Building carbon emission reduction intelligent accounting method based on intelligent algorithm

Abstract

The invention provides a building carbon emission reduction intelligent accounting method based on an intelligent algorithm, and the method comprises the steps: constructing a multi-mode sensor detection network, and collecting the temperature, heat flow, humidity, infrared thermal image and meteorological boundary condition data of a building envelope; forming a time-space consistent multi-modal clean data set through unified time and space calibration; the method comprises the following steps: identifying a heat bridge suspected region and a moisture-containing salient region by utilizing the characteristics of a temperature difference field, a heat flow gradient field and a dew point temperature difference, generating a heat-moisture coupling sensitive partition mask, establishing a heat-moisture coupling heat transfer model under partition constraint, and identifying a reference heat transfer coefficient, a heat bridge influence parameter and a moisture-containing parameter by adopting a Bayesian inversion method; according to the method, high-precision identification of the influence factors of the thermal performance of the building and dynamic, traceable and quantitative accounting of the carbon emission are realized, the accuracy and auditing credibility of carbon emission reduction evaluation are improved, and the method is suitable for scenes of green building evaluation, energy-saving reconstruction evaluation, carbon asset accounting and the like.

Description

Technical Field

[0001] This invention relates to the field of intelligent algorithm technology, and in particular to an intelligent calculation method and system for building carbon emission reduction based on intelligent algorithms. Background Technology

[0002] With the comprehensive advancement of the dual-carbon strategy, energy consumption and carbon emissions in the building sector have become important components in achieving carbon peaking and carbon neutrality. Building energy consumption accounts for a high proportion of total social energy consumption, and the thermal performance of the building envelope (such as exterior walls, roofs, doors, and windows) directly determines the building's energy consumption level and is a key factor in building energy conservation and carbon emission control.

[0003] Current methods for testing the thermal performance of buildings mostly employ wired installations and single-point static measurements, which suffer from problems such as cumbersome wiring, significant construction interference, long testing cycles, and the inability to conduct continuous monitoring. These methods fail to reflect the dynamic thermal behavior of the building envelope under actual operating conditions. Furthermore, the identification and quantification of coupling factors such as thermal bridging and humidity levels still rely on experience or indirect estimation, lacking high-precision and verifiable technical means.

[0004] Therefore, we propose an intelligent accounting method and system for building carbon emission reduction based on intelligent algorithms.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to address the shortcomings of existing technologies by providing an intelligent calculation method for building carbon emission reduction based on intelligent algorithms, thereby solving the technical problems mentioned in the background section.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] A smart calculation method for building carbon emission reduction based on intelligent algorithms includes the following steps:

[0009] S1. Establish paired measuring points consisting of indoor surface temperature, outdoor surface temperature, heat flux density, and surface relative humidity; deploy paired sensors and infrared imaging devices in the detection area, record meteorological conditions and operating load as boundary conditions, and complete the consistency calibration of time reference and coordinate reference.

[0010] S2. Obtain the time series of temperature, heat flow, humidity, weather and infrared thermal image corresponding to the detection area, perform time alignment, spatial registration and outlier removal to form a multimodal clean dataset for parameter identification and area determination;

[0011] S3. The infrared thermal image surface temperature difference texture is fused and analyzed with the heat flow and temperature and humidity changes of paired measurement points to locate suspected thermal bridge areas and quantify the thermal bridge characteristic intensity; based on the humidity time series and surface temperature time series, the humidity state index is extracted, and the thermal bridge partition mask and the humidity partition mask are output.

[0012] S4. Using the multimodal cleaning dataset, the thermal bridge partition mask, and the moisture-containing partition mask as constraints, perform thermal and moisture coupling parameter identification to obtain a set of thermal parameters for the building envelope, and provide an uncertainty assessment of the set of thermal parameters.

[0013] S5. Convert the effects of thermal bridging and humidity into equivalent heat transfer coefficients, and generate the time series and spatial distribution of the equivalent heat transfer coefficients; perform systematic deviation correction on the thermal resistance, heat transfer coefficient and energy consumption estimates obtained based on steady-state assumptions, and form the corrected thermal performance results and their confidence intervals.

[0014] S6. Using the corrected thermal performance results and boundary conditions as input, calculate the carbon emissions of the baseline operation and the carbon emissions after the implementation of energy-saving measures to obtain the time series and statistical results of carbon emission reduction; generate the accounting confidence interval, abnormal period identifier and traceability report based on the uncertainty assessment as the accounting output of building carbon emission reduction.

[0015] S1 includes: determining the geometric boundaries and measurement point layout range of the building envelope detection area, dividing the sensor layout grid according to the structural characteristics of the building's exterior walls, doors, windows, or roof, and ensuring that each measurement point corresponds to a clear spatial coordinate;

[0016] Temperature sensors, heat flow sensors, and humidity sensors are arranged in pairs on the inner and outer surfaces of the detection area, and an infrared imaging device is set up to ensure the consistency between infrared imaging and the spatial distribution of measurement points.

[0017] Meteorological boundary conditions corresponding to the collection and detection area, including outdoor temperature, relative humidity, wind speed, and solar radiation intensity information, are used as external drivers for subsequent calculations.

[0018] Establish a unified time and spatial coordinate reference, and synchronously calibrate the sampling frequency, sampling time and coordinate position of each measuring point and infrared device to ensure the spatiotemporal consistency of subsequent data fusion;

[0019] The output includes an observation configuration file containing sensor configuration parameters, sampling frequency, and spatial point coordinates, which can be used for subsequent multimodal acquisition.

[0020] S2 includes: starting the sensor network and infrared imaging device according to the observation configuration file, and collecting raw time-series data of indoor surface temperature, outdoor surface temperature, heat flux density, surface relative humidity, infrared thermal image and meteorological boundary conditions;

[0021] Time alignment and spatial registration are performed on the collected data to ensure that each physical quantity has a one-to-one corresponding sampling position in space at each time step;

[0022] The system performs outlier removal and signal correction on the acquired data, including removing acquisition jitter, drift, and missing measurement segment interpolation markers, and generating a continuous and usable time series of physical quantities.

[0023] The processed temperature, heat flux, humidity, infrared thermal image, and meteorological boundary condition data are fused into a multimodal clean dataset, preserving the correspondence between timestamps, spatial coordinates, and corresponding physical attributes.

[0024] Output a multimodal cleaning dataset for subsequent thermal bridge and moisture content identification.

[0025] S3 includes: based on a multimodal cleaning dataset, extracting surface temperature difference features and heat flow change features at each measuring point, and constructing a temperature difference field and heat flow field corresponding to the infrared thermal image;

[0026] By utilizing the temperature difference texture features in infrared thermal images, matching the heat flux density changes at the measurement points, identifying areas of abnormally concentrated temperature gradients, locating suspected thermal bridge areas, and extracting thermal bridge morphological features;

[0027] Based on the time series of indoor and outdoor surface temperature and humidity, the humidity state index is calculated, the area with significant local humidity is identified, and the humidity zoning results are generated.

[0028] By combining the spatial overlap between suspected thermal bridge regions and significantly humid regions, thermal bridge partition masks and humid region masks are obtained, which are used for regional constraints in subsequent parameter identification.

[0029] The output includes the recognition results of the thermal bridge partition mask, the wet partition mask, and local feature parameters, ensuring complete consistency with the coordinate baseline of the dataset.

[0030] S4 includes: constructing a thermal and moisture coupled heat transfer model under the regional constraints of thermal bridge partition mask and humidified partition mask. The model parameters include the baseline heat transfer coefficient, thermal bridge influence parameters and humidified state parameters.

[0031] The model parameters are retrieved using multi-time period heat flow, temperature and humidity data to form a set of thermal parameters with consistent temporal and spatial distribution.

[0032] Quantitative values ​​and confidence intervals of thermal bridge influence parameters and moisture content parameters are obtained through least squares fitting, Bayesian estimation, or other identification algorithms.

[0033] The identified set of thermal parameters is associated with the coordinate datum to ensure that the parameter value of each measuring point corresponds to its spatial location.

[0034] The output includes a set of thermal parameters including the reference heat transfer coefficient, thermal bridge effect parameters, humid state parameters, and uncertainties.

[0035] S5 includes: based on the set of thermal parameters, converting the thermal bridge influence parameters and the humid state parameters into corresponding heat transfer coefficient increments, forming thermal bridge increment terms and humidity increment terms;

[0036] The thermal bridge increment term, the moisture content increment term, and the reference heat transfer coefficient are weighted and superimposed to calculate the spatiotemporal distribution of the equivalent heat transfer coefficient at different measurement points and time steps.

[0037] Based on the equivalent heat transfer coefficient, a systematic deviation correction is made to the thermal resistance and energy consumption estimation results under the traditional steady-state assumption;

[0038] Confidence interval analysis is performed on the corrected thermal performance results to form an assessment of the consistency between the equivalent heat transfer coefficient distribution and the corrected results.

[0039] The output corrected thermal performance results and corresponding confidence intervals provide a realistic input for carbon emission reduction calculation.

[0040] S6 includes: calculating baseline operating carbon emissions using corrected thermal performance results and meteorological boundary conditions as inputs;

[0041] Calculate the carbon emissions after energy saving based on the thermal performance status after implementing energy-saving measures.

[0042] A time series of carbon emission reductions is generated by comparing the carbon emissions during baseline operation with those after energy conservation.

[0043] Based on the uncertainty of the thermal and wet coupling parameters, calculate the confidence interval of carbon emission reduction and identify abnormal time periods and abnormal measurement points;

[0044] Output carbon emission reduction calculation results, confidence intervals, and traceable reports to achieve accurate, stable, and verifiable calculation of building carbon emission reductions.

[0045] The beneficial effects of this invention are as follows:

[0046] This invention employs a multimodal wireless sensor array to collect temperature, heat flow, humidity, infrared thermal imaging, and environmental meteorological information in a unified manner. Through self-organizing network communication, it enables rapid deployment and flexible allocation, significantly reducing construction interference. The system can continuously collect data at a minute-level time resolution, achieving dynamic monitoring of thermal resistance fluctuations, thermal bridge formation, and humidity migration. This significantly improves the timeliness and spatial coverage of the detection, making the data representative for engineering applications. This high-frequency dynamic monitoring method can obtain richer measured information without damaging the building structure, providing a solid data foundation for subsequent intelligent accounting.

[0047] This invention utilizes multimodal features such as temperature difference field gradient, heat flux density anomaly, and dew point temperature difference threshold, combined with spatial overlap analysis and adaptive threshold algorithm, to accurately locate potential thermal bridges and moisture-containing anomaly areas in the building envelope. By constructing a coupling-sensitive partition mask, the complex physical field problem is simplified into a feature partitioning problem, providing a high-confidence spatial constraint for subsequent thermal-humidity coupling parameter identification, significantly reducing model convergence time and parameter uncertainty, overcoming the low-resolution estimation methods based on macroscopic mean or empirical parameters in traditional methods, and making building carbon emission reduction calculations closer to the real physical state.

[0048] This invention introduces Bayesian inversion and MCMC sampling algorithms into carbon emission accounting for the first time. Through joint inference using prior knowledge and measured data, the system can automatically identify key physical quantities such as heat transfer coefficient, thermal bridge-related heat transfer terms, and moisture permeability parameters. The algorithm sets up residual convergence criteria and confidence interval evaluation mechanisms, which can significantly suppress computational drift caused by data noise and parameter coupling, making the identification results statistically robust. Compared with traditional single-model fitting, it can maintain high accuracy and high robustness under complex operating conditions.

[0049] This invention corrects for deviations in energy consumption calculations by converting thermal bridges and humidity effects into equivalent heat transfer coefficients, making them highly close to actual energy consumption. Furthermore, it uses an error propagation model to perform stratified calculation and quantification of uncertainties, making previously difficult-to-track deviations explicit. This not only enhances the scientific rigor and comparability of the calculation results but also provides quantitative evidence for carbon asset auditing, avoiding the problem of "unaccountability for estimation deviations" in the current accounting system.

[0050] This invention introduces a point-by-point calculation and spatial aggregation mechanism into the carbon emission calculation process, enabling precise quantification of carbon emission differences between baseline and retrofit scenarios at different locations within the building envelope. By coupling carbon emission factors with energy consumption calculation results, a carbon emission reduction accounting link with higher spatiotemporal resolution and the ability to trace back to specific locations is formed. Through mechanisms such as 2σ threshold anomaly identification, uncertainty labeling, and automatic generation of accounting audit reports, the results possess high traceability and verifiability. Even in complex environments or in the event of data loss, the system maintains stability and reliability. Attached Figure Description

[0051] Figure 1 This is a schematic diagram of an intelligent accounting method for building carbon emission reduction based on intelligent algorithms according to the present invention;

[0052] Figure 2 This is a schematic diagram of the framework of an intelligent accounting system for building carbon emission reduction based on intelligent algorithms according to the present invention. Detailed Implementation

[0053] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0054] Example 1: As Figure 1 As shown, this embodiment provides an intelligent accounting method for building carbon emission reduction based on intelligent algorithms, including the following steps:

[0055] S1. Sensor and observation configuration: Determine the detection area of ​​the target enclosure structure, and establish paired measurement points consisting of indoor surface temperature, outdoor surface temperature, heat flux density, and surface relative humidity; deploy paired sensors and set up infrared imaging devices in the detection area, record meteorological conditions and operating load as boundary conditions, and complete the consistency calibration of time reference and coordinate reference.

[0056] S2. Multimodal synchronous acquisition and preprocessing: Acquire the synchronous time series of temperature, heat flow, humidity, meteorology and infrared thermography corresponding to the detection area, perform time alignment, spatial registration and outlier removal to form a multimodal clean dataset for parameter identification and area determination.

[0057] S3. Thermal Bridge and Moisture Content Feature Identification: Based on the multimodal cleaning dataset, the surface temperature difference texture of the infrared thermal image and the heat flow-temperature and humidity changes of paired measurement points are fused and analyzed to locate suspected thermal bridge areas and quantify the thermal bridge feature intensity to obtain thermal bridge partitioning results; at the same time, moisture content indicators are extracted based on humidity time series and surface temperature time series to obtain moisture content partitioning results; thermal bridge partitioning masks and moisture content partitioning masks are output for subsequent parameter identification.

[0058] S4. Thermal-humidity coupling parameter identification: Using the multimodal clean dataset, the thermal bridge partition mask, and the humid partition mask as constraints, thermal-humidity coupling parameter identification is performed to obtain a set of thermal parameters for the building envelope. The set of thermal parameters includes at least the spatiotemporal distribution of the reference heat transfer coefficient, thermal bridge influence parameters, and humidity state index, and an uncertainty assessment is given for the set of thermal parameters.

[0059] S5. Calculation and deviation correction of equivalent heat transfer coefficient: Based on the set of thermal parameters, the effects of thermal bridge and moisture content are converted into equivalent heat transfer coefficients, and the time series and spatial distribution of equivalent heat transfer coefficients are generated; on the basis of the equivalent heat transfer coefficients, the thermal resistance, heat transfer coefficient and energy consumption estimates obtained based on the steady-state assumption are systematically corrected to form the corrected thermal performance results and their confidence intervals.

[0060] S6. Intelligent calculation and output of carbon emission reduction: Using the corrected thermal performance results and the boundary conditions as input, calculate the baseline operating carbon emissions and the carbon emissions after the implementation of energy-saving measures to obtain the carbon emission reduction time series and statistical results; and generate the calculation confidence interval, abnormal period identifier and traceability report based on the uncertainty assessment as the calculation output of building carbon emission reduction.

[0061] S1 specifically includes the following sub-steps:

[0062] S110. Determining the inspection area and dividing the grid: Determine the inspection area of ​​the target enclosure structure. The inspection area is applicable to enclosure components with clear inner and outer surface boundaries, such as concrete exterior walls, brick exterior walls, glass curtain walls, and roof structures.

[0063] The selection of the detection area should ensure that a stable temperature difference (preferably ≥5℃) is formed within the detection cycle, and exclude areas with strong heat source interference or strong local airflow disturbance.

[0064] Within the detection area, a coordinate grid is established based on structural geometric information, and sensors are evenly or equivalently distributed with a spacing of 0.3m to 1.0m to ensure spatial coverage and distinguishability of heat flow characteristics within the detection area.

[0065] S120. Sensor and infrared thermal imaging device deployment: At each measuring point location of the above grid, a temperature sensor, a heat flow sensor, and a humidity sensor are installed in pairs on the inner and outer surfaces, respectively.

[0066] The sensor is installed by fixing it tightly to the surface of the enclosure structure to ensure good thermal contact and avoid measurement errors caused by contact thermal resistance.

[0067] The field of view of the infrared thermal imaging device completely overlaps with the sensor placement area, and high reflectivity calibration target points are placed on the outer edge of the detection area to form a spatial calibration reference that can be used to correspond one-to-one between image coordinates and actual physical coordinates.

[0068] The performance parameters of various sensors are as follows: Temperature sensor: range -20℃~80℃, accuracy ±0.1℃; Heat flow sensor: range 0~200W / m², accuracy ±5%; Humidity sensor: range 0%~100%RH, accuracy ±2%.

[0069] S130. External meteorological boundary condition acquisition: Simultaneously, during the detection period, meteorological boundary conditions corresponding to the detection area are acquired, including outdoor temperature, relative humidity, wind speed and solar radiation intensity. The acquisition frequency is synchronized with the sensor sampling frequency.

[0070] The shared timestamps of external meteorological boundary conditions and monitoring data provide environmental-driven inputs for thermal parameter calculations and carbon emission accounting.

[0071] S140. Time reference and spatial calibration: Establish a unified time reference for all sensors and infrared thermal imaging equipment, and set the sampling frequency to 1 to 10 minutes, which can be adjusted according to the field conditions.

[0072] The time synchronization tolerance is limited to no more than 1 minute. When any sensor timestamp offset exceeds the tolerance range, linear interpolation and time difference compensation algorithms are automatically used for correction, and the synchronization deviation log is recorded for subsequent traceability. Spatial calibration adopts the "target point geometric projection calibration" method: based on the target position calibration of three or more points, a one-to-one correspondence mapping matrix between infrared image pixel coordinates and sensor physical coordinates is established to achieve spatiotemporal consistency of multi-source data.

[0073] S150, Observation Configuration File Output: After completing the point placement, calibration, and time synchronization, an observation configuration file is generated. The configuration file must contain at least the following content:

[0074] Sensor placement coordinate matrix Sampling frequency And the total detection time T; the infrared thermal image and sensor coordinate mapping matrix; meteorological boundary condition acquisition rules and interface paths; time synchronization and deviation correction parameters.

[0075] This observation configuration file serves as the basic input for subsequent multimodal data synchronous acquisition and thermal-humidity coupling parameter identification, ensuring a closed-loop correspondence between the detection configuration and subsequent calculation processes.

[0076] S2 specifically includes the following sub-steps:

[0077] S210. Synchronous acquisition of multi-source data: Start the multi-source sensor network and infrared thermal imaging device according to the point coordinates and sampling frequency of the observation configuration file.

[0078] The data collected included indoor surface temperature. Outdoor surface temperature Heat flux density q, relative humidity RH, infrared thermal image features IR_feature, and meteorological boundary conditions meteo.

[0079] The sampling frequency is set to 1 to 10 minutes, and the collection duration covers at least 24 hours to ensure the integrity of the diurnal thermal response characteristics; all acquisition devices are started synchronously under a unified time reference and timestamps are recorded to ensure timing consistency.

[0080] S220, Time Alignment and Spatial Registration: Perform time alignment and spatial registration on all acquired data. The time synchronization tolerance is no more than 1 minute. When the timestamp offset of any measurement point is less than or equal to 3 sampling periods, linear interpolation is used for correction; if the offset exceeds 3 sampling periods, it is automatically marked as an "invalid time period" and will not participate in subsequent modeling and calculation.

[0081] Spatial registration is based on pre-deployed calibration target points, establishing a one-to-one mapping relationship between infrared image pixel coordinates and sensor physical coordinates through geometric projection transformation. The mapping matrix has been calibrated before acquisition to ensure complete spatial correspondence of the data.

[0082] S230, Anomaly Detection and Removal: Perform anomaly detection and removal on various types of acquired signals; for temperature, heat flow, and humidity signals: use ±3σ threshold to remove anomalies that exceed the normal fluctuation range, and also remove values ​​that exceed the physical range of the sensor.

[0083] For infrared thermal imaging signals: Based on the local neighborhood mean μ and standard deviation σ, abnormal pixels exceeding μ+3σ are marked and removed, and adjacent pixels are interpolated and corrected; the missing data after removal are not used for thermal-humidity coupling parameter identification, and only the markings are retained for data quality tracking.

[0084] S240. Data resampling and fusion: When the sampling frequencies of different types of sensors are inconsistent, all signals are resampled to a unified time axis by linear interpolation based on the minimum sampling time step to ensure time series alignment.

[0085] The fused data contains synchronized values ​​of temperature, heat flux, humidity, infrared thermal imaging features, and meteorological boundary conditions at each time step, and forms a one-to-one correspondence with the sensor spatial coordinates, thus constituting a spatiotemporally consistent multimodal clean dataset.

[0086] S250, Clean Dataset Output: The data after time alignment, spatial registration, outlier removal and resampling is output as a multimodal clean dataset.

[0087] The minimum set of fields in a dataset is defined as:

[0088]

[0089] Where time is a unified timestamp. The output consists of sensor physical coordinates and other fields representing the corresponding physical quantities. The output format can be CSV, JSON, or binary structured format, and includes synchronization accuracy and anomaly marking information. This clean dataset provides standardized and synchronized input for thermal bridge and moisture feature identification (S310–S350) and thermal-moisture coupling parameter identification (S410–S450).

[0090] S3 specifically includes the following sub-steps:

[0091] S310. Multimodal feature extraction: Based on the multimodal cleaning dataset, extract feature information related to thermal bridges and humid conditions;

[0092] Temperature difference characteristics through indoor surface temperature With outdoor surface temperature Calculation of the difference:

[0093]

[0094] in To the location of sensor deployment The difference between indoor and outdoor surface temperatures; Temperature of the inner surface of the building envelope; Temperature of the outer surface of the building envelope; This represents the planar coordinate position of the temperature sensor, used to correspond with the spatial positioning of thermal bridges and humid areas.

[0095] Heat flux characteristics are obtained by calculating heat flux density. Spatial gradient:

[0096]

[0097] in and These represent the partial derivatives of the heat flux density along the x and y directions, respectively, reflecting the local rate of change of the heat flux distribution in space; The spatial gradient of heat flux density is used to identify thermal bridges and areas of abnormal humidity; the infrared thermal image temperature difference field is obtained through the infrared image pixel temperature matrix. The spatial resolution is consistent with the sensor spacing (0.3m to 1.0m) obtained by differential analysis with the outdoor temperature field. The extracted temperature difference field, heat flow gradient field and infrared temperature difference field are synchronized with the multimodal clean dataset in time and space, providing a basis for subsequent thermal bridge and humidity identification.

[0098] S320. Identification of suspected thermal bridge areas: Based on the temperature difference field and heat flow gradient field, temperature difference threshold and gradient concentration criteria are used to identify suspected thermal bridge areas.

[0099] The temperature difference threshold is set as follows: ; The temperature difference threshold is defined as follows: The region of abnormal heat flow concentration is defined as:

[0100]

[0101] in This is the average value of the heat flux gradient. The standard deviation of the heat flux gradient is used; when a local heat flux gradient exceeds twice the standard deviation of the mean, the heat flux at that location is considered abnormally concentrated; spatial connectivity requirement: areas where more than three consecutive adjacent measuring points meet the above conditions are marked as suspected thermal bridge areas, and isolated anomalies are automatically ignored; the identified suspected thermal bridge areas are represented by a binary matrix. This indicates that the value is 0 or 1, and the spatial resolution is consistent with the sensor placement.

[0102] S330. Identification of Significantly Humid Areas: Calculate the dew point temperature from the temperature and humidity information in the multimodal cleaning dataset and determine the significantly humid areas.

[0103] Dew point temperature is calculated using the Magnus formula:

[0104]

[0105]

[0106] in is the Magnus function used to calculate the air dew point temperature as an intermediate variable; T is the temperature (unit: °C); RH is the relative humidity (dimensionless, ranging from 0 to 1); a and b are the empirical coefficients of the Magnus formula, where a = 17.27 and b = 237.7 °C. The dew point temperature is (°C).

[0107] Dew point temperature difference is defined as:

[0108]

[0109] in For dew point temperature difference, The measured temperature (°C) of the building's interior surface. The dew point temperature (°C) is used; when ΔTd ≤ 1°C and RH ≥ 80%, the measuring point is marked as a significantly humid region; the output humidity zone is presented as a binary matrix. The value is 0 or 1, where 1 represents a significantly humid area and 0 represents a non-humid area.

[0110] S340, Thermal-Moisture Overlay Partition Mask Generation: Masking suspected thermal bridge areas Masking of areas with significant moisture Perform spatial overlay and calculate the degree of overlap:

[0111]

[0112] Where Overlap represents the spatial overlap between the thermal bridge region and the humid region, and the value ranges from 0 to 1; Area of ​​suspected thermal bridge region (m²); Area of ​​the significantly humid region (m²); This represents the area of ​​the intersection between the thermal bridge region and the humid region. This represents the combined area of ​​the thermal bridge region and the humid region.

[0113] when At that time, the region is determined to be a region with significant spatial coupling of thermal and humidity features, and is marked as a thermal-humidity coupling sensitive region. The final generated region mask matrix is ​​denoted as... Its value is 0 or 1, where 1 represents the heat-humidity coupling sensitive area and 0 represents the non-coupling area.

[0114] S350, Feature Output: Outputs thermal bridge partition mask, humidified partition mask and thermal-humid coupling sensitive area mask, and simultaneously outputs local feature parameter matrix.

[0115] The output field is defined as follows:

[0116]

[0117] Where time represents the timestamp of the corresponding detection data collection; Due to the temperature difference between indoors and outdoors, The heat flux gradient is expressed in W / m³. Here, RH represents the dew point temperature difference (in °C) and relative humidity. , , These represent binary mask matrices for the thermal bridge region, the humid region, and the thermal-humidity coupling sensitive region, respectively.

[0118] The mask matrix is ​​stored in 0 / 1 binary matrix form or floating-point weight matrix, with the resolution consistent with the sensor layout. The output format supports CSV, JSON or binary structured format. The output result is the direct input of thermal-humid coupling parameter identification (S410–S450), ensuring the consistency of terminology and logical closure of the steps before and after.

[0119] S4 specifically includes the following sub-steps:

[0120] S410. Establishment of the thermal-humidity coupling model: Under the spatial constraints of the thermal bridge partition mask and the humidified partition mask, a thermal-humidity coupling heat transfer model is established for each detection area:

[0121]

[0122] in Equivalent heat transfer coefficient, in units of W / (m²·K); : Reference heat transfer coefficient, unit W / (m²·K); : Thermal bridge influence parameter, dimensionless; : Moisture content parameter, dimensionless; , Weighting factor, unit W / (m²·K), initial value range is 0.1~2.0; Sensor placement coordinates;

[0123] Mask matrix and Used to limit the parameter identification region, inversion calculations are performed only in the region where mask=1.

[0124] S420, Parameter Identification Algorithm: Bayesian inversion method is used for... Perform parameter estimation.

[0125] Target posterior distribution:

[0126]

[0127] Where D represents the measured temperature, heat flow, and humidity data; It is the likelihood function; The prior distribution is given; the likelihood function is set to Gaussian form:

[0128]

[0129] in The measured heat flux over time t. Based on The model heat flow, To measure the noise standard deviation; the prior distribution is taken as a non-information-based normal distribution: N is the number of time series sampling points; parameter inversion is achieved through MCMC (Markov Chain Monte Carlo) sampling, and the sampling step size can be adaptively adjusted according to the residual.

[0130] S430. Convergence and computational stability control: Residual definition:

[0131]

[0132] Convergence threshold: residual ≤ 0.05 W / m²; Maximum number of iterations: 100 steps; Early stopping strategy: terminate early when the residual decreases by less than 1% for 5 consecutive iterations; Stability strategy: automatically reset parameters if the residual increases for 3 consecutive iterations. Return to the initial state and resample.

[0133] S440, Spatial Registration and Confidence Interval Calculation:

[0134] After completing the parameter inversion, the estimated Spatial coordinates Alignment, resolution consistent with sensor placement.

[0135] Calculate the 95% confidence interval based on the posterior distribution variance:

[0136]

[0137] in , Let the standard deviation of the parameter be . It is a parameter The estimated value (e.g., posterior mean or median). It is the 95th percentile coefficient of the normal distribution. The confidence interval is calculated point by point. The output CI matrix is ​​consistent with the spatial resolution of the sensor, providing statistical boundary conditions for subsequent uncertainty propagation analysis and carbon emission reduction accounting.

[0138] S450, Thermal Parameter Set Output: Outputs a set of thermal parameters, including:

[0139]

[0140] in Spatial coordinates; These are estimated values ​​for the heat transfer coefficient, thermal bridge effect parameters, and humid state parameters, respectively. , , These are the upper and lower limits of the 95% confidence interval for the corresponding parameters, calculated based on S440.

[0141] The output format supports two forms:

[0142] Matrix form (spatial distribution): Each parameter and confidence interval corresponds one-to-one with the sensor grid, which is suitable for spatial visualization and regional thermal characteristic analysis;

[0143] Vectorized form (row list structure): Each row contains a coordinate point. Its parameters and confidence intervals are applicable to subsequent bias correction and numerical optimization; the output dataset can be directly used as input for S510–S550 bias correction and carbon emission reduction calculation, realizing seamless connection between the front and back processes and ensuring the traceability and consistency of thermal characteristic analysis and carbon calculation results.

[0144] S5 specifically includes the following sub-steps:

[0145] S510, Thermal Bridge and Moisture Increment Calculation: Thermal Bridge Influence Parameters Obtained from S410–S450 and moisture content parameters The incremental term for calculating the equivalent heat transfer coefficient.

[0146] Thermal bridge increment:

[0147]

[0148] Moisture content increase:

[0149]

[0150] in , , where W is the weighting factor and the unit is W / (m²·K); and The units are all W / (m²·K), and the coordinates of each sensor point are calculated. Execute point by point; when Exceeding the reference heat transfer coefficient If the value is within ±30%, the point is marked as an outlier and will not be included in subsequent correction calculations.

[0151] S520, Equivalent heat transfer coefficient calculation: Calculate the equivalent heat transfer coefficient for each spatial measuring point:

[0152]

[0153] in The baseline heat transfer coefficient is used; calculations are performed point-by-point at each sensor location, and the output is... To maintain consistency with the sensor's spatial resolution, a two-dimensional spatial matrix field with equivalent heat transfer coefficients is formed, providing an input basis for energy consumption and carbon emission accounting;

[0154] For the locations identified as outliers in S510 Its corresponding It will not participate in the subsequent carbon emission calculations for S530–S550, but will be retained in the results set for quality traceability and closed-loop management of monitoring data.

[0155] S530, Thermal Resistance and Energy Consumption Calculation: After obtaining the equivalent heat transfer coefficient Then, calculate the corresponding thermal resistance. :

[0156]

[0157] And calculate the energy consumption per unit area. ; Continuous-time integral form:

[0158]

[0159] Where τ is the time variable (s). This is a temperature difference-weighted average value; calculated based on temperature data collected from S210–S250.

[0160] Discrete sampling uses a summation method:

[0161]

[0162] in Let be the indoor and outdoor temperature difference at the nth sampling time; t represents the sampling interval; t represents the cumulative collection time in hours.

[0163] S540. Calculation of Energy Consumption Uncertainty: Based on error propagation theory, calculate energy consumption per unit area. uncertainty .

[0164] when and When they are independent, then:

[0165]

[0166] in Confidence intervals derived from S410–S450 parameter identification. This is due to temperature measurement error.

[0167] like and If a correlation exists, add a covariance term:

[0168]

[0169] in Equivalent heat transfer coefficient The covariance with temperature difference ΔT is used to account for the case where there is a statistical correlation between the two.

[0170] The partial derivative can be expressed as:

[0171]

[0172]

[0173] Uncertainty calculations are performed point-by-point at each sensor location, and the output is... Spatial coordinates A one-to-one correspondence is formed to create a spatial distribution matrix of energy consumption uncertainty, providing basic data for the uncertainty analysis of S550 carbon emissions.

[0174] S550, Corrected Thermal Performance Output: Outputs corrected thermal performance results, including equivalent heat transfer coefficient, thermal resistance, energy consumption, and uncertainty information.

[0175] The output field set is defined as follows:

[0176]

[0177] Where x and y represent spatial coordinates; R is the equivalent heat transfer coefficient, in W / (m²·K); R is the thermal resistance, in (m²·K) / W; Q is the energy consumption per unit area, in J / m². It is the uncertainty of the equivalent heat transfer coefficient; It represents the uncertainty in energy consumption calculation; flag is a data status indicator, with 0 indicating normal data points and 1 indicating abnormal data points.

[0178] Anomalies are defined as follows: Measurement points exceeding the ±30% threshold of the baseline heat transfer coefficient are not included in the S610–S650 carbon emission calculation, but their information is retained in the output data for quality traceability and statistical analysis. The output format supports matrix (spatial distribution) and vectorized structure (table format), which can be directly used as input to the S610–S650 carbon emission calculation module to realize the calculation of carbon emission reduction after thermal-humidity deviation correction.

[0179] This step establishes a standardized interface between thermal parameter correction results, energy consumption information, and carbon emission accounting, enabling seamless integration from thermal and moisture deviation correction to carbon emission reduction accounting.

[0180] S6 specifically includes the following sub-steps:

[0181] S610, Baseline Operating Carbon Emissions Calculated Point-by-Point: Based on Corrected Energy Consumption per Unit Area Output from S550 Based on this, and using carbon emission factors (unit (Based on authoritative national or regional guidelines, such as IPCC guidelines or national standards), carbon emissions are calculated point-by-point for baseline operation:

[0182]

[0183] in The point-by-point cumulative energy consumption (kWh) is calculated according to S530 under the baseline scenario. Clearly indicate the data source and validity period; if a regional factor is used, the version number or publication date should be recorded.

[0184] Point-by-point calculation is only performed on normal measurement points with flag=0; for abnormal measurement points with flag=1, It is denoted as Null and is not included in the total calculation, but is retained in the output data for anomaly diagnosis and subsequent statistical analysis.

[0185] The output results form a spatial distribution matrix of carbon emissions. It can correspond one-to-one with the sensor coordinate space, providing baseline emission data for subsequent energy conservation and emission reduction accounting.

[0186] S620. Carbon emissions after retrofitting are calculated point-by-point: based on the energy consumption per unit area obtained after implementing energy-saving measures. Based on this, calculate the carbon emissions after the renovation point by point:

[0187]

[0188] in: Based on S520 Operating conditions after modification (actual or predicted) Calculated under the sequence;

[0189] If the energy structure changes after the modification (such as fuel replacement or power source switching), the corresponding scenario should be invoked. The scenario ID is recorded for subsequent emission reduction analysis and carbon accounting audit; point-by-point calculations are performed only for normal monitoring points with flag=0; for abnormal monitoring points with flag=1, Recorded as Null, it is not included in the carbon emission aggregation; the output results are generated. The spatial distribution matrix provides the input basis for calculating the energy saving and emission reduction of S630.

[0190] S630, Carbon Emission Reduction Point-by-Point and Summary Calculation: Based on the carbon emission results output by S610 and S620, the carbon emission difference before and after the energy-saving renovation is calculated point-by-point, and weighted summaries are performed at the regional scale.

[0191] Point-by-point carbon emission reduction is defined as:

[0192]

[0193] The total carbon emission reduction is:

[0194]

[0195] If the areas corresponding to the measuring points are inconsistent, a weighted summation can be used:

[0196]

[0197] The summary calculation only includes normal measurement points with flag=0; for abnormal points with flag=1, It is recorded as Null and excluded from the total calculation, but listed separately in the report with an explanation of the reason.

[0198] The output includes: The spatial distribution matrix; Total regional carbon emission reduction; list and explanation of anomalies; the summary results can be directly used for spatial visualization on GIS / BIM platforms, and can also be connected to the carbon accounting audit module as the final basis for evaluating the effectiveness of energy-saving renovation and emission reduction.

[0199] S640. Calculation of Uncertainty in Carbon Emission Reduction and Identification of Abnormal Periods / Points: The uncertainty propagation method is used to estimate the uncertainty of point-by-point and aggregated carbon emission reductions.

[0200] Pointwise uncertainty:

[0201]

[0202]

[0203]

[0204] , It is the variance of carbon emissions; It is the variance of carbon emission reduction at a single measurement point; if and Since there is a correlation, a covariance term should be added:

[0205]

[0206] in This refers to the covariance of carbon emissions before and after the retrofit; the total uncertainty of regional carbon reduction is calculated using the independent error approximation or the Monte Carlo method.

[0207]

[0208] Anomaly detection rules:

[0209] Point-by-point anomaly: when Or, if the measuring point was previously marked as an anomaly when flag=1 was in S550; Time-period anomaly: when the total carbon emission reduction within a certain time window... When the value exceeds the mean ±2σ, it is identified as an abnormal period. The reasons for both abnormal points and abnormal periods are recorded (such as sensor failure, extreme weather events, changes in operating conditions), and an audit log is generated for manual review.

[0210] S650, Carbon Emission Reduction Calculation Output and Traceability Report: Outputs calculation results at two levels: point-by-point and summary, and generates a traceability report. Output fields are:

[0211]

[0212] Also output summary items:

[0213]

[0214] in The uncertainty lies in the carbon emissions before the retrofit. The uncertainty lies in the carbon emissions after the renovation. It is the uncertainty of the carbon emission reduction (difference); This is the normal number of measurement points; This refers to the number of abnormal measurement points;

[0215] The report includes, but is not limited to:

[0216] Time series maps and spatial distribution heatmaps of carbon emission reductions at specific points and in specific regions;

[0217] Uncertainty analysis and confidence intervals (default 95% CI);

[0218] List of abnormal time periods / abnormal points and explanations of their causes;

[0219] Carbon emission factors used Explanation of source, version, and applicable period;

[0220] Input data summary (S210–S250 output clean dataset identifier), key parameters (S410–S450 output thermal parameter set identifier), and calculation version number to ensure traceability and reproducibility of results; output formats support CSV, JSON, PDF (report), and structured database writing, and reports should include signatures and timestamps for auditing purposes.

[0221] Example 2: Figure 2 As shown, this embodiment provides an intelligent accounting system for building carbon emission reduction based on intelligent algorithms, including:

[0222] The multimodal sensing acquisition module is used to deploy wireless sensing nodes on the inner and outer surfaces of the building envelope and at environmental boundary conditions to collect temperature, heat flow, humidity, infrared thermal images and meteorological data in real time, and to perform time synchronization and spatial calibration on the data to form a structured detection dataset.

[0223] The thermal-humidity coupling feature recognition module is used to identify suspected thermal bridge areas and significantly humid areas in the building envelope based on features such as temperature difference field, heat flux density gradient field and dew point temperature difference, and to construct a coupling sensitive partition mask to constrain the subsequent modeling area.

[0224] The thermal-humidity coupled parameter inversion module is used to perform joint inversion of heat transfer coefficient, thermal bridge additional heat transfer term and moisture permeability parameter under the constraint of the sensitive partition mask, using Bayesian inversion and Markov chain Monte Carlo sampling algorithm, and calculate the set of thermal performance parameters based on residual convergence control and confidence interval.

[0225] The energy consumption deviation correction and uncertainty propagation module is used to convert the thermal bridge effect and the moisture effect into the equivalent heat transfer coefficient, correct the deviation of the thermal resistance and energy consumption calculation results of the building envelope, and calculate the energy consumption uncertainty through the error propagation model.

[0226] The carbon emission reduction calculation module is used to calculate the carbon emission under the baseline scenario and the transformation scenario point by point based on the corrected energy consumption results and carbon emission factors, and obtain the total carbon emission reduction by point-by-point difference and spatial summarization, while identifying and removing abnormal periods and abnormal measurement points.

[0227] The output and audit module is used to output accounting results including point-by-point carbon emissions, carbon emission reductions, uncertainties, and anomaly indicators, and generate an accounting audit report with timestamps and parameter version numbers, so as to realize the traceability and reproducibility of the carbon emission accounting process.

[0228] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0229] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0230] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for intelligently calculating building carbon emission reductions based on intelligent algorithms, characterized in that, Includes the following steps: S1. Establish paired measuring points consisting of indoor surface temperature, outdoor surface temperature, heat flux density, and surface relative humidity; deploy paired sensors and infrared imaging devices in the detection area, record meteorological conditions and operating load as boundary conditions, and complete the consistency calibration of time reference and coordinate reference. S2. Obtain the time series of temperature, heat flow, humidity, weather and infrared thermal image corresponding to the detection area, perform time alignment, spatial registration and outlier removal to form a multimodal clean dataset for parameter identification and area determination; S3. The surface temperature difference texture of infrared thermal images is fused and analyzed with the heat flow and temperature and humidity changes of paired measurement points to locate suspected thermal bridge areas and quantify the thermal bridge characteristic intensity. Based on the humidity time series and surface temperature time series, extract the humidity state index and output the thermal bridge partition mask and the humidity partition mask; S4. Using the multimodal cleaning dataset, the thermal bridge partition mask, and the moisture-containing partition mask as constraints, perform thermal and moisture coupling parameter identification to obtain a set of thermal parameters for the building envelope, and provide an uncertainty assessment of the set of thermal parameters. S5. Convert the effects of thermal bridging and humidity into equivalent heat transfer coefficients, and generate the time series and spatial distribution of the equivalent heat transfer coefficients; perform systematic deviation correction on the thermal resistance, heat transfer coefficient and energy consumption estimates obtained based on steady-state assumptions, and form the corrected thermal performance results and their confidence intervals.

2. The intelligent accounting method for building carbon emission reduction based on intelligent algorithms according to claim 1, characterized in that, It also includes S6, which uses the corrected thermal performance results and boundary conditions as input to calculate the carbon emissions of the baseline operation and the carbon emissions after the implementation of energy-saving measures, and obtains the time series and statistical results of carbon emission reduction; and generates the accounting confidence interval, abnormal period identifier and traceability report based on the uncertainty assessment, as the accounting output of building carbon emission reduction.

3. The intelligent accounting method for building carbon emission reduction based on intelligent algorithms according to claim 1, characterized in that, S1 includes: Determine the geometric boundaries and measurement point layout range of the building envelope detection area, and divide the sensor layout grid according to the structural characteristics of the building's exterior walls, doors, windows, or roof to ensure that each measurement point corresponds to a clear spatial coordinate; Temperature sensors, heat flow sensors, and humidity sensors are arranged in pairs on the inner and outer surfaces of the detection area, and an infrared imaging device is set up to ensure the consistency between infrared imaging and the spatial distribution of measurement points. Meteorological boundary conditions corresponding to the collection and detection area, including outdoor temperature, relative humidity, wind speed, and solar radiation intensity information, are used as external drivers for subsequent calculations.

4. The intelligent accounting method for building carbon emission reduction based on intelligent algorithms according to claim 3, characterized in that, S1 also includes: Establish a unified time and spatial coordinate reference, and synchronously calibrate the sampling frequency, sampling time and coordinate position of each measuring point and infrared device to ensure the spatiotemporal consistency of subsequent data fusion; The output includes an observation configuration file containing sensor configuration parameters, sampling frequency, and spatial point coordinates, which can be used for subsequent multimodal acquisition.

5. The intelligent accounting method for building carbon emission reduction based on intelligent algorithms according to claim 1, characterized in that, S2 include: The sensor network and infrared imaging device are activated according to the observation configuration file to collect raw time-series data of indoor surface temperature, outdoor surface temperature, heat flux density, surface relative humidity, infrared thermal image and meteorological boundary conditions. Time alignment and spatial registration are performed on the collected data to ensure that each physical quantity has a one-to-one corresponding sampling position in space at each time step; The system performs outlier removal and signal correction on the acquired data, including removing acquisition jitter, drift, and missing measurement segment interpolation markers, and generating a continuous and usable time series of physical quantities. The processed temperature, heat flux, humidity, infrared thermal image, and meteorological boundary condition data are fused into a multimodal clean dataset, preserving the correspondence between timestamps, spatial coordinates, and corresponding physical attributes. Output a multimodal cleaning dataset for subsequent thermal bridge and moisture content identification.

6. The intelligent accounting method for building carbon emission reduction based on intelligent algorithms according to claim 1, characterized in that, S3 includes: Based on the multimodal cleaning dataset, the surface temperature difference characteristics and heat flow change characteristics of each measuring point are extracted, and the temperature difference field and heat flow field corresponding to the infrared thermal image are constructed. By utilizing the temperature difference texture features in infrared thermal images, matching the heat flux density changes at the measurement points, identifying areas of abnormally concentrated temperature gradients, locating suspected thermal bridge areas, and extracting thermal bridge morphological features; Based on the time series of indoor and outdoor surface temperature and humidity, the humidity state index is calculated, the areas with significant local humidity are identified, and the humidity zoning results are generated.

7. The intelligent accounting method for building carbon emission reduction based on intelligent algorithms according to claim 6, characterized in that, S3 also includes: By combining the spatial overlap between suspected thermal bridge regions and significantly humid regions, thermal bridge partition masks and humid region masks are obtained, which are used for regional constraints in subsequent parameter identification. The output includes the recognition results of the thermal bridge partition mask, the wet partition mask, and local feature parameters, ensuring complete consistency with the coordinate baseline of the dataset.

8. The intelligent calculation method for building carbon emission reduction based on intelligent algorithms according to claim 1, characterized in that, S4 includes: Under the regional constraints of thermal bridge partition mask and humid partition mask, a heat and moisture coupled heat transfer model is constructed. The model parameters include the baseline heat transfer coefficient, thermal bridge influence parameters and humid state parameters. The model parameters are retrieved using multi-time period heat flow, temperature and humidity data to form a set of thermal parameters with consistent temporal and spatial distribution. Quantitative values ​​and confidence intervals of thermal bridge influence parameters and moisture content parameters are obtained through least squares fitting, Bayesian estimation, or other identification algorithms. The identified set of thermal parameters is associated with the coordinate datum to ensure that the parameter value of each measuring point corresponds to its spatial location. The output includes a set of thermal parameters including the reference heat transfer coefficient, thermal bridge effect parameters, humid state parameters, and uncertainties.

9. The intelligent calculation method for building carbon emission reduction based on intelligent algorithms according to claim 1, characterized in that, S5 include: Based on the set of thermal parameters, the thermal bridge influence parameters and the humid state parameters are converted into corresponding heat transfer coefficient increments, forming thermal bridge increment terms and humidity increment terms; The thermal bridge increment term, the moisture content increment term, and the reference heat transfer coefficient are weighted and superimposed to calculate the spatiotemporal distribution of the equivalent heat transfer coefficient at different measurement points and time steps. Based on the equivalent heat transfer coefficient, a systematic deviation correction is made to the thermal resistance and energy consumption estimation results under the traditional steady-state assumption; Confidence interval analysis is performed on the corrected thermal performance results to form an assessment of the consistency between the equivalent heat transfer coefficient distribution and the corrected results. The output corrected thermal performance results and corresponding confidence intervals provide a realistic input for carbon emission reduction calculation.

10. The intelligent calculation method for building carbon emission reduction based on intelligent algorithms according to claim 2, characterized in that, S6 include: Using the corrected thermal performance results and meteorological boundary conditions as inputs, the baseline operating carbon emissions are calculated. Calculate the carbon emissions after energy saving based on the thermal performance status after implementing energy-saving measures. A time series of carbon emission reductions is generated by comparing the carbon emissions during baseline operation with those after energy conservation. Based on the uncertainty of the thermal and wet coupling parameters, calculate the confidence interval of carbon emission reduction and identify abnormal time periods and abnormal measurement points; Output carbon emission reduction calculation results, confidence intervals, and traceable reports to achieve accurate, stable, and verifiable calculation of building carbon emission reductions.

Images