improved, and reliable thermal performance data support is provided for related work of a building.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of building material testing, in particular to a method for testing the thermal performance of building wall materials based on multi-modal sensing. BACKGROUND

[0002] In the development process of the construction industry, the thermal performance of wall materials has a direct and key impact on the energy consumption level of buildings and the indoor thermal comfort environment. Accurately mastering the thermal performance of wall materials is an indispensable part of building energy-saving design, energy-saving renovation of existing buildings, and wall material research and development. Currently, there are various methods for testing the thermal performance of building wall materials, such as heat flow meter method and guarded hot box method.

[0003] The heat flow meter method usually arranges heat flow meters and temperature sensors on the surface of the wall, and calculates the thermal performance parameters by collecting heat flux density and temperature difference. However, this method has obvious limitations in practical application. It has high requirements for the stability of the test environment. Once the environmental temperature, wind speed and other factors fluctuate, the collected data will deviate greatly, making it difficult to accurately reflect the thermal performance of the wall in actual use; this method can only test the wall as a whole, and cannot accurately obtain the thermal performance differences of different sections of the wall. When there are local thermal bridges or uneven material distribution, it is difficult to accurately identify the problem, which brings great difficulty to subsequent building energy-saving optimization and troubleshooting.

[0004] Although the guarded hot box method has improved the testing accuracy compared to the heat flow meter method, it can accurately measure the thermal performance of the wall in the laboratory environment, but this method usually requires cutting the wall sample for testing in a specific laboratory device, and cannot realize the in-situ thermal performance testing of the building wall. Since the installation method of the wall in the actual building, the surrounding environment and the connection relationship with other building components will affect its thermal performance, the test results in the laboratory environment are often different from the actual thermal performance of the wall in the actual building, and it is difficult to truly reflect the performance of the wall in the actual use scenario. In addition, the testing equipment of the guarded hot box method is large, the testing process is complex, the testing period is long, and the testing cost is relatively high, which makes it difficult to meet the demand of large-scale, rapid detection of the thermal performance of building walls. SUMMARY

[0005] The present application relates to the technical field of building material testing, in particular to a method for testing the thermal performance of building wall materials based on multi-modal sensing.

[0006] To achieve the above-mentioned purpose, the present application provides a method for testing the thermal performance of building wall materials based on multi-modal sensing, which comprises: A multi-modal sensor is arranged on a building wall to collect wall thermal performance data; the building wall is segmented into multiple test sections; historical thermal data of each test section is obtained; a thermal performance parameter is calculated based on the historical thermal data and the real-time collected wall thermal performance data; a thermal performance level of each test section is determined; a test plan is formulated based on the thermal performance level; environmental interference is detected during the test process; real-time thermal data is analyzed; the test process is corrected based on the analysis result; a thermal performance index of the building wall material is calculated.

[0007] Preferably, the collection of wall thermal performance data comprises: obtaining wall surface temperature distribution data and heat flux density data through the multi-modal sensor; The segmentation of the building wall into multiple test sections comprises: dividing the test sections according to the structural characteristics of the wall, each test section corresponding to an independent thermal performance evaluation unit.

[0008] Preferably, the obtaining of the historical thermal data of each test section comprises: extracting past temperature change records and thermal resistance historical values of each test section from a storage unit; The calculation of the thermal performance parameter based on the historical thermal data and the real-time collected wall thermal performance data comprises: combining the past temperature change records and the wall surface temperature distribution data, performing thermal deformation calculation to generate a thermal strain parameter.

[0009] Preferably, the determination of the thermal performance level of each test section comprises: calculating a thermal performance coefficient according to the thermal strain parameter; comparing the thermal performance coefficient with a preset threshold value to determine the thermal performance level of each test section as a first performance level, a second performance level or a third performance level.

[0010] Preferably, the formulation of the test plan based on the thermal performance level comprises: using the thermal performance level as the primary selection factor, and using the distance between the test section and the sensor reference point as the secondary selection factor; if the thermal performance levels of multiple test sections are equal, the test section closest to the sensor reference point is preferentially selected to perform the test.

[0011] Preferably, the detection of environmental interference during the test process comprises: obtaining center position and interference range data of an environmental interference source; calculating the real-time distance between the real-time test position and the center position of the environmental interference source; When the real-time distance is less than or equal to the interference range, an interference confirmation signal is generated.

[0012] Preferably, the analyzing real-time thermal data includes: capturing temperature abnormality areas in the real-time thermal data; Through multimodal data fusion technology, the temperature abnormality area and heat flux density data are integrated to generate thermal abnormality signals.

[0013] Preferably, the test process of correcting the analysis results includes: determining a test error based on the thermal abnormality signal; Based on an adaptive adjustment algorithm, a test adjustment signal is generated to correct the test plan.

[0014] Preferably, the calculation of thermal performance indicators of building wall materials includes: Based on the calibrated test process and the thermal strain parameters, a thermal permeability calculation is performed to generate a thermal performance index value.

[0015] Preferably, performing thermal permeability calculation includes: The thermal performance evaluation unit of each test section is updated based on the thermal performance index value and historical thermal data.

[0016] Compared with the prior art, the present invention has the following beneficial effects: By installing multimodal sensors on building walls to collect thermal performance data, we can overcome the limitations of traditional single sensors and obtain multi-dimensional data related to thermal performance, such as wall temperature, heat flow, and humidity. This makes the collected data more comprehensive and rich, and more accurately reflects the thermal state of the wall. The use of multimodal sensors can reduce data deviations caused by the inherent characteristics of a single sensor or the influence of external factors, improve data reliability at the source of data collection, and provide a higher-quality data foundation for the subsequent calculation of thermal performance parameters.

[0017] Dividing building walls into multiple test sections and acquiring historical thermal data for each test section changes the traditional testing method, which primarily tests the entire wall, and enables precise monitoring and analysis of the thermal performance of different wall sections. By comparing historical thermal data from different test sections with real-time data, differences in thermal performance between wall sections can be clearly identified, allowing for the timely detection of areas with abnormal thermal performance, such as localized thermal bridges or areas with reduced thermal performance due to material aging. This provides clear guidance for subsequent targeted optimization measures or maintenance and renovation, avoiding the drawback of traditional overall testing, which makes it difficult to locate localized problems.

[0018] The thermal performance parameters are calculated based on historical thermal data and real-time collected wall thermal performance data, fully combining long-term performance information contained in the historical data and current performance state reflected by the real-time data, so that the calculated thermal performance parameters are more comprehensive and accurate. The historical data can reflect the thermal performance change rule of the wall under different seasons and different environmental conditions, and combined with the real-time data, the current wall thermal performance can be more accurately evaluated, the calculation deviation caused by single dependence on real-time data or historical data is reduced, and the thermal performance parameters can better fit the actual performance of the wall.

[0019] The thermal performance grade of each test section is determined, and a test plan is made based on the grade, so that the difference and pertinence of the test process can be realized. For the section with good thermal performance grade and stable performance, the test frequency and test time can be appropriately adjusted to reduce unnecessary test resource consumption; and for the section with low thermal performance grade, performance fluctuation or abnormality, the test frequency can be increased, the test time can be prolonged, the thermal performance change of the section can be monitored, the test resource can be reasonably distributed, the test efficiency can be improved, and the thermal performance of the key section can be fully monitored.

[0020] The environmental interference is detected in the test process, and the test process is corrected based on the analysis result, effectively solving the problem that the traditional test method is easily affected by environmental factors. The environmental temperature, wind speed, sunshine and other external factors will interfere with the test result of the wall thermal performance, the method detects these environmental interference factors in real time, analyzes the influence degree of the real-time thermal data, and then takes corresponding correction measures to eliminate or reduce the adverse effects of environmental interference on the test result, so that the test result can still accurately reflect the thermal performance of the wall material itself under complex and changeable environmental conditions, and the reliability and stability of the test result are improved.

[0021] By analyzing the real-time thermal data and correcting the test process, the thermal performance indicators of the building wall material are calculated again, forming a complete closed-loop test process. The analysis of the real-time thermal data can timely find the abnormal situation in the test process, and the test process correction combined with the environmental interference detection result can continuously optimize the test link and reduce the test error, so that the finally calculated thermal performance indicators of the building wall material can more accurately and objectively reflect the thermal performance level of the wall material, and provide more reliable basis for building energy saving design, existing building energy saving reconstruction, wall material quality evaluation and other work, which is helpful to promote the further development of the building industry in the field of energy saving, and adapt to the demand of intelligent building for precise thermal performance data. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 A working principle diagram of the test method for the thermal performance of the building wall material based on the multi-modal sensing is shown in the figure. Figure 2 Flow chart for determining thermal performance rating; Figure 3 Flow chart for detecting environmental interference; Figure 4 Flow chart for analyzing real-time thermal data. DETAILED DESCRIPTION

[0023] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0024] Please refer to Figure 1 The present application provides a method for testing the thermal performance of building wall materials based on multi-modal sensing, and the overall implementation scheme is as follows: A multi-modal sensor array is deployed on the surface of the building wall to collect thermal performance data of the wall; the overall wall is divided into multiple independent test sections according to the structural characteristics of the wall; historical thermal data of each test section is called from a data storage system; fusion calculation is performed based on the historical data and the real-time collected thermal performance data of the wall to obtain thermal performance parameters; the thermal performance rating of each test section is determined according to the parameter calculation results; a differentiated test plan is formulated according to the performance rating; environmental interference factors are monitored in real time during the test execution; multi-dimensional analysis is performed on the collected real-time thermal data; the test flow is dynamically corrected according to the analysis results; and finally the comprehensive thermal performance index of the building wall material is calculated.

[0025] In the implementation process of the multi-modal sensor-based building wall material thermal performance test method, the deployment of the multi-modal sensor adopts a combination scheme of an infrared thermal imager and a heat flow sensor. The infrared thermal imager is installed on an adjustable gimbal, and the gimbal is kept parallel to the wall surface through a fixed base. The installation position needs to ensure that the lens field of view covers the entire area to be tested. The working waveband of the infrared thermal imager is in the long-wave infrared range, and the temperature measurement accuracy is controlled within a reasonable range. The sampling frequency is set to one complete temperature distribution image per minute. Each temperature distribution image contains temperature data of millions of pixels, and each pixel corresponds to the temperature value of a specific position on the wall surface. The data format is a floating-point matrix. The heat flow sensor is installed in a patch form and directly attached to the wall surface. The sensor sensing surface is completely attached to the wall surface, and thermal grease is applied in between to reduce the contact thermal resistance. The heat flow sensor collects heat flux density data at a frequency of five times per second. The data includes heat intensity values and direction vectors. In the multi-modal sensor deployment stage, in addition to the combination of the infrared thermal imager and the heat flow sensor, the layout density of the sensor array is determined according to the wall area and structural complexity. For walls with larger areas or more door and window openings, structural columns and other complex structures, the number of sensors should be appropriately increased to ensure that at least two or more sensors cover each test section. This avoids data collection blind spots caused by insufficient sensor coverage and further ensures the spatial continuity of wall surface temperature distribution data and heat flux density data, laying a foundation for subsequent segmented testing and accurate data analysis. At the same time, during the sensor installation process, appropriate installation methods are selected according to the characteristics of different wall materials. For example, for hard wall materials such as concrete, expansion screws are used to assist in fixing the sensor base. For lightweight wall materials such as insulation boards, special adhesives are used for pasting to ensure that the sensor is tightly attached to the wall surface and to reduce the impact of improper installation methods on data collection accuracy.

[0026] The division of the test section is based on the structural characteristics of the wall. A high-resolution visible light camera is used to first obtain a panoramic image of the wall. The image is transmitted to an image processing system for edge detection and feature recognition. The system automatically identifies significant feature lines such as material joints, door and window opening edges, and structural column boundaries on the wall. Based on these feature lines, the wall surface is divided into several regular rectangular regions. Each rectangular region is assigned a unique number, and the numbering rule is generated in the order from left to right and from top to bottom. Each rectangular region serves as an independent thermal performance evaluation unit, and the unit number is mapped to the data collection channel of the multi-modal sensor. The mapping relationship is stored in a configuration file, which contains the geometric center coordinates, area data, and boundary coordinate set of each evaluation unit.

[0027] During data acquisition, the infrared thermal imager transmits temperature distribution data in real time to the central processing unit via a Gigabit Ethernet interface. Temperature distribution data is stored in a matrix format, with the row and column indices of the matrix corresponding to the two-dimensional coordinate positions on the wall surface, and the matrix element values ​​representing the temperature measurements at that position. The heat flux sensor transmits heat flux density data to the data acquisition card via the RS-485 bus. The acquisition card performs analog-to-digital conversion on the data and uploads it to the central processing unit. Heat flux density data is stored in a time series format, with each data point containing a timestamp, heat flux intensity value, and heat flux direction angle. The central processing unit establishes an independent data cache for each test section. The cache uses a ring buffer structure to store real-time data from the last several hours.

[0028] Temperature distribution data processing involves background temperature compensation and radiation correction. Background temperature compensation is achieved using a reference blackbody placed near the sensor, while radiation correction accounts for the effects of ambient radiation reflection and atmospheric transmittance. Heat flux data processing involves signal filtering and unit conversion. Filtering uses a digital filter to eliminate high-frequency noise, while unit conversion converts the raw voltage signal into standard heat flux units. The processed data is stored in a distributed database system, which is organized into separate tables based on the test segment number. Temperature data for each test segment is stored as a time series matrix, and heat flux data as a multidimensional time series.

[0029] The geometric information of the test sections is stored in a spatial database, which records the vertex coordinates, center coordinates, and area of ​​each rectangular area. The sensor reference point coordinates are measured using a global positioning system (GPS) with centimeter-level accuracy. The reference point coordinates, along with the coordinates of each test section's center point, are stored in the positioning database. The database is spatially indexed to support fast distance queries. Distance calculations utilize a planar projection algorithm, which accounts for the planar nature of wall surfaces and simplifies three-dimensional distance calculations into two-dimensional distance calculations.

[0030] Multimodal sensors are calibrated regularly, with the calibration cycle dynamically adjusted based on the equipment's operating status. Infrared thermal imagers are calibrated using a standard temperature source, collecting calibration data at different temperature points to establish a temperature-voltage relationship table. Heat flow sensors are calibrated using a standard heat flow source, and the calibration process includes zero-point and span calibration. Calibration data is stored in the equipment archive database for online correction during real-time data acquisition. Data acquisition software automatically adjusts acquisition parameters based on this calibration data to ensure measurement accuracy.

[0031] Quality monitoring is implemented during data collection, and the monitoring indicators include data integrity, data validity, and data consistency. Data integrity checks whether the data packet contains all required fields, data validity checks whether the measured value is within a reasonable range, and data consistency checks whether there is a logical conflict in the synchronized data collected by different sensors. When an abnormal data is detected, the system automatically records the abnormal event and triggers the data re-collection mechanism. The re-collection mechanism adopts different strategies according to the type of abnormality. For transient abnormalities, immediate re-collection is adopted, and for continuous abnormalities, the sensor working parameters are adjusted before re-collection.

[0032] A redundancy check mechanism is used for data transmission, and a check code is added to each data packet, which is checked by the receiving end. Data packets that fail the check require the sending end to retransmit, and when the number of transmission failures exceeds the threshold, a fault diagnosis program is started. The fault diagnosis program detects the network connection state, sensor working state, and data acquisition card state, and the diagnosis results are recorded into the system log. The system log records all operation events and data flow states in real time, and the log files are stored by date rotation, with a storage period that meets the relevant requirements.

[0033] A perfect data traceability mechanism is established throughout the implementation process, and the operation time and operation results are recorded at each link from data collection to storage. Data traceability is achieved through a unique identifier, and each data packet carries information such as collection time, sensor number, and test section number. These information is always associated during data processing, ensuring the traceability of data sources. Data storage uses a multi-backup strategy, and real-time data is stored in memory cache, disk storage, and network storage at three levels, and the data at different levels is periodically checked for consistency.

[0034] Example 2: refer to Figure 2The storage unit adopts a distributed database architecture to manage historical thermal data. The database system is composed of multiple nodes, each responsible for storing historical data of a specific test section. In terms of historical thermal data management, the distributed database not only stores past temperature change records and thermal resistance history values by test section, but also annotates and classifies historical data according to the season and weather type (such as sunny, cloudy, rainy) during testing. In this way, when extracting historical data for fusion calculation with real-time data, it can more accurately filter out historical data similar to the current test environment conditions, reducing the calculation deviation of thermal performance parameters caused by large differences in environmental conditions. Past temperature change records are stored in time series, recording content including temperature extreme value data collected during historical testing, daily average temperature fluctuation curve, and temperature change trend data for each test section. Temperature extreme value data includes daily maximum temperature, minimum temperature, and their occurrence timestamps, and the daily average temperature fluctuation curve is generated by smoothing twenty-four continuous sampling data. Thermal resistance history values come from the calculation results of previous test records, stored as a set of equivalent thermal resistance values for each test section, including thermal resistance value records under different seasons and weather conditions.

[0035] During data extraction, the system calls data from the corresponding partition of the database according to the test section number. The time range for extraction is set to the last twelve complete test periods, each test period containing all data records from the start to the end of the test. The data query statement includes time range filtering conditions, test section number matching conditions, and data integrity verification conditions. The extracted data packet includes a temperature time series array and a thermal resistance value array, each array element with an accurate timestamp and data quality identifier.

[0036] Before thermal performance parameter calculation, the system performs data preprocessing operations. Preprocessing includes two main steps: time axis alignment and spatial coordinate matching. Time axis alignment unifies the sampling frequency of historical temperature data to the sampling frequency of real-time data through interpolation algorithms, and spatial coordinate matching unifies the pixel grid of historical temperature data with the pixel grid of real-time temperature distribution data. In the alignment process, the nearest neighbor interpolation algorithm is used to handle pixel position deviation, ensuring that the historical data of each pixel point has a consistent spatial correspondence with real-time data.

[0037] The thermal deformation calculation is realized by using a difference algorithm to calculate the pixel difference between the real-time temperature data matrix and the historical reference temperature data matrix. The reference temperature data is selected from the test records under the same environmental conditions in the historical data, including environmental temperature, humidity, solar intensity, and other parameters. In the thermal deformation calculation process, in addition to combining the past temperature change records and the wall surface temperature distribution data, the aging degree factor of the wall material is also considered. By associating the construction time and the past maintenance records of each test section wall in the database, a material aging correction factor is appropriately introduced for the wall with a long service time in the thermal deformation calculation, so that the generated thermal strain parameters are more consistent with the actual thermal deformation of the wall under the service state. The thermal strain parameter matrix generated by the difference calculation contains the strain value of each pixel point, which reflects the deformation characteristics of the material caused by temperature change. The thermal strain parameter not only contains the basic linear expansion coefficient, but also contains derived parameters such as temperature gradient change rate and thermal strain distribution uniformity.

[0038] The thermal performance coefficient calculation is completed by using a multiple regression model, and the input parameters of the model include the index values in the thermal strain parameter matrix. The weight coefficients of the model are dynamically configured according to the type of wall material, and different material types correspond to different sets of weight coefficients. The weight coefficients are stored in the material attribute database, which contains the physical characteristic parameters of common wall materials. The regression model outputs a comprehensive performance score value, which is the thermal performance coefficient.

[0039] The performance level determination process compares the calculated thermal performance coefficient with the preset threshold interval. The threshold interval is divided into three clear ranges: the first performance level corresponds to the interval from zero to zero point three, the second performance level corresponds to the interval from zero point three to zero point seven, and the third performance level corresponds to the interval from zero point seven to one point zero. The comparison algorithm uses the interval inclusion detection method to determine which threshold interval range the thermal performance coefficient value falls into. The determination result generates a performance level identifier, which is represented in the form of a digital code.

[0040] The performance level determination result is written into a special data table in the test section attribute database, which records the number of each test section, the test time, the thermal performance coefficient value, and the performance level code. At the same time, the system associates the performance level information with the corresponding area in the three-dimensional wall model. The three-dimensional wall model is constructed using the building information modeling technology, and each test section has a corresponding geometric entity in the model. The association operation is realized by spatial position matching, and the performance level data is assigned to the attribute field of the corresponding geometric entity.

[0041] The visualization correlation process adopts a color coding scheme, different performance levels correspond to different display colors. The first performance level area is displayed in green, the second performance level area is displayed in yellow, and the third performance level area is displayed in red. The color rendering is accelerated by the graphics processing unit, and the color distribution of the wall model surface is updated in real time. The visualization system supports multi-angle viewing and detail zooming, allowing operators to observe the detailed performance characteristics of each test section.

[0042] The entire implementation process establishes a complete data pipeline, from historical data extraction to performance level visualization presentation, forming a closed-loop process. Each link of data processing records operation logs, which include processing time, data volume, processing result status, etc. The data pipeline sets multiple quality checkpoints to verify the effectiveness of intermediate results, and initiates a reprocessing mechanism when data anomalies are found. The reprocessing mechanism selects to re-execute a single processing step or the entire processing flow according to the type of anomaly, ensuring the reliability of the final result.

[0043] The system periodically archives historical data, which is compressed and encrypted before being transferred to long-term storage media. The archiving strategy is based on data timestamps, and data exceeding a certain period of time is automatically entered into the archiving process. Archiving data retains all original information and processing records, supporting data backtracking and analysis when needed. The data management interface provides historical data query and retrieval functions, allowing users to find the required data records by specifying time range, test section number, etc.

[0044] Example 3: Refer to Figure 3 The test plan is based on two key factors: thermal performance level and spatial distance. The system arranges all test sections in descending order according to thermal performance level, with higher level values indicating higher test priority. When multiple test sections have the same thermal performance level, the system calculates the spatial distance between the center point of these sections and the sensor reference point, and selects the closest section for testing first. For example, the east outer wall of a building is divided into 12 test sections, with their thermal performance levels and distance data as shown in Table 1.

[0045] Table 1: Test section priority ranking table.

[0046]

[0047] The distance calculation uses coordinate data obtained by the global positioning system, and after conversion by the Gauss-Kruger projection, the planar distance between the center point of each test section and the sensor reference point is calculated. The sensor reference point is set at the geometric center of the wall, and its coordinates are obtained by differential GPS measurement, with a planar coordinate accuracy of centimeter level. When calculating the distance, the planar characteristics of the wall surface are considered, and the three-dimensional space distance is projected onto a two-dimensional plane for calculation. The projection plane is parallel to the wall surface.

[0048] The environmental interference detection is realized by a distributed monitoring network deployed in the test site. The network consists of 8 ultrasonic sensors and 6 electromagnetic field sensors, which are evenly distributed around the test building. In the distributed monitoring network for environmental interference detection, in addition to ultrasonic sensors and electromagnetic field sensors, temperature sensors and wind speed sensors are added to collect real-time temperature changes and wind speed data of the test environment, which are used as supplementary basis for environmental interference assessment. When the environmental temperature fluctuates greatly in a short time or the wind speed exceeds the set threshold, an interference warning signal is generated, further improving the environmental interference detection system and enhancing the identification ability of complex environmental interference. The monitoring range of the ultrasonic sensor covers an area with a radius of 50 meters, and the sampling frequency is 10 times per second, which can detect ultrasonic interference generated by mechanical vibration, personnel activity, etc. The electromagnetic field sensor works in the frequency band of 50Hz-2.4GHz, with a sensitivity of microtesla level, which can monitor electromagnetic interference generated by power equipment, wireless communication equipment, etc.

[0049] The interference monitoring network collects environmental data in real time and transmits it to the central processing unit through a special data bus. The data processing algorithm identifies the characteristics of the interference signal and determines the center position coordinates of the interference source. The position calculation uses a multi-point positioning algorithm to calculate the spatial coordinates of the interference source by the time difference and intensity difference of the signals received by multiple sensors. The interference range data is calculated by a signal attenuation model, which considers the influence of obstacles on the signal propagation path and gives the three-dimensional spatial distribution range of the interference field.

[0050] The real-time test position is obtained by the Beidou / GPS dual-mode positioning module installed on the mobile test equipment, with a positioning data update frequency of 1 time per second and a positioning accuracy of sub-meter level. The positioning module outputs latitude and longitude coordinates, which are converted into a planar coordinate system consistent with the interference monitoring network by a coordinate conversion algorithm.

[0051] The real-time distance calculation uses the plane Euclidean distance formula to calculate the straight-line distance between the current position of the test equipment and the center point of the interference source. In the spatial distance calculation link of the test plan preparation, in addition to using the plane Euclidean distance formula to calculate the distance between the test section center point and the sensor reference point, the distance calculation needs to avoid the protruding components (such as window sills, waist lines, etc.) on the surface of the wall. If the line connecting the test section center point and the sensor reference point passes through the protruding component, the broken-line distance calculation method is used, that is, the shortest path distance from the test section center point to the sensor reference point bypassing the protruding component, to ensure that the distance calculation result can truly reflect the actual moving path length of the test equipment, and provide more accurate spatial distance basis for test sequence optimization. During the calculation process, the system real-time acquires the dynamic position coordinates of the test equipment and the updated position data of the interference source. When the calculated real-time distance value is less than or equal to the interference range radius, the system generates a digital interference confirmation signal.

[0052] The interference confirmation signal adopts a standard data format, including fields such as interference type code, interference source intensity, interference duration, and impact range. The interference type code is classified according to international standards, including categories such as mechanical vibration, electromagnetic radiation, and thermal radiation. The interference source intensity is represented by a normalized numerical value, ranging from zero to one. The interference duration records the length of time from the occurrence of the interference to the present. The impact range field records the three-dimensional spatial dimensions of the interference field.

[0053] When the interference confirmation signal is generated, the system automatically triggers the test process adjustment mechanism. The adjustment measures include pausing data collection, reducing sensor sensitivity, activating shielding devices, etc. For transient interference, the system records the interference time period, and continues testing after the interference disappears; for continuous interference, the system recalculates the test path to avoid the interference impact area.

[0054] During the test process, the system continuously monitors the environmental interference conditions, updating the interference source database once every second. The database records all detected interference source information, including occurrence time, duration, intensity change curve, etc. These data are used for interference removal processing during subsequent test quality evaluation and data analysis.

[0055] The entire test plan execution process implements a dynamic adjustment mechanism. The system re-evaluates the priority ranking of the test section every five minutes, updates the thermal performance level according to the latest collected data, and adjusts the test sequence accordingly. At the same time, the system real-time updates the interference source information, dynamically optimizes the test path, and ensures that the test process is carried out under optimal conditions. All adjustment operations are recorded in the system log, including adjustment time, adjustment reason, and detailed information such as parameter changes before and after adjustment, forming a complete test process traceability record.

[0056] Example 4: see Figure 4In the implementation process of the building wall material thermal performance test method based on multi-modal sensing, real-time thermal data analysis is processed using a sliding time window mechanism. Each time window contains sixty consecutive sampling points of temperature data, and the window width is dynamically adjusted according to the sampling frequency. In the sliding time window mechanism of real-time thermal data analysis, in addition to dynamically adjusting the window width according to the sampling frequency, the setting rule of the window sliding step is as follows: for the test stage with relatively gentle temperature change, a larger sliding step is adopted to reduce the data processing amount and improve the analysis efficiency; for the test stage with relatively frequent temperature fluctuation, the sliding step is reduced to increase the data sampling density and ensure that subtle temperature abnormal changes can be captured in time. Temperature abnormal area detection is realized through an improved statistical outlier identification algorithm, which calculates the deviation of each sampling point from the average temperature value in the window, and establishes a dynamic threshold model based on historical data. When the temperature value of a sampling point exceeds the threshold range, the point is marked as a potential abnormal point.

[0057] The multi-modal data fusion technology adopts an improved Kalman filter algorithm, which spatiotemporally matches the spatial coordinate information of the temperature abnormal area with the synchronously collected heat flux density data. The fusion process first performs spatial interpolation processing on the heat flux density data to make the spatial resolution consistent with the temperature distribution data. In the multi-modal data fusion process, in addition to the spatial interpolation processing of the heat flux density data, the temperature abnormal area data is also preprocessed by smoothing to remove isolated abnormal temperature points caused by instantaneous sensor errors, avoiding the interference of isolated abnormal points on the calculation results of the heat flux-temperature coupling model, so that the generated thermal abnormality signal can more accurately reflect the real thermal abnormality of the wall. By establishing a heat flux-temperature coupling model, the correlation index of heat flux and temperature at each spatial position point is calculated:

[0058] wherein: represents the thermal abnormality correlation coefficient, represents the number of sampling points, is the heat flux density measurement value of the i th sampling point, represents the average heat flux density in the window, is the standard deviation of the heat flux density data, is the temperature measurement value of the i th sampling point, represents the average temperature in the window, is the standard deviation of the temperature data. When the value is lower than the set threshold, the system determines that there is a thermal abnormality in the area, and generates a thermal abnormality signal containing the abnormal area boundary coordinates and the abnormal intensity level.

[0059] The test error determination process establishes an error propagation model based on the thermotechnical anomaly signal. The model maps the anomaly intensity value to a confidence bias value of the test data, and the mapping relationship is realized through a pre-established calibration curve. The calibration curve is established based on a large amount of experimental data and contains the measurement error range corresponding to different anomaly intensities. The confidence bias value is expressed in percentage, and the larger the value, the lower the reliability of the test data.

[0060] The adaptive adjustment algorithm is implemented using fuzzy control logic, which includes multiple input variables and output variables. The input variables include the confidence bias value, the anomaly duration, the anomaly area proportion, and other parameters. The output variables include the sampling frequency adjustment amount, the test duration correction coefficient, and the sensor gain adjustment parameter. The fuzzy control rule base contains a series of empirical rules, which are established based on expert knowledge and historical debugging data. In the fuzzy control rule base of the adaptive adjustment algorithm, in addition to the existing rules based on the confidence bias value and the anomaly duration, control rules based on the anomaly region position are added: if the thermotechnical anomaly region is located in the key load-bearing part of the wall or the node part prone to thermal bridge, more aggressive adjustment measures (such as significantly increasing the sampling frequency, prolonging the test duration) are preferred to ensure more detailed monitoring and analysis of the thermotechnical anomaly in the key region.

[0061] The control rules are in the form of if-then sentences, for example: if the confidence bias value is large and the anomaly duration is long, then significantly reduce the sampling frequency. The fuzzy reasoning process uses the Mamdani reasoning method to convert fuzzy output to precise control instructions through defuzzification processing. These instructions are issued in the form of digital control signals, including specific parameter adjustment values and effective time points.

[0062] The execution of the test adjustment signal adopts a gradual adjustment strategy to avoid the impact of parameter mutation on the test process. The sampling frequency adjustment is achieved by changing the clock frequency of the data acquisition card, and the adjustment step is dynamically determined according to the current frequency value. The test duration correction is achieved by recalculating the remaining test time, considering the anomaly influence degree and test progress factors. The sensor gain adjustment is achieved by changing the amplifier circuit parameters, and the adjustment process includes a calibration verification link.

[0063] During the entire adjustment process, the system monitors the adjustment effect in real time and optimizes the adjustment parameters through a feedback mechanism. The effect monitoring indicators include the data quality index, the signal stability coefficient, and the measurement consistency index. These indicators are calculated every second and compared with the baseline values before adjustment. If the adjustment effect does not meet the expected target, the system starts a secondary adjustment process and uses different control rules for re-optimization.

[0064] All adjustment operations are recorded in the system log, including adjustment time, adjustment parameters, adjustment reasons, and adjustment effects, etc. Detailed information. Log data is stored in a structured format, supporting subsequent queries and analysis. The system also establishes an adjustment effect evaluation mechanism, continuously optimizing fuzzy control rules through machine learning algorithms to improve the accuracy and efficiency of adaptive adjustment. All parameter changes involved in the adjustment process are displayed in real-time through a visual interface, making it easy for operators to monitor and understand the system state.

[0065] In the implementation process of the building wall material thermal performance test method based on multi-modal sensing, the thermal permeability calculation uses a numerical solution method to handle the unsteady heat conduction problem. At the beginning of the calculation, the system first establishes a physical model of the wall material, which includes the basic parameters of the material such as thermal conductivity, specific heat capacity, and density. In the physical model construction stage of thermal permeability calculation, in addition to the basic parameters of the material such as thermal conductivity, specific heat capacity, and density, the construction level information of the wall is also included. If the wall is composed of different materials in layers, the thickness of each layer of material and the interface contact resistance parameters are explicitly defined in the model, and a layered calculation method is used to handle the unsteady heat conduction problem, making the calculation results more in line with the multi-layer construction characteristics of the actual wall. The corrected test process data is input as boundary conditions, including the temperature change history of the wall surface and the heat flow boundary conditions. The thermal strain parameters are used as material property adjustment variables in the calculation, reflecting the influence of temperature changes on the thermal physical properties of the material.

[0066] The calculation area is discretized into a fine grid system, and each grid cell is assigned a corresponding material attribute value. The time step is dynamically adjusted according to the temperature change rate, with a smaller step size when the temperature changes rapidly and a larger step size when the temperature changes slowly. During the solution process, the temperature and heat flow values of each grid cell at each time step are recorded, which are used for subsequent thermal performance index calculation.

[0067] The thermal performance index value is obtained by analyzing the calculated temperature field and heat flow field data. The system calculates the heat storage capacity and heat transfer capacity of the wall within a certain time period, and integrates these data into thermal performance indicators. The index value contains multiple dimensions, including thermal inertia index, thermal response rate, thermal decay multiple, and other parameters. Each parameter is normalized, and finally a comprehensive performance score in the range of zero to one is generated.

[0068] The historical thermal data update adopts an incremental learning mechanism. The newly calculated thermal performance index values are weighted and fused with the historical values stored in the database. The weight coefficients are determined according to the data timestamp. In the incremental learning mechanism of historical thermal data update, in addition to determining the weight coefficients according to the data timestamp, the weight is also adjusted in combination with the test environment similarity of historical data. For historical data with high similarity to the current test environment conditions, the weight proportion is appropriately increased, and vice versa. The weight is reduced, so that the fused historical data can provide more effective reference for the current thermal performance index calculation. The weight of recent data is higher, and the weight of long-term data is lower. The weight allocation adopts an exponential decay function. The new value after fusion replaces the original value and is stored in the database, while the historical version record is also retained.

[0069] The update operation of the thermal performance evaluation unit is carried out at two levels of data and model. At the data level, the system writes the updated thermal performance index values into a special field of the evaluation unit attribute database. The database records the timestamp of each update, the value before update, the value after update, and the data source for update. At the model level, the system re-trains the thermal performance prediction model using the updated data set. The training process adopts a supervised learning algorithm, and the input features include environmental parameters, material parameters and historical performance data, and the output is the predicted thermal performance index.

[0070] After the model training is completed, the system uses the new model to re-predict the thermal performance level of all test sections. The prediction results are compared with the existing level, and if significant differences are found, the level adjustment process is triggered. The level adjustment is based on the consistency of the prediction results and the actual measured data, and the adjustment range is determined by pre-defined rules.

[0071] The updated evaluation unit generates a new thermal performance level label, which includes the level code and the confidence index. The confidence index reflects the credibility of the prediction results, which is calculated based on the model prediction error and the consistency of historical data. The new label is written into the evaluation unit attribute database and is re-associated with the three-dimensional wall model.

[0072] The reset calculation of the test plan is automatically started after the update of the evaluation unit is completed. The reset process maintains the principle of taking the thermal performance level as the primary selection element, but uses the updated level data for priority sorting. The system recalculates the test priority of all test sections and generates a new test sequence. The new test plan considers the latest performance level distribution and test resource constraints, optimizes the test path and test time allocation.

[0073] The whole updating process forms a complete closed-loop control system. The system periodically collects new test data, updates the thermal performance indicators, re-trains the prediction model, adjusts the performance levels, and optimizes the test plan. This closed-loop design enables the system to adapt to the characteristics of the thermal performance of wall materials changing over time, maintaining the accuracy and reliability of the test results. Each closed-loop run produces a detailed operation log, recording the complete process from data collection to plan update, supporting subsequent audit and optimization analysis.

[0074] The system also establishes a version control mechanism to track all changes to data and models. Each update generates a new version number, and the version information contains metadata such as update time, change content, and operator. The version history supports backtracking to any historical state, facilitating comparison of performance change trends at different times. This design provides technical support for long-term monitoring of the evolution of building wall thermal performance.

[0075] It should be noted that the relational terms herein such as first and second and the like are used solely to distinguish one from another entity or action, without necessarily requiring or implying that any such entity or action is either the first or second in time. Also, the terms "comprises", "comprising", or any other variation thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can also include other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0076] Although embodiments of the present application have been shown and described, it is to be understood that various modifications, substitutions, combinations, and variations can be made by those skilled in the art without departing from the principles and spirits of the present application, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for testing the thermal performance of building wall materials based on multimodal sensing, characterized in that: The application relates to a method for testing thermal performance of a building wall. The method comprises: setting up a multi-modal sensor on the building wall and collecting thermal performance data of the wall; segmenting the building wall into multiple test sections; obtaining historical thermal performance data of each test section, calculating thermal performance parameters based on the historical thermal performance data and the real-time collected thermal performance data of the wall, determining the thermal performance grade of each test section, and formulating a test plan based on the thermal performance grade; detecting environmental interference during the test; analyzing real-time thermal performance data; correcting the test process based on the analysis result; calculating the thermal performance index of the building wall material; the collection of the thermal performance data of the wall comprises: obtaining wall surface temperature distribution data and heat flux density data through the multi-modal sensor; the segmentation of the building wall into multiple test sections comprises:

2. The method of testing the thermal performance of a building wall material based on multi-modal sensing of claim 1, wherein, dividing the test sections according to the structural features of the wall, and each test section corresponds to an independent thermal performance evaluation unit. the obtaining of the historical thermal performance data of each test section comprises: extracting the past temperature change record and the thermal resistance historical value of each test section from a storage unit; the calculation of the thermal performance parameters based on the historical thermal performance data and the real-time collected thermal performance data of the wall comprises:

3. The method of testing the thermal performance of a building wall material based on multi-modal sensing of claim 2, wherein, combining the past temperature change record and the wall surface temperature distribution data, performing thermal deformation calculation to generate thermal strain parameters. the determination of the thermal performance grade of each test section comprises: calculating the thermal performance coefficient according to the thermal strain parameters; 4. The method of testing the thermal performance of a building wall material based on multi-modal sensing of claim 3, wherein, comparing the thermal performance coefficient with a preset threshold value to determine the thermal performance grade of each test section as a first performance grade, a second performance grade or a third performance grade. the formulation of the test plan based on the thermal performance grade comprises: taking the thermal performance grade as the primary selection factor and taking the distance between the test section and the sensor reference point as the secondary selection factor; 5. The method of testing the thermal performance of a building wall material based on multi-modal sensing of claim 4, wherein, if the thermal performance grades of multiple test sections are equal, the test section closest to the sensor reference point is preferentially selected to perform the test. the detection of environmental interference during the test comprises: obtaining the center position and the interference range data of the environmental interference source; calculating the real-time distance between the real-time test position and the center position of the environmental interference source; 6. The method of testing the thermal performance of a building wall material based on multi-modal sensing of claim 5, wherein, generating an interference confirmation signal when the real-time distance is less than or equal to the interference range. the analysis of the real-time thermal performance data comprises: capturing the temperature abnormal area in the real-time thermal performance data; 7. The method of testing the thermal performance of a building wall material based on multi-modal sensing of claim 6, wherein, integrating the temperature abnormal area and the heat flux density data through a multi-modal data fusion technology to generate a thermal abnormality signal. the correction of the test process based on the analysis result comprises: determining the test error according to the thermal abnormality signal; 8. The method of testing the thermal performance of a building wall material based on multi-modal sensing of claim 7, wherein, generating a test adjustment signal based on an adaptive adjustment algorithm to correct the test plan. the calculation of the thermal performance index of the building wall material comprises:

9. The method of testing the thermal performance of a building wall material based on multi-modal sensing of claim 8, wherein, performing thermal permeability calculation based on the corrected test process and the thermal strain parameters to generate a thermal performance index value. the thermal permeability calculation comprises: combining the thermal performance index value and the historical thermal performance data to update the thermal performance evaluation unit of each test section.

Images