External wall heat transfer coefficient field detection method and system based on infrared thermal imaging technology

Abstract

The invention provides an outer wall heat transfer coefficient field detection method and system based on an infrared thermal imaging technology. An infrared thermal imager is used for rapidly scanning an outer wall in a non-contact mode, initial temperature distribution is efficiently obtained, time and labor are saved, a dynamic environment correction mechanism is constructed, wall temperature, the sun, wind speed and other data are collected for 12 hours, phase difference characteristics are extracted, and the outer wall heat transfer coefficient field detection method and system based on the infrared thermal imaging technology are obtained through amplitude and phase spectrograms. Lagging interference of solar radiation and wind speed fluctuation is accurately eliminated through an instant correction coefficient and a phase offset, temperature data are made to be more real, a heat transfer coefficient calculated based on the corrected temperature data is accurate and reliable in result, a key basis can be provided for building energy-saving transformation and judgment of whether the standard is reached or not, and the method is suitable for different environments and outer wall types, high in universality and easy to popularize. And energy-saving evaluation and quality control in the building industry are facilitated.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of building wall detection, in particular to a kind of on-site detection method of outer wall heat transfer coefficient based on infrared thermal imaging technology, a kind of on-site detection system of outer wall heat transfer coefficient based on infrared thermal imaging technology. BACKGROUND

[0002] The heat transfer coefficient of building outer wall is the core index to measure building energy-saving performance, and its accurate detection is of great significance to building energy consumption evaluation and energy-saving reconstruction scheme formulation. At present, the detection methods of outer wall heat transfer coefficient mainly include laboratory detection method and on-site detection method. The laboratory detection method tests the wall specimen by constructing a simulated environment, which can obtain relatively accurate results under controllable conditions, but has obvious limitations: on the one hand, there may be differences between the specimen and the actual building outer wall in structure and construction technology, resulting in a deviation between the test results and the actual situation; on the other hand, this method cannot directly detect the heat transfer coefficient of the outer wall of the built building, and has narrow applicability. The on-site detection method mainly includes heat flow meter method and heat box method. The heat flow meter method directly measures the heat flux density and temperature difference by arranging heat flow meters and temperature sensors on the wall surface to calculate the heat transfer coefficient, but its detection period is long (usually more than 72 hours of continuous monitoring), and the arrangement of sensors will damage the integrity of the wall surface, causing certain impact on the appearance of the building. The heat box method forms a temperature difference by constructing an artificial thermal environment on both sides of the wall to calculate the heat transfer coefficient. This method is cumbersome and complicated to operate, and is greatly disturbed by environmental factors, so the detection accuracy is difficult to guarantee. With the development of infrared thermal imaging technology, it has been preliminarily applied in building wall detection due to its advantages of non-contact and rapid acquisition of large-area temperature field. However, the existing detection methods based on infrared thermal imaging technology still have obvious defects: since the building outer wall is exposed to the natural environment for a long time, its surface temperature is easily affected by dynamic environmental factors such as solar radiation and wind speed, resulting in large errors in the initial temperature data obtained by the infrared thermal imager; at the same time, the influence of solar radiation and wind speed on the wall temperature has a lag, and the existing methods lack effective correction mechanism, which is difficult to eliminate such interference, so that the final calculated heat transfer coefficient has low precision and cannot meet the needs of engineering practice. SUMMARY

[0003] In view of the above problems, the present application is proposed to provide a kind of on-site detection method of outer wall heat transfer coefficient based on infrared thermal imaging technology and corresponding on-site detection system of outer wall heat transfer coefficient based on infrared thermal imaging technology, which can overcome the above problems or at least partially solve the above problems.

[0004] The present application discloses an on-site detection method of outer wall heat transfer coefficient based on infrared thermal imaging technology, which comprises: An infrared thermal imager is used to detect the temperature of the outer wall surface of a building to obtain an initial temperature of the outer wall surface; the initial temperature of the outer wall surface includes an initial temperature of the outer surface of the outer wall; The temperature data of the outer wall surface and the environmental parameter data within a preset time period are synchronously collected; Based on the temperature data of the outer wall surface and the environmental parameter data within the preset time period, a frequency domain analysis method is used to analyze the dynamic relationship between the changes in the environmental parameter data and the changes in the temperature of the outer wall surface; Based on the dynamic relationship between the changes in the environmental parameter data and the changes in the temperature of the outer wall surface, a dynamic correction factor for correcting the initial temperature of the outer wall surface is calculated, and the initial temperature of the outer wall surface is corrected by using the dynamic correction factor to obtain a corrected temperature of the outer wall surface; the corrected temperature of the outer wall surface includes a corrected temperature of the outer surface of the outer wall; The temperature of the inner surface of the outer wall, the indoor air temperature and the outdoor air temperature are collected, and a machine learning algorithm is used to calculate the heat transfer coefficient of the outer wall of the building based on the corrected temperature of the outer surface of the outer wall, the temperature of the inner surface of the outer wall, the indoor air temperature and the outdoor air temperature.

[0005] Optionally, the environmental parameter data at least includes solar radiation intensity and wind speed; based on the temperature data of the outer wall surface and the environmental parameter data within the preset time period, a frequency domain analysis method is used to analyze the dynamic relationship between the changes in the environmental parameter data and the changes in the temperature of the outer wall surface, which includes: Based on the temperature data of the outer wall surface, the solar radiation intensity data and the wind speed data within the preset time period, a temperature change curve of the outer wall surface is constructed; the temperature change curve of the outer wall surface shows the dynamic evolution law of temperature with time, and directly reflects the time sequence corresponding relationship between the key environmental factors of solar radiation intensity and wind speed and the temperature change of the wall; Based on the temperature change curve of the outer wall surface, amplitude spectrum and phase spectrum of each frequency component are generated; Based on the amplitude spectrum and phase spectrum of each frequency component, the measured amplitude peak value, the phase difference between the changes in solar radiation intensity and the changes in the temperature of the outer wall surface, and the phase difference between the changes in wind speed and the changes in the temperature of the outer wall surface are extracted.

[0006] Optionally, based on the dynamic relationship between the changes in the environmental parameter data and the changes in the temperature of the outer wall surface, a dynamic correction factor for correcting the initial temperature of the outer wall surface is calculated, and the initial temperature of the outer wall surface is corrected by using the dynamic correction factor to obtain a corrected temperature of the outer wall surface, which includes: Based on the amplitude spectrum and phase spectrum, the solar activity characteristics and wind speed fluctuation characteristics corresponding to the frequency components are interpreted; If the solar activity feature is high-frequency fluctuation, an instant correction coefficient is calculated based on a measured amplitude peak value and a preset wind speed fluctuation feature coefficient, the instant correction coefficient is used to correct the initial temperature of the outer wall surface to obtain a corrected temperature of the outer wall surface; the preset wind speed fluctuation feature coefficient represents a leading or lag relationship between a wind speed fluctuation phase and a solar radiation intensity phase, and corresponds to an attenuation coefficient or an enhancement coefficient; If the solar activity feature is low-frequency fluctuation, a phase offset is calculated based on a phase difference between solar radiation intensity variation and outer wall surface temperature variation and a phase difference between wind speed variation and outer wall surface temperature variation, the phase offset is used to correct the initial temperature of the outer wall surface to obtain a corrected temperature of the outer wall surface.

[0007] Optionally, based on the outer wall surface temperature data, the solar radiation intensity data and the wind speed data of the preset time length, an outer wall surface temperature variation curve is constructed, including: Abnormal values are removed from the outer wall surface temperature data, the solar radiation intensity data and the wind speed data of the preset time length, short-time missing values are filled by using a linear interpolation method, and average temperature data of the outer wall surface temperature is obtained by averaging temperature data of different monitoring points of the wall body. A smooth curve is fitted by taking time as the horizontal axis and the average temperature data of the outer wall surface temperature as the vertical axis, and solar radiation intensity and wind speed peak points at corresponding time points are marked beside the curve to obtain the outer wall surface temperature variation curve.

[0008] Optionally, based on the outer wall surface temperature variation curve, amplitude spectrum and phase spectrum of each frequency component are generated, including: The outer wall surface temperature data, the solar radiation intensity data and the wind speed data are extracted from the outer wall surface temperature variation curve, and the time series of the three are synchronized by aligning time stamps; The outer wall surface temperature data, the solar radiation intensity data and the wind speed data are subjected to fast Fourier transform to convert time domain signals into frequency domain signals, and amplitude spectrum and phase spectrum of each frequency component are obtained.

[0009] Optionally, the outer wall surface temperature data, the solar radiation intensity data and the wind speed data are subjected to fast Fourier transform to convert time domain signals into frequency domain signals, and amplitude spectrum and phase spectrum of each frequency component are obtained, including: The outer wall surface temperature data, the solar radiation intensity data and the wind speed data are preprocessed, FFT algorithm is executed on the preprocessed data, samples of even indexes and odd indexes are separated, FFT of the two sub-sequences is recursively calculated, and the complete complex FFT result is obtained by merging the results; Calculate the modulus, phase of the complex FFT result, and the frequency axis based on the sampling frequency and data length, and draw the amplitude spectrum and phase spectrum of the external wall surface temperature data, solar radiation intensity data and wind speed data based on the calculated modulus, phase and frequency axis.

[0010] Optionally, based on the amplitude spectrum and phase spectrum of each frequency component, the measured amplitude peak value, the phase difference between the solar radiation intensity change and the external wall surface temperature change, and the phase difference between the wind speed change and the external wall surface temperature change are extracted, including: Identify the main frequency components from the amplitude spectrum and phase spectrum, select the frequencies with an amplitude ratio within the first preset proportion as characteristic frequencies, and extract the phase angle of the solar radiation intensity phase spectrum and the phase angle of the external wall surface temperature phase spectrum at the characteristic frequencies. The difference between the phase angle of the solar radiation intensity phase spectrum and the phase angle of the external wall surface temperature phase spectrum is the phase difference between the solar radiation intensity change and the external wall surface temperature change. Identify the main frequency components from the amplitude spectrum and phase spectrum, select the frequencies with an amplitude ratio within the first preset proportion as characteristic frequencies, and extract the phase angle of the wind speed phase spectrum and the phase angle of the external wall surface temperature phase spectrum at the characteristic frequencies. The difference between the phase angle of the wind speed phase spectrum and the phase angle of the external wall surface temperature phase spectrum is the phase difference between the wind speed change and the external wall surface temperature change. Wherein, the phase difference is positive, indicating that the environmental parameter change leads the external wall surface temperature change; the phase difference is negative, indicating that the external wall surface temperature change leads the environmental parameter change, and the phase difference statistics at each characteristic frequency are counted. Optionally, after the step of correcting the initial temperature of the external wall surface using a dynamic correction factor to obtain the corrected temperature of the external wall surface, the method further comprises: Compare the initial temperature spectrum of the external wall surface before and after correction and the corrected temperature spectrum of the external wall surface. If the high-frequency amplitude decreases to the background noise level and the low-frequency phase difference narrows to within the preset phase difference threshold range, it is determined that the correction is effective.

[0011] Optionally, a machine learning algorithm is used to calculate the heat transfer coefficient of the building external wall based on the corrected temperature of the external wall surface, the internal temperature of the external wall, the indoor air temperature and the outdoor air temperature, including: Standardize the corrected temperature of the external wall surface, the internal temperature of the external wall, the indoor air temperature and the outdoor air temperature; Input the standardized corrected temperature of the external wall surface, the internal temperature of the external wall, the indoor air temperature and the outdoor air temperature into a pre-trained artificial neural network model, and output the heat transfer coefficient of the building external wall.

[0012] The application also discloses an external wall heat transfer coefficient on-site detection system based on infrared thermal imaging technology, which comprises: The exterior wall surface initial temperature detection module is used to detect the temperature of the building's exterior wall surface using an infrared thermal imager to obtain the initial temperature of the exterior wall surface. The temperature and environmental parameter data acquisition module is used to synchronously collect exterior wall surface temperature data and environmental parameter data for a preset duration; The module for analyzing the dynamic relationship between temperature and environmental parameters is used to analyze the dynamic relationship between changes in environmental parameters and changes in external wall surface temperature based on the preset duration of external wall surface temperature data and environmental parameter data, using frequency domain analysis methods. The initial temperature correction module is used to calculate a dynamic correction factor for correcting the initial temperature of the external wall surface based on the dynamic relationship between changes in environmental parameter data and changes in the external wall surface temperature, and to correct the initial temperature of the external wall surface using the dynamic correction factor to obtain the corrected temperature of the external wall surface. The heat transfer coefficient calculation module is used to collect the temperature of the inner surface of the exterior wall, the indoor air temperature, and the outdoor air temperature. It uses machine learning algorithms to calculate the heat transfer coefficient of the building exterior wall based on the corrected temperature of the outer surface of the exterior wall, the temperature of the inner surface of the exterior wall, the indoor air temperature, and the outdoor air temperature.

[0013] This invention has the following advantages: This invention presents a method for on-site detection of exterior wall heat transfer coefficient based on infrared thermal imaging technology. It uses an infrared thermal imager to rapidly scan the exterior wall non-contactly, efficiently acquiring the initial temperature distribution, saving time and effort. A dynamic environmental correction mechanism is constructed, collecting 12 hours of wall temperature data along with solar and wind speed data. Phase difference characteristics are extracted, and amplitude and phase spectra are used to instantly correct coefficients and phase shifts, accurately eliminating the hysteresis interference from solar radiation and wind speed fluctuations, resulting in more accurate temperature data. The heat transfer coefficient calculated based on the corrected temperature data is accurate and reliable, providing crucial evidence for building energy-saving renovations and determining compliance. Furthermore, it is adaptable to different environments and exterior wall types, exhibiting strong versatility and assisting in energy-saving assessments and quality control in the building industry. Attached Figure Description

[0014] Figure 1 This is a flowchart illustrating the steps of an on-site detection method for the heat transfer coefficient of an external wall based on infrared thermal imaging technology, provided in an embodiment of the present invention. Detailed Implementation

[0015] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0016] Reference Figure 1 The diagram illustrates a flowchart of a method for on-site detection of the heat transfer coefficient of an external wall based on infrared thermal imaging technology, as provided in an embodiment of the present invention. Specifically, it may include the following steps: S1, an infrared thermal imager is used to detect the temperature of the building outer wall surface, and an initial temperature of the outer wall surface is obtained; S2, the temperature data of the outer wall surface and the environmental parameter data within a preset time length are synchronously collected; S3, based on the temperature data of the outer wall surface and the environmental parameter data within the preset time length, the dynamic relationship between the change of the environmental parameter data and the change of the temperature of the outer wall surface is analyzed by a frequency domain analysis method; S4, a dynamic correction factor for correcting the initial temperature of the outer wall surface is calculated based on the dynamic relationship between the change of the environmental parameter data and the change of the temperature of the outer wall surface, and the initial temperature of the outer wall surface is corrected by using the dynamic correction factor, so that a corrected temperature of the outer wall surface is obtained; S5, the temperature of the inner surface of the outer wall, the indoor air temperature and the outdoor air temperature are collected, and a machine learning algorithm is used to calculate the heat transfer coefficient of the building outer wall based on the corrected temperature of the outer surface of the outer wall, the temperature of the inner surface of the outer wall, the indoor air temperature and the outdoor air temperature.

[0017] The application effectively eliminates the interference of solar radiation, wind speed and the like by a dynamic environmental parameter correction mechanism and multi-dimensional data fusion, significantly improves the detection accuracy, and provides reliable data for heat transfer coefficient calculation; relying on 12 hours of monitoring data and a machine learning algorithm, subjective bias is reduced, and the stability and reliability of the results are enhanced; the infrared thermal imager quickly collects data combined with automatic processing, shortens the detection period, reduces repetitive work, and improves efficiency; accurate heat transfer coefficient can accurately evaluate the thermal insulation performance of the outer wall, provide scientific basis for building energy saving reconstruction, help reduce building energy consumption, and promote the development of energy saving technology.

[0018] In step S1, a sunny day is selected for detection, the infrared thermal imager is kept perpendicular to the wall surface, the center of the lens is aligned with the center of the detection area, the distance is controlled within the range of 5-20m, the building outer wall is divided into several 10m*10m detection areas according to the facade, more than 3 thermal images are taken for each area, the shooting is focused on the wall surface, the initial temperature data obtained by shooting is averaged and reserved.

[0019] The operation specification of the application can guarantee the detection quality from multiple aspects, the selection of a sunny day can reduce the interference of light fluctuation, and provide stable basic data for subsequent correction; the 10m*10m area is divided and more than 3 thermal images are taken, which can avoid omission and eliminate accidental errors after averaging, and improve the representativeness of data; the lens is perpendicular to the wall, the center is aligned, and the distance is controlled within 5-20m, which can reduce the measurement deviation caused by angle and distance, and ensure the temperature measurement accuracy; focusing on the wall can avoid background interference and ensure clear temperature signal.

[0020] The unified operation specification makes the data comparable, facilitates overall analysis, and the average processing can also reduce the subsequent correction pressure, these details reduce interference through the standardized process, ensure that the initial data is real and stable, and provide reliable support for subsequent correction and calculation.

[0021] In the S2 step, the external wall surface temperature data and the environmental parameter data of a preset time length are synchronously collected: Three temperature monitoring points are uniformly arranged on the wall surface, thermocouple sensors are used, a solar radiation sensor and an anemometer are synchronously erected, and 12 hours of data collection is performed. In the S3 step, based on the external wall surface temperature data and the environmental parameter data of the preset time length, an external wall surface temperature change curve is constructed, the change characteristics of the environmental parameter data and the change characteristics of the external wall surface temperature are analyzed through a frequency domain analysis method, the dynamic relationship between the two is analyzed, and a dynamic environmental parameter correction mechanism is established.

[0022] In an optional embodiment of the present application, based on the external wall surface temperature data and the environmental parameter data of the preset time length, the dynamic relationship between the change of the environmental parameter data and the change of the external wall surface temperature is analyzed through a frequency domain analysis method, including: Based on the external wall surface temperature data, the solar radiation intensity data and the wind speed data of the preset time length, an external wall surface temperature change curve is constructed; the external wall surface temperature change curve shows the dynamic evolution law of temperature with time, and directly reflects the time sequence corresponding relationship between the key environmental factors of solar radiation intensity and wind speed and the wall temperature change; Based on the external wall surface temperature change curve, amplitude spectrum and phase spectrum of each frequency component are generated; Based on the amplitude spectrum and the phase spectrum of each frequency component, the measured amplitude peak value, the phase difference between the change of solar radiation intensity and the change of the external wall surface temperature, and the phase difference between the change of wind speed and the change of the external wall surface temperature are extracted.

[0023] In an optional embodiment of the present application, based on the external wall surface temperature data, the solar radiation intensity data and the wind speed data of the preset time length, an external wall surface temperature change curve is constructed, including: Abnormal values are removed from the external wall surface temperature data, the solar radiation intensity data and the wind speed data of the preset time length, linear interpolation method is used to fill in short missing values, and the temperature data of different monitoring points of the same wall is averaged to obtain the average temperature data of the external wall surface temperature; Taking time as the horizontal axis and the average temperature data of the external wall surface temperature as the vertical axis, a smooth curve is fitted and generated, the solar radiation intensity and the wind speed peak point at the corresponding moment are marked beside the curve, and the external wall surface temperature change curve is obtained.

[0024] In an optional embodiment of the present application, based on the external wall surface temperature change curve, amplitude spectrum and phase spectrum of each frequency component are generated, including: Extract the external wall surface temperature data, solar radiation intensity data and wind speed data in the external wall surface temperature change curve, align them according to the time stamp, and synchronize the time series of the three; Perform fast Fourier transform on the external wall surface temperature data, solar radiation intensity data and wind speed data to convert the time domain signal to the frequency domain signal, and obtain the amplitude spectrum and phase spectrum of each frequency component.

[0025] In an optional embodiment of the present application, the fast Fourier transform is performed on the external wall surface temperature data, solar radiation intensity data and wind speed data to convert the time domain signal to the frequency domain signal, and obtain the amplitude spectrum and phase spectrum of each frequency component, including: The external wall surface temperature data, solar radiation intensity data and wind speed data are preprocessed, the FFT algorithm is executed on the preprocessed data, the samples of even index and odd index are separated, the FFT of the two sub-sequences is recursively calculated, and the complete complex FFT result is obtained by merging the results; The modulus and phase of the complex FFT result are calculated, the frequency axis is calculated based on the sampling frequency and data length, and the amplitude spectrum and phase spectrum of the external wall surface temperature data, solar radiation intensity data and wind speed data are plotted based on the calculated modulus, phase and frequency axis.

[0026] In an optional embodiment of the present application, the measured amplitude peak value, the phase difference between the solar radiation intensity change and the external wall surface temperature change, and the phase difference between the wind speed change and the external wall surface temperature change are extracted based on the amplitude spectrum and phase spectrum of each frequency component, including: The main frequency components are identified from the amplitude spectrum and phase spectrum, the frequencies with an amplitude ratio within the first preset proportion are selected as characteristic frequencies, the phase angle of the solar radiation intensity phase spectrum and the phase angle of the external wall surface temperature phase spectrum are extracted at the characteristic frequencies, and the difference between the phase angle of the solar radiation intensity phase spectrum and the phase angle of the external wall surface temperature phase spectrum is the phase difference between the solar radiation intensity change and the external wall surface temperature change; The main frequency components are identified from the amplitude spectrum and phase spectrum, the frequencies with an amplitude ratio within the first preset proportion are selected as characteristic frequencies, the phase angle of the wind speed phase spectrum and the phase angle of the external wall surface temperature phase spectrum are extracted at the characteristic frequencies, and the difference between the phase angle of the wind speed phase spectrum and the phase angle of the external wall surface temperature phase spectrum is the phase difference between the wind speed change and the external wall surface temperature change; wherein, the phase difference is positive, indicating that the environmental parameter change leads the external wall surface temperature change; the phase difference is negative, indicating that the external wall surface temperature change leads the environmental parameter change, and the phase difference statistics of each characteristic frequency are counted. The present application first establishes a coordinate system framework with time as the horizontal axis (accurate to the minute level, covering the complete 12-hour collection period), wall average temperature as the vertical axis (unit: ℃, range set according to actual monitoring range) when constructing the outer wall surface temperature change curve, providing a clear dimensional reference for data visualization. The pre-processed wall average temperature data every 5 minutes (abnormal value elimination, missing value filling and multi-point average calculation have been completed) is imported into professional data processing software, and the software will preliminarily connect based on these discrete data points to form an original curve reflecting the fluctuation of temperature with time.

[0027] Since the original curve may contain high-frequency noise (manifested as irregular sharp fluctuations in the local curve) caused by instantaneous airflow and sensor micro-vibration factors, moving average method is used for smoothing processing: a reasonable sliding window (such as 3 consecutive data points as a window) is set, the temperature average value in each window is calculated and the original value of the window center data point is replaced, and the iteration processing of the whole period data is completed by sliding the window point by point. This process can effectively filter short-period and small-amplitude temperature fluctuations, making the curve shape more gentle and highlighting the overall trend of wall temperature changes with the environment (such as the temperature rise stage affected by solar radiation and the temperature drop stage of night heat dissipation). To intuitively present the correlation between environmental parameters and wall temperature, the solar radiation intensity (unit: W / ㎡, using numerical labels or broken line secondary axis form) and wind speed peak point (unit: m / s, marked with arrows or special symbols) need to be accurately marked beside the corresponding time nodes of the smoothed curve: for example, at the curve position around 12 o'clock noon, the maximum value of solar radiation intensity at this period is marked simultaneously; at the time of the gust at 3 o'clock in the afternoon, the wind speed mutation point is marked with a peak symbol. Through this multi-parameter linkage marking, the final outer wall surface temperature change curve not only clearly shows the dynamic evolution law of temperature with time, but also intuitively reflects the time sequence corresponding relationship between solar radiation intensity, wind speed key environmental factors and wall temperature change, providing visual analysis basis for subsequent extraction of phase difference features and temperature data correction.

[0028] The steps for extracting amplitude and phase difference features are: Extract wall temperature data, solar radiation intensity data and wind speed data in the temperature change curve, align them by timestamp, and synchronize the three time series; perform fast Fourier transform on the three sets of data, and perform fast Fourier transform (FFT) on the three sets of time series data. The core role of this transformation is to convert the signal originally presented in the time domain (with time as the variable) to the frequency domain (with frequency as the variable), thereby revealing the periodicity hidden in the data. Through the transformation, the amplitude spectrum and phase spectrum corresponding to each frequency component can be obtained: the amplitude spectrum reflects the energy proportion of different frequency components in the data, and the larger the amplitude, the more significant the periodic change corresponding to the frequency; the phase spectrum represents the starting position of each frequency component in time, i.e., the phase angle, which can reflect the time difference of different signals in periodic change.

[0029] To focus on key periodic features, the frequency components need to be screened. Identify all frequency components from the amplitude spectrum, calculate the proportion of the amplitude of each frequency component in the total amplitude, and select the top 30% of the frequency with the largest amplitude as the characteristic frequency. These characteristic frequencies correspond to the strongest energy and most representative periodic fluctuations in the data, effectively reflecting the main rules of solar radiation, wind speed and wall temperature changes, and excluding the interference of secondary frequency components.

[0030] Under the determined characteristic frequency, the phase angles in the solar radiation intensity phase spectrum and the wall temperature phase spectrum are extracted. Since the phase angle reflects the position of the signal in the cycle, the difference between the two is the phase difference of solar radiation and wall temperature change at the characteristic frequency. Similarly, the phase angles of the wind speed phase spectrum and the wall temperature phase spectrum are extracted and the difference is calculated to obtain the phase difference of wind speed and wall temperature change. The positive and negative of the phase difference has a clear physical meaning: when the phase difference is positive, it means that in the periodic change corresponding to the characteristic frequency, the change of the environmental parameter (solar radiation or wind speed) leads the wall temperature change, i.e., the fluctuation of the environmental parameter occurs first, and the wall temperature responds subsequently; when the phase difference is negative, it means that the wall temperature change leads the environmental parameter change, i.e., the fluctuation of the wall temperature occurs before the fluctuation of the environmental parameter. The phase difference of solar radiation and wall temperature, and the phase difference of wind speed and wall temperature obtained for all characteristic frequencies are statistically analyzed to form a systematic phase difference feature data set, which provides a key phase relationship basis for subsequent temperature data correction.

[0031] Before performing the fast Fourier transform (FFT), the data of each group (wall temperature, solar radiation intensity, wind speed data) needs to be preprocessed. This step aims to eliminate the trend items and direct current components that may exist in the data, such as removing the slow drift of wall temperature data over time through linear fitting, or subtracting the average value of wind speed data to exclude the interference of constant airflow, to ensure that the subsequent transformation results accurately reflect the periodic fluctuation characteristics of the data. If the data length does not meet the requirement of being an integer power of 2, zero padding needs to be performed to meet the data length requirement of the FFT algorithm, improving the calculation efficiency and spectral resolution. After preprocessing, the FFT algorithm is executed for each group of data: first, the data samples are split into two sub-sequences according to the parity of the index, for example, the original sequence x (0), x (1), x (2),..., x (N-1) is divided into even sub-sequence x (0), x (2),..., x (N-2) and odd sub-sequence x (1), x (3),..., x (N-1), then recursive FFT calculation is performed on the two sub-sequences until the sub-sequence length is reduced to 1; in the recursive process, the transformation results of the two sub-sequences are merged using the rotation factor (based on the periodicity of the complex exponential function), and the butterfly operation structure is used to reduce repeated calculations, finally the complete complex FFT result is obtained, which contains the amplitude and phase information of the data at different frequencies.

[0032] Based on the complex FFT result, three key parameters are further calculated: first, the modulus of the complex result (i.e. amplitude), which is obtained by taking the square root of the sum of the squares of the real and imaginary parts, it reflects the energy intensity of the corresponding frequency component; second, the phase of the result, which is calculated by the argument of the complex number, representing the starting position of the frequency component on the time axis; third, the frequency axis, according to the sampling frequency (such as 1 sample every 5 minutes, corresponding to 0.00033 Hz) and the data length (N value after zero padding), the actual frequency corresponding to each FFT result is calculated according to the formula "frequency = sampling frequency x index / N", ensuring that the horizontal scale of the spectrum diagram accurately corresponds to the physical frequency.

[0033] With frequency as the horizontal axis and amplitude / phase as the vertical axis, the amplitude spectrum and phase spectrum of the three groups of data are drawn respectively: in the amplitude spectrum, the higher the peak value, the more significant the periodic fluctuation of the corresponding frequency (such as the peak value of solar radiation data near 0.0417 Hz corresponding to the diurnal cycle); in the phase spectrum, the phase values at different frequencies directly show the time offset characteristics of the signal. Through these graphs, the main frequency components and their energy distribution in the data can be clearly identified, providing visual analysis basis for subsequent feature frequency selection and phase difference calculation.

[0034] In an optional embodiment of the present invention, a dynamic correction factor for correcting the initial temperature of the exterior wall surface is calculated based on the dynamic relationship between changes in environmental parameter data and changes in the exterior wall surface temperature, and the initial temperature of the exterior wall surface is corrected using the dynamic correction factor to obtain the corrected temperature of the exterior wall surface, including: Interpreting solar activity characteristics and wind speed fluctuation characteristics corresponding to frequency components based on amplitude and phase spectra; If the solar activity characteristics are high-frequency fluctuations, then the instantaneous correction coefficient is calculated based on the measured peak amplitude and the preset wind speed fluctuation characteristic coefficient. The instantaneous correction coefficient is used to correct the initial temperature of the outer wall surface to obtain the corrected temperature of the outer wall surface. The preset wind speed fluctuation characteristic coefficient represents the leading or lagging relationship between the wind speed fluctuation phase and the solar radiation intensity phase, and corresponds to the attenuation coefficient or the enhancement coefficient. If the solar activity is characterized by low-frequency fluctuations, the phase shift is calculated based on the phase difference between the changes in solar radiation intensity and the changes in the external wall surface temperature, and the phase difference between the changes in wind speed and the changes in the external wall surface temperature. The phase shift is then used to correct the initial temperature of the external wall surface, resulting in the corrected temperature of the external wall surface.

[0035] This invention interprets solar activity corresponding to frequency components based on amplitude and phase spectra. It uses an instant correction coefficient for high-frequency fluctuations and introduces a phase offset for low-frequency fluctuations to correct the initial temperature data of the wall surface, thereby eliminating the lag interference of solar radiation and wind speed fluctuations on the detection results. (1) Real-time correction of high-frequency fluctuations: To capture pulsed disturbances in solar transient activity with high-frequency fluctuations >1Hz, the relative energy ratio is calculated by continuously monitoring the amplitude peak value in this frequency band, with the correction factor being: K = 1 + (measured amplitude / reference amplitude - 1) × 0.6; Where K is the correction coefficient, the reference amplitude is the average peak value under the same period of calm weather in history, and the coefficient of 0.6 is used to balance the risk of overcorrection; when the wind speed fluctuation phase leads the solar radiation by more than 0.2π, the K value is multiplied by the attenuation coefficient of 0.8; otherwise, it is multiplied by the enhancement coefficient of 1.2. The real-time update frequency is once per second to correct the initial temperature data of the exterior wall surface.

[0036] This invention precisely addresses the pulse-like interference from solar transients. High-frequency fluctuations >1Hz can easily distort initial temperature data. By continuously monitoring the peak amplitude of this frequency band and calculating the relative energy ratio, the temperature data can be reasonably adjusted based on the correction coefficient formula, thus balancing the risk of overcorrection.

[0037] When the phase difference between the wind speed and the solar radiation is different, the interaction of the two can be considered by multiplying the attenuation or enhancement coefficient, so that the correction is more in line with the actual situation, and the real-time update of 1 second can timely follow the high-frequency fluctuation change and ensure the timeliness of the correction, so as to finally improve the accuracy of the initial temperature data and lay a reliable foundation for the calculation of the subsequent heat transfer coefficient.

[0038] (2) Introduce the phase offset amount by low-frequency fluctuation: Calculate the offset amount Δt according to the phase difference in the low-frequency band: Δt = (measured phase difference / 2π) x period; Wherein the period takes the 12-hour basic value; Build a temperature compensation model: The corrected temperature T'=Tinitial+(Tinitial(t+Δt)-Tinitial(t) x 0.7; The cumulative error of long-period correction is weakened by the weight coefficient of 0.7, and the offset amount is updated once an hour to correct the temperature.

[0039] The change rhythm of the low-frequency fluctuation is slow (such as the trend change in 12 hours), and the offset amount Δt is updated once an hour, which can timely respond to the slow change of the phase difference (such as the phase difference fine adjustment generated by the sun angle and day and night alternation), and avoid unnecessary calculation consumption caused by high-frequency update. The update frequency is matched with the time scale of the low-frequency fluctuation, so that the correction parameter is always synchronized with the lag characteristics of the current environment, and the rationality of the correction is further improved.

[0040] The correction method quantifies the time offset, balances the correction weight, and dynamically updates the parameter, accurately solves the lag problem of the environmental parameter and the wall temperature in the low-frequency fluctuation, retains the trend characteristics of the temperature data, effectively controls the error accumulation of the long-period correction, makes the corrected temperature data more in line with the actual heat response law of the wall, and provides a key guarantee for the accurate calculation of the heat transfer coefficient in the S5 step.

[0041] In an optional embodiment of the present application, after the step of correcting the initial temperature of the outer wall surface by using the dynamic correction factor to obtain the corrected temperature of the outer wall surface, the method further comprises: Compare the initial temperature spectrum of the outer wall surface before and after correction and the corrected temperature spectrum of the outer wall surface, if the high-frequency amplitude is reduced to the background noise level and the low-frequency phase difference is reduced to the preset phase difference threshold range, it is judged that the correction is effective.

[0042] Compare the data spectrum before and after correction, if the high-frequency amplitude is reduced to the background noise level and the low-frequency phase difference is reduced to ±0.1π, it is judged that the correction is effective.

[0043] In an optional embodiment of the present application, a machine learning algorithm is used to calculate the heat transfer coefficient of the building outer wall based on the outer wall surface correction temperature, the outer wall surface temperature, the indoor air temperature and the outdoor air temperature, comprising: standardizing the outer wall surface correction temperature, the outer wall surface temperature, the indoor air temperature and the outdoor air temperature; inputting the standardized outer wall surface correction temperature, the outer wall surface temperature, the indoor air temperature and the outdoor air temperature into a pre-trained artificial neural network model, and outputting the heat transfer coefficient of the building outer wall.

[0044] The outer wall surface temperature in step S5 is scanned and collected by an infrared thermal imager, capturing the overall temperature distribution of the wall facing the indoor. The wall surface temperature correction excludes the interference of direct sunlight and weather factors. The indoor air temperature collection point is selected at the center of the space away from doors and windows and heat sources, and the height is controlled at 1.2-1.5m for collection. The outdoor air temperature is measured at a place with good ventilation and sunshade around the building.

[0045] The artificial neural network is used to calculate the wall heat transfer coefficient K value from the temperature parameters. The four collected temperature variables are inputted, and through the process of "data standardization → feature mapping → nonlinear conversion → result output", the four temperature variables are outputted as the wall heat transfer coefficient K value. The model is trained based on the measured data.

[0046] In the data standardization stage, the four temperature variables need to be uniformly processed. Since there are differences in the numerical ranges of different temperature parameters, such as the large indoor and outdoor air temperature difference and the relatively small wall surface temperature difference, direct input into the model will affect the calculation accuracy. Through standardization processing, these temperature values are mapped to the same numerical interval, which not only preserves the relative relationship between the temperatures, such as the temperature difference trend between indoor air and inner surface, and the temperature difference characteristics between outer surface and outdoor air, but also eliminates the interference of different magnitudes, so that the model can focus more on the internal relationship between temperature parameters.

[0047] The feature mapping link is to input the standardized temperature data into the input layer of the neural network, and then realize feature transmission through the weight connection between the input layer and the hidden layer. After receiving the four temperature variables, the input layer will distribute the information to each neuron of the hidden layer according to the preset weight matrix. This process is not a simple numerical transmission, but a preliminary screening and reorganization of temperature features combined with the rules learned in the training process, such as highlighting the temperature combination features that have a significant impact on K value, laying a foundation for subsequent deep operation.

[0048] The nonlinear conversion is mainly completed in the hidden layer of the neural network, the hidden layer processes the received features through a nonlinear activation function, and the processing can simulate the complex nonlinear relationship between the temperature parameters and the K value. In the heat transfer process, the K value is not simply linearly fluctuated with the temperature, but is cooperatively affected by multiple temperature parameters, and presents a complex change rule. The multi-layer structure of the hidden layer can gradually deepen the processing of the features, and extract high-order features from the basic temperature values, such as the influence of the interaction between different temperature gradients on the K value, and constantly approach the real heat transfer rule.

[0049] After the processing of the multiple hidden layers, the information is finally transmitted to the output layer, the output layer calculates and outputs the wall heat transfer coefficient K value, the output layer will combine the optimized weight parameters according to the feature information transmitted by the last hidden layer, and obtain the specific K value result through a linear activation function, so that the output value meets the actual physical meaning and is within a reasonable range.

[0050] In an optional embodiment of the present application, the collected initial temperature of the outer wall surface includes the initial temperature of the outer surface of the outer wall and the initial temperature of the inner surface of the outer wall; the initial temperature of the outer surface of the outer wall and the initial temperature of the inner surface of the outer wall are corrected to obtain the corrected temperature of the outer surface of the outer wall and the corrected temperature of the inner surface of the outer wall; and a machine learning algorithm is used to calculate the heat transfer coefficient of the building outer wall based on the corrected temperature of the outer surface of the outer wall, the corrected temperature of the inner surface of the outer wall, the indoor air temperature and the outdoor air temperature.

[0051] It should be noted that, for the method embodiments, in order to simply describe, they are all expressed as a series of action combinations, but those skilled in the art should know that the embodiments of the present application are not limited by the action sequence described, because according to the embodiments of the present application, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should know that the embodiments described in the specification all belong to preferred embodiments, and the actions involved are not necessarily necessary for the embodiments of the present application.

[0052] The infrared thermal imaging technology-based outer wall heat transfer coefficient field detection system provided in the embodiments of the present application can specifically include the following modules: The outer wall surface initial temperature detection module is used to detect the temperature of the outer wall surface of the building by using an infrared thermal imager to obtain the initial temperature of the outer wall surface; The temperature and environmental parameter data acquisition module is used to synchronously acquire the outer wall surface temperature data and the environmental parameter data for a preset time length; The temperature and environmental parameter dynamic relationship analysis module is used to analyze the dynamic relationship between the change of the environmental parameter data and the change of the outer wall surface temperature based on the outer wall surface temperature data and the environmental parameter data for the preset time length through a frequency domain analysis method; An initial temperature correction module is configured to calculate a dynamic correction factor for correcting the initial temperature of the outer wall surface based on a dynamic relationship between the environmental parameter data and the change in the outer wall surface temperature, and correct the initial temperature of the outer wall surface by using the dynamic correction factor to obtain a corrected temperature of the outer wall surface; A heat transfer coefficient calculation module is configured to collect the inner surface temperature of the outer wall, the indoor air temperature and the outdoor air temperature, and calculate the heat transfer coefficient of the building outer wall based on the corrected temperature of the outer wall surface, the inner surface temperature of the outer wall, the indoor air temperature and the outdoor air temperature by using a machine learning algorithm.

[0053] Optionally, the environmental parameter data at least includes solar radiation intensity and wind speed; and the temperature and environmental parameter dynamic relationship analysis module is further configured to: Based on the outer wall surface temperature data, the solar radiation intensity data and the wind speed data of the preset time length, a wall surface temperature change curve is constructed, which shows the dynamic evolution law of temperature with time and directly reflects the time sequence corresponding relationship between the key environmental factors of solar radiation intensity and wind speed and the wall temperature change; Based on the wall surface temperature change curve, amplitude spectrum and phase spectrum of each frequency component are generated; Based on the amplitude spectrum and phase spectrum of each frequency component, the measured amplitude peak value, the phase difference between the solar radiation intensity change and the outer wall surface temperature change, and the phase difference between the wind speed change and the outer wall surface temperature change are extracted.

[0054] Optionally, the initial temperature correction module is further configured to: Based on the amplitude spectrum and phase spectrum, the solar activity characteristics and wind speed fluctuation characteristics corresponding to the frequency component are interpreted; If the solar activity characteristics are high-frequency fluctuations, an instant correction coefficient is calculated based on the measured amplitude peak value and a preset wind speed fluctuation characteristic coefficient, and the initial temperature of the outer wall surface is corrected by using the instant correction coefficient to obtain the corrected temperature of the outer wall surface; the preset wind speed fluctuation characteristic coefficient represents the leading or lagging relationship between the wind speed fluctuation phase and the solar radiation intensity phase, and corresponds to a decay coefficient or an enhancement coefficient; If the solar activity characteristics are low-frequency fluctuations, a phase offset is calculated based on the phase difference between the solar radiation intensity change and the outer wall surface temperature change, and the phase difference between the wind speed change and the outer wall surface temperature change, and the initial temperature of the outer wall surface is corrected by using the phase offset to obtain the corrected temperature of the outer wall surface.

[0055] Optionally, the temperature and environmental parameter dynamic relationship analysis module is further configured to: Abnormal values are eliminated from the outer wall surface temperature data, the solar radiation intensity data and the wind speed data of the preset time length, and linear interpolation is used to fill in short-time missing values, and the temperature data of different monitoring points of the same wall is averaged to obtain average temperature data of the outer wall surface temperature. Take time as the horizontal axis, and the average temperature data of the external wall surface temperature as the vertical axis, generate a smooth curve by fitting, and mark the solar radiation intensity and the peak point of the wind speed at the corresponding time on the curve to obtain the external wall surface temperature change curve.

[0056] Optionally, the temperature and environmental parameter dynamic relationship analysis module is further used for: Extracting the external wall surface temperature data, the solar radiation intensity data, and the wind speed data in the external wall surface temperature change curve, aligning them according to the time stamp, and synchronizing the time series of the three; Performing fast Fourier transform on the external wall surface temperature data, the solar radiation intensity data, and the wind speed data to convert the time domain signals into frequency domain signals, and obtaining the amplitude spectrum and the phase spectrum of each frequency component.

[0057] Optionally, the temperature and environmental parameter dynamic relationship analysis module is further used for: Performing preprocessing on the external wall surface temperature data, the solar radiation intensity data, and the wind speed data, executing the FFT algorithm on the preprocessed data, separating the samples with even indexes and the samples with odd indexes, recursively calculating the FFT of the two subsequences, and merging the results to obtain the complete complex FFT result; Calculating the modulus and the phase of the complex FFT result, calculating the frequency axis based on the sampling frequency and the data length, and drawing the amplitude spectrum and the phase spectrum of the external wall surface temperature data, the solar radiation intensity data, and the wind speed data based on the calculated modulus, phase, and frequency axis.

[0058] Optionally, the temperature and environmental parameter dynamic relationship analysis module is further used for: Identifying the main frequency components from the amplitude spectrum and the phase spectrum, screening out the frequencies with an amplitude proportion within a preset proportion as characteristic frequencies, extracting the phase angle of the solar radiation intensity phase spectrum and the phase angle of the external wall surface temperature phase spectrum at the characteristic frequencies, and taking the difference between the phase angle of the solar radiation intensity phase spectrum and the phase angle of the external wall surface temperature phase spectrum as the phase difference between the solar radiation intensity change and the external wall surface temperature change; Identifying the main frequency components from the amplitude spectrum and the phase spectrum, screening out the frequencies with an amplitude proportion within a preset proportion as characteristic frequencies, extracting the phase angle of the wind speed phase spectrum and the phase angle of the external wall surface temperature phase spectrum at the characteristic frequencies, and taking the difference between the phase angle of the wind speed phase spectrum and the phase angle of the external wall surface temperature phase spectrum as the phase difference between the wind speed change and the external wall surface temperature change; wherein, the phase difference being positive indicates that the environmental parameter change leads the external wall surface temperature change, and the phase difference being negative indicates that the external wall surface temperature change leads the environmental parameter change, and the phase difference statistics at each characteristic frequency are counted. Optionally, the system further comprises a correction validity verification module, and the correction validity verification module is used for: The initial temperature spectrum of the outer wall surface before correction and the corrected temperature spectrum of the outer wall surface are compared, if the amplitude of the high frequency band is reduced to the background noise level and the phase difference of the low frequency band is reduced to within the preset phase difference threshold range, it is determined that the correction is effective.

[0059] Optionally, the heat transfer coefficient calculation module is further configured to: standardize the corrected temperature of the outer surface of the outer wall, the temperature of the inner surface of the outer wall, the indoor air temperature, and the outdoor air temperature; input the standardized corrected temperature of the outer surface of the outer wall, the temperature of the inner surface of the outer wall, the indoor air temperature, and the outdoor air temperature into a pre-trained artificial neural network model, and output the heat transfer coefficient of the building outer wall.

[0060] For the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts are referred to the part of the method embodiment.

[0061] It should be noted that, in this document, the terms such as first and second are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment. Without more limitations, the element defined by the statement "including a" does not exclude the presence of another identical element in the process, method, article or equipment including the element.

[0062] Each embodiment in the specification is described in a related manner, and the same and similar parts between each embodiment can be referred to each other. Each embodiment focuses on the difference from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts are referred to the part of the method embodiment.

[0063] The above only describes the preferred embodiments of the present application, and is not used to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application is included in the protection scope of the present application.

Claims

1. A method for detecting the heat transfer coefficient of an external wall in situ based on infrared thermography, characterized in that, The method comprises: The method comprises: Synchronously collecting the temperature data of the outer wall surface and the environmental parameter data for a preset time length; Based on the temperature data of the outer wall surface and the environmental parameter data for the preset time length, the dynamic relationship between the change of the environmental parameter data and the change of the temperature of the outer wall surface is analyzed by a frequency domain analysis method; Based on the dynamic relationship between the change of the environmental parameter data and the change of the temperature of the outer wall surface, a dynamic correction factor for correcting the initial temperature of the outer wall surface is calculated, and the initial temperature of the outer wall surface is corrected by using the dynamic correction factor to obtain a corrected temperature of the outer wall surface; the corrected temperature of the outer wall surface comprises a corrected temperature of the outer surface of the outer wall; The temperature of the inner surface of the outer wall, the indoor air temperature and the outdoor air temperature are collected, and a machine learning algorithm is used to calculate the heat transfer coefficient of the building outer wall based on the corrected temperature of the outer surface of the outer wall, the temperature of the inner surface of the outer wall, the indoor air temperature and the outdoor air temperature.

2. The method of claim 1, wherein, The environmental parameter data at least comprises solar radiation intensity and wind speed; based on the temperature data of the outer wall surface and the environmental parameter data for the preset time length, the dynamic relationship between the change of the environmental parameter data and the change of the temperature of the outer wall surface is analyzed by a frequency domain analysis method, which comprises: Based on the temperature data of the outer wall surface, the solar radiation intensity data and the wind speed data for the preset time length, a temperature change curve of the outer wall surface is constructed; the temperature change curve of the outer wall surface shows the dynamic evolution law of temperature with time, and directly reflects the time sequence corresponding relationship between the key environmental factors of solar radiation intensity and wind speed and the temperature change of the wall; Based on the temperature change curve of the outer wall surface, amplitude spectrum and phase spectrum of each frequency component are generated; Based on the amplitude spectrum and the phase spectrum of each frequency component, the measured amplitude peak value, the phase difference between the change of solar radiation intensity and the change of the temperature of the outer wall surface, and the phase difference between the change of wind speed and the change of the temperature of the outer wall surface are extracted.

3. The method of claim 2, wherein, Based on the dynamic relationship between the change of the environmental parameter data and the change of the temperature of the outer wall surface, a dynamic correction factor for correcting the initial temperature of the outer wall surface is calculated, and the initial temperature of the outer wall surface is corrected by using the dynamic correction factor to obtain a corrected temperature of the outer wall surface, which comprises: Based on the amplitude spectrum and the phase spectrum, the solar activity characteristics and the wind speed fluctuation characteristics corresponding to the frequency component are interpreted; If the solar activity characteristics are high-frequency fluctuations, an instant correction coefficient is calculated based on the measured amplitude peak value and a preset wind speed fluctuation characteristic coefficient, and the initial temperature of the outer wall surface is corrected by using the instant correction coefficient to obtain the corrected temperature of the outer wall surface; the preset wind speed fluctuation characteristic coefficient represents the leading or lagging relationship between the wind speed fluctuation phase and the solar radiation intensity phase, which corresponds to a decay coefficient or an enhancement coefficient; If the solar activity characteristics are low-frequency fluctuations, a phase offset is calculated based on the phase difference between the change of solar radiation intensity and the change of the temperature of the outer wall surface, and the phase difference between the change of wind speed and the change of the temperature of the outer wall surface, and the initial temperature of the outer wall surface is corrected by using the phase offset to obtain the corrected temperature of the outer wall surface.

4. The method of claim 2, wherein, Based on the temperature data of the outer wall surface, the solar radiation intensity data and the wind speed data for the preset time length, a temperature change curve of the outer wall surface is constructed, which comprises: The average temperature data of the external wall surface temperature is obtained by removing outliers from the preset length of the external wall surface temperature data, the solar radiation intensity data and the wind speed data, filling in short missing values by using linear interpolation method, and averaging the temperature data of different monitoring points of the wall body; Taking time as the horizontal axis and the average temperature data of the external wall surface temperature as the vertical axis, a smooth curve is fitted to obtain the external wall surface temperature change curve, and the solar radiation intensity and the wind speed peak point at the corresponding time are marked on the curve.

5. The method of claim 2, wherein, Based on the external wall surface temperature change curve, the amplitude spectrum and the phase spectrum of each frequency component are generated, including: The external wall surface temperature data, the solar radiation intensity data and the wind speed data are extracted from the external wall surface temperature change curve, and the three time series are synchronized by aligning the time stamps; The external wall surface temperature data, the solar radiation intensity data and the wind speed data are subjected to fast Fourier transform to convert time domain signals into frequency domain signals, and the amplitude spectrum and the phase spectrum of each frequency component are obtained.

6. The method of claim 5, wherein, The external wall surface temperature data, the solar radiation intensity data and the wind speed data are subjected to fast Fourier transform to convert time domain signals into frequency domain signals, and the amplitude spectrum and the phase spectrum of each frequency component are obtained, including: The external wall surface temperature data, the solar radiation intensity data and the wind speed data are preprocessed, and the FFT algorithm is executed on the preprocessed data, the samples with even and odd indexes are separated, the FFT of the two sub-sequences is calculated recursively, and the complete complex FFT result is obtained by merging the results; The modulus and phase of the complex FFT result are calculated, the frequency axis is calculated based on the sampling frequency and the data length, and the amplitude spectrum and the phase spectrum of the external wall surface temperature data, the solar radiation intensity data and the wind speed data are plotted based on the calculated modulus, phase and frequency axis.

7. The method of claim 2, wherein, Based on the amplitude spectrum and the phase spectrum of each frequency component, the measured amplitude peak value, the phase difference between the solar radiation intensity change and the external wall surface temperature change, and the phase difference between the wind speed change and the external wall surface temperature change are extracted, including: The main frequency components are identified from the amplitude spectrum and the phase spectrum, and the frequencies with an amplitude ratio within a preset proportion are selected as characteristic frequencies. At the characteristic frequencies, the phase angle of the solar radiation intensity phase spectrum and the phase angle of the external wall surface temperature phase spectrum are extracted, and the difference between the phase angle of the solar radiation intensity phase spectrum and the phase angle of the external wall surface temperature phase spectrum is the phase difference between the solar radiation intensity change and the external wall surface temperature change. The main frequency components are identified from the amplitude spectrum and the phase spectrum, and the frequencies with an amplitude ratio within a preset proportion are selected as characteristic frequencies. At the characteristic frequencies, the phase angle of the wind speed phase spectrum and the phase angle of the external wall surface temperature phase spectrum are extracted, and the difference between the phase angle of the wind speed phase spectrum and the phase angle of the external wall surface temperature phase spectrum is the phase difference between the wind speed change and the external wall surface temperature change. Wherein, the phase difference is positive, indicating that the environmental parameter change leads the external wall surface temperature change; the phase difference is negative, indicating that the external wall surface temperature change leads the environmental parameter change, and the phase difference statistics of each characteristic frequency are counted.

8. The method of claim 1, wherein, After the step of correcting the initial temperature of the external wall surface using a dynamic correction factor to obtain the corrected temperature of the external wall surface, the method further includes: The initial temperature spectrum of the outer wall surface before correction and the corrected temperature spectrum of the outer wall surface are compared, and if the amplitude of the high frequency band is reduced to the background noise level and the phase difference of the low frequency band is reduced to within the preset phase difference threshold range, it is determined that the correction is effective.

9. The method of claim 1, wherein, The heat transfer coefficient of the building outer wall is calculated based on the corrected temperature of the outer surface of the outer wall, the inner surface temperature of the outer wall, the indoor air temperature and the outdoor air temperature by using a machine learning algorithm, including: The corrected temperature of the outer surface of the outer wall, the inner surface temperature of the outer wall, the indoor air temperature and the outdoor air temperature are standardized; The corrected temperature of the outer surface of the outer wall, the inner surface temperature of the outer wall, the indoor air temperature and the outdoor air temperature are standardized; 10. A field detection system for the heat transfer coefficient of an external wall based on infrared thermography technology, characterized in that it comprises: The corrected temperature of the outer surface of the outer wall, the inner surface temperature of the outer wall, the indoor air temperature and the outdoor air temperature are standardized; The system comprises: The outer wall surface initial temperature detection module is used to detect the temperature of the outer wall surface of the building by using an infrared thermal imager to obtain the initial temperature of the outer wall surface; the initial temperature of the outer wall surface includes the initial temperature of the outer surface of the outer wall; The temperature and environmental parameter data acquisition module is used to synchronously acquire the temperature data and environmental parameter data of the outer wall surface for a preset time length; The temperature and environmental parameter dynamic relationship analysis module is used to analyze the dynamic relationship between the change of the environmental parameter data and the change of the temperature of the outer wall surface based on the temperature data and environmental parameter data of the outer wall surface for the preset time length by using a frequency domain analysis method; The initial temperature correction module is used to calculate a dynamic correction factor for correcting the initial temperature of the outer wall surface based on the dynamic relationship between the change of the environmental parameter data and the change of the temperature of the outer wall surface, and to correct the initial temperature of the outer wall surface by using the dynamic correction factor to obtain the corrected temperature of the outer wall surface; the corrected temperature of the outer wall surface includes the corrected temperature of the outer surface of the outer wall; The heat transfer coefficient calculation module is used to acquire the inner surface temperature of the outer wall, the indoor air temperature and the outdoor air temperature, and to calculate the heat transfer coefficient of the building outer wall based on the corrected temperature of the outer surface of the outer wall, the inner surface temperature of the outer wall, the indoor air temperature and the outdoor air temperature by using a machine learning algorithm.

Images