A plant system carbon emission dynamic monitoring and tracing method and system

Abstract

This invention discloses a method and system for dynamic monitoring and source tracing of carbon emissions in a plant system, relating to the field of data processing technology. The method includes: discretizing the plant space into multiple micro-environment control volumes; performing asymmetric directed spatial interpolation mapping on carbon emission data using real-time HVAC airflow vectors to obtain reconstructed carbon emission data for each control volume; obtaining a disorder penalty parameter based on spatial distribution probability and its logarithmic operation result; extracting the degradation ratio by comparing real-time and historical data to obtain a decay parameter characterizing equipment aging; coupling the disorder penalty parameter and decay parameter with a risk vector, combined with comprehensive production load rate constraints, to generate an adaptive safety envelope; and generating source tracing and location results based on the comparison results and the numerical dominance of the parameters. This invention solves the technical problems of low monitoring accuracy and difficulty in decoupling and tracing the risks of sudden failures and gradual equipment aging in existing monitoring systems, achieving precise location and full-process control of carbon emission monitoring.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, specifically to a method and system for dynamic monitoring and tracing of carbon emissions in a plant system. Background Technology

[0002] During the high-load continuous operation of modern industrial plant systems, the monitoring and anomaly tracing of carbon emissions face complex dynamic environmental challenges. Existing monitoring methods usually directly adopt the natural diffusion model of the outdoor macro-meteorological field or conventional static geometric grid division, ignoring the forced convection effect dominated by the HVAC system in the indoor confined space. This leads to deviations in the spatial interpolation algorithm based on omnidirectional symmetric distance attenuation when reconstructing the carbon emission concentration distribution of the plant's underlying environment, making it difficult to accurately restore the diffusion pattern of pollutants.

[0003] Meanwhile, existing technologies, when assessing sudden anomalies in spatial consistency, largely rely on the ratio of range to mean. This calculation method is susceptible to interference from single-point data anomalies or overall system shutdowns, making it difficult to effectively quantify the increase in spatial distribution polarization and disorder caused by local pipeline leaks or sudden equipment shutdowns. Furthermore, when identifying the gradual increase in carbon emissions from core power equipment due to long-term mechanical wear or pipeline aging, existing methods typically rely on manually set preset fluctuation tolerances. This makes it difficult to isolate the interference from environmental baselines and load fluctuations, and cannot adaptively extract the proportion contributed by equipment degradation itself, resulting in limitations in the diagnostic accuracy of chronic hidden dangers.

[0004] Furthermore, existing early warning mechanisms often use fixed alarm boundaries or simple linear addition and subtraction to adjust thresholds, failing to establish a correlation between carbon emission baselines and the actual comprehensive production load rate of the factory. This can easily lead to missed alarms at low loads and false alarms at full production and high loads. Moreover, facing the dual risks of sudden spatial imbalances and gradual time-based losses, existing methods lack mathematical norm constraints and multi-dimensional risk decoupling mechanisms. This makes it difficult for the system to distinguish whether the abnormal cause is a sudden failure of the plant support system or the physical aging of the core power unit after detecting excessive carbon emissions. Consequently, it cannot provide accurate directional tracing and guidance for on-site operation and maintenance, thus restricting the efficiency of low-carbon operation and emission reduction of the plant system. Summary of the Invention

[0005] To address the technical problems mentioned above, this invention provides a method and system for dynamic monitoring and tracing of carbon emissions in a plant system. This solves the problems of low accuracy in carbon emission monitoring caused by the use of static monitoring diffusion models and fixed alarm thresholds in complex indoor plant environments, as well as the inability to effectively decouple and trace the dual risks of sudden spatial failures and gradual equipment aging.

[0006] A method for dynamic monitoring and source tracing of carbon emissions in a plant system includes: discretizing the plant space based on the plant's physical boundaries to obtain multiple micro-environment control bodies; performing asymmetric directed spatial interpolation mapping on the real-time raw carbon emission data collected by monitoring nodes based on real-time HVAC airflow vectors to obtain real-time carbon emission reconstruction data for each micro-environment control body; obtaining spatial distribution probability based on the real-time carbon emission reconstruction data; obtaining a disorder penalty parameter based on the spatial distribution probability and its logarithmic operation result; obtaining the percentage of individual unit degradation increment based on the real-time carbon emission reconstruction data and corresponding historical data; obtaining a decay parameter based on the percentage of degradation increment for each individual unit across the entire region; obtaining a reasonable emission baseline based on the carbon emission baseline quota and the overall production load rate; obtaining a sum of squares based on the disorder penalty parameter and the decay parameter; obtaining a root mean square value based on the sum of squares; obtaining a safe contraction ratio by subtracting the root mean square value from a constant; generating an adaptive safety envelope based on the reasonable emission baseline and the safe contraction ratio; locking the corresponding physical location when the real-time carbon emission reconstruction data is greater than the adaptive safety envelope; and generating source tracing results based on the numerical comparison of the disorder penalty parameter and the decay parameter.

[0007] Optionally, based on the physical boundary of the plant, the plant space is discretized to obtain multiple micro-environment control volumes, including: dividing the plant space into virtual data analysis and control volume elements that are logically and thermodynamically topologically independent according to the physical boundary of the plant and the distribution of power facilities, thereby obtaining multiple micro-environment control volumes.

[0008] Optionally, based on the real-time HVAC airflow vector, an asymmetric directed spatial interpolation mapping is performed on the raw carbon emission data collected by the monitoring nodes, including: determining the position vector pointing from the raw data collection location to the geometric center of the microenvironment control volume; obtaining the spatial angle based on the position vector and the real-time HVAC airflow vector; obtaining the downwind influence factor based on the spatial angle; obtaining the wind speed percentage based on the real-time wind speed scalar and the rated maximum wind speed scalar; obtaining the distance reduction factor based on the wind speed percentage and the downwind influence factor; and obtaining the aerodynamic effective distance based on the distance reduction factor and the straight-line distance between the raw data collection location and the microenvironment control volume.

[0009] Optionally, the asymmetric directed spatial interpolation mapping includes: performing a cosine operation on the spatial angle to obtain a cosine value; taking the maximum value between the cosine value and the constant zero to obtain the downwind influence factor; dividing the real-time wind speed scalar by the rated maximum wind speed scalar to obtain the wind speed proportion; subtracting the product of the wind speed proportion and the downwind influence factor from the constant one to obtain the distance reduction coefficient; multiplying the straight-line distance by the distance reduction coefficient to obtain the aerodynamic effective distance; calculating the negative square of the aerodynamic effective distance to obtain the distance attenuation weight; multiplying the real-time carbon emission raw data of each monitoring node by the corresponding distance attenuation weight and summing the results to obtain the weighted emission; summing the distance attenuation weights of each monitoring node to obtain the weighted sum; and dividing the weighted emission by the weighted sum to obtain the real-time carbon emission reconstructed data.

[0010] Optionally, the spatial distribution probability is obtained based on the real-time carbon emission reconstruction data, including: summing the real-time carbon emission reconstruction data of all microenvironment control bodies in the entire domain at the current detection time point to obtain the total carbon emission data of the entire domain; and dividing the real-time carbon emission reconstruction data of each microenvironment control body by the total carbon emission data of the entire domain to obtain the spatial distribution probability of each microenvironment control body.

[0011] Optionally, the disorder penalty parameter is obtained based on the spatial distribution probability and its logarithmic operation result, including: performing natural logarithmic operation on each spatial distribution probability to obtain the natural logarithmic operation result; obtaining the negative spatial distribution discrete term based on each spatial distribution probability and the corresponding natural logarithmic operation result; obtaining the dispersion ratio based on the natural logarithmic operation result of the spatial distribution discrete term and the total number of micro-environment control volumes; and adding the constant to the dispersion ratio to obtain the disorder penalty parameter.

[0012] Optionally, the percentage of individual cell degradation increment is obtained based on real-time carbon emission reconstruction data and corresponding historical data, including: subtracting the historical data at a preset historical time point from the real-time carbon emission reconstruction data at the current detection time point to obtain the drift increment; if the drift increment is positive, dividing the drift increment by the real-time carbon emission reconstruction data at the current detection time point to obtain the percentage of individual cell degradation increment; if the drift increment does not exceed zero, setting the percentage of individual cell degradation increment to zero.

[0013] Optionally, the attenuation parameter is obtained based on the percentage of degradation increment of each individual unit in the entire domain, including: performing a maximum value calculation on the percentage of degradation increment of each individual unit in all microenvironment control bodies in the entire domain to obtain the attenuation parameter.

[0014] Optionally, a reasonable emission baseline can be obtained based on the carbon emission baseline allowance and the overall production load rate, including: multiplying the carbon emission baseline allowance by the overall production load rate to obtain a reasonable emission baseline constrained by the production load.

[0015] Optionally, generating an adaptive safety envelope includes: squaring and summing the disorder penalty parameter and the attenuation parameter to obtain a sum of squares; dividing the sum of squares by a constant two and then taking the square root to obtain the root mean square value; subtracting the root mean square value from the constant one to obtain the safe contraction ratio; and multiplying the reasonable emission baseline by the safe contraction ratio to generate the adaptive safety envelope.

[0016] Optionally, the corresponding physical location is locked, and the source tracing result is generated based on the numerical comparison of the disorder penalty parameter and the attenuation parameter, including: when the real-time carbon emission reconstruction data of any microenvironment control body is greater than the adaptive safety envelope, the physical location corresponding to the microenvironment control body is locked; if the disorder penalty parameter is greater than the attenuation parameter, the source tracing result representing the sudden failure of the supporting environment is output; if the attenuation parameter is greater than the disorder penalty parameter, the source tracing result representing the physical aging of the core power equipment is output.

[0017] A dynamic carbon emission monitoring and tracing system for plant operations is also provided. This system implements dynamic carbon emission monitoring and tracing methods for plant operations. The system includes: a spatial reconstruction module for acquiring real-time carbon emission reconstruction data; a disorder assessment module for acquiring disorder penalty parameters; a decay analysis module for acquiring decay parameters; and a dynamic envelope decision module for generating an adaptive safety envelope. This dynamic envelope decision module also generates tracing and location results.

[0018] The beneficial effects of this invention are reflected in: In the overall plant system's dynamic carbon emission monitoring and source tracing method, firstly, by introducing real-time HVAC airflow vectors to construct an asymmetric directed spatial interpolation mapping, the data from physical monitoring nodes is transformed into reconstructed data for the micro-environment control volume. This overcomes the limitations of existing omnidirectional symmetric interpolation in confined indoor spaces. Higher weights are assigned to upwind nodes based on the airflow direction, ensuring the data flow reconstruction conforms to aerodynamic diffusion laws and reducing spatial data reconstruction errors. Secondly, spatial distribution probability and logarithmic operations are used to extract disorder penalty parameters, measuring the polarization state of carbon emission spatial distribution from the information entropy dimension, enabling the identification of spatial imbalances caused by sudden leaks, etc. Thirdly, the attenuation parameter is extracted using the drift increment ratio between current and historical data, without the need for preset tolerances. Under this premise, the progressive aging degree of equipment caused by mechanical wear is quantified, reducing the interference of environmental background noise and normal production fluctuations. Furthermore, this solution binds the carbon emission control boundary with the comprehensive production load rate and uses a risk vector coupling mechanism to fuse the disorder penalty parameter and the attenuation parameter to generate an adaptive safety envelope. This achieves boundary sinking under low load and boundary rising under high load, and quantitative adjustments are made when risks are detected, reducing the false alarm and false negative rates caused by fixed thresholds. Finally, by comparing the results of exceeding limits and the numerical values ​​of two-dimensional risk factors, it is possible to distinguish whether the anomaly originates from a sudden failure of the supporting environment or physical loss of the core unit, providing executable traceability and location instructions for the industrial site and improving the low-carbon operation and maintenance efficiency of the plant. Attached Figure Description

[0019] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0020] Figure 1 This is a schematic diagram illustrating the steps of the dynamic monitoring and source tracing method for carbon emissions in the plant system of the present invention; Figure 2 This is a schematic diagram of a portion of steps S1 in the carbon emission dynamic monitoring and tracing method of the plant system of the present invention; Figure 3 This is a schematic diagram of a portion of step S2 in the carbon emission dynamic monitoring and tracing method of the plant system of the present invention; Figure 4 This is a schematic diagram of a portion of step S3 in the plant system carbon emission dynamic monitoring and tracing method of the present invention; Figure 5 This is a schematic diagram of a portion of step S4 in the dynamic monitoring and tracing method for carbon emissions in the plant system of the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0022] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0023] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, the terms first, second, etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0024] This invention provides a method for dynamic monitoring and source tracing of carbon emissions in a plant system, such as... Figure 1 As shown, in one embodiment, the method includes: S1. Based on the physical boundaries of the plant, the space within the plant system is discretized into multiple micro-environment control volumes. Real-time raw carbon emission data collected by monitoring nodes and real-time HVAC airflow vectors of the plant are acquired. Based on the real-time HVAC airflow vectors, asymmetric directed spatial interpolation mapping is performed on the real-time raw carbon emission data to obtain the reconstructed real-time carbon emission data for each micro-environment control volume. In this step, a micro-environment control volume refers to a virtual data analysis and control volume element that is logically and thermodynamically topologically independent, divided within the plant system based on the physical boundaries of the plant and the distribution of power facilities. This volume is used to accurately pinpoint the localized carbon emission characteristics of independent functional areas such as refrigeration, gas supply, and power supply. Asymmetric directed spatial interpolation mapping is a spatial reconstruction algorithm that inverts and reconstructs the raw discrete data of local monitoring nodes into high-density grid-distributed data. Its characteristic is that the convection direction and wind speed of the indoor HVAC airflow are incorporated into the distance attenuation coefficient, making the spatial reconstruction shape of carbon concentration conform to the aerodynamic diffusion law of fluid dynamics.

[0025] S2. Based on the real-time carbon emission reconstruction data of each microenvironment control body, obtain its spatial distribution probability in the entire domain. Based on the spatial distribution probability and its logarithmic operation, obtain the disorder penalty parameter characterizing the spatial imbalance state of the system. In this step, the disorder penalty parameter is a dimensionless risk quantification factor dynamically calculated by measuring the polarization state of the spatial probability distribution of carbon concentration in the entire domain using the Shannon information entropy principle in statistical thermodynamics. When the carbon concentration in the entire domain accumulates in a large number of local areas and the spatial distribution becomes unbalanced due to sudden pipeline leakage or single-machine failure, this parameter adaptively increases to measure the degree of disorder caused by sudden spatial consistency deterioration.

[0026] S3. Compare the real-time carbon emission reconstruction data of each microenvironment control body with its corresponding historical data to extract the percentage of individual degradation increment. Based on the percentage of individual degradation increment of each microenvironment control body in the entire domain, obtain the attenuation parameter characterizing equipment aging. In this step, the percentage of individual degradation increment refers to the dimensionless percentage calculated by dividing the positive emission increase of a specific microenvironment control body from the current detection time point to the beginning of the sliding time window by the real-time carbon emission reconstruction data of that microenvironment control body at the current time. Essentially, it is the absolute percentage contributed to the current emissions by the medium- and long-term slow variables such as mechanical wear or pipeline aging of the microenvironment control body itself after stripping away the environmental baseline and macro fluctuations. The attenuation parameter is a global risk indicator determined by scanning and extracting the maximum value of the percentage of individual degradation increment of all microenvironment control bodies in the entire domain. It is used to maintain a high diagnostic sensitivity for the most severe local chronic wear and progressive failure risks across the entire plant, without any objectively preset thresholds.

[0027] S4. Obtain the overall production load rate and carbon emission baseline quota of the plant management system. Based on the constraint of the overall production load rate, couple the disorder penalty parameter and the attenuation parameter with a risk vector to generate an adaptive safety envelope. Compare the current real-time carbon emission reconstruction data of each micro-environment control body with the adaptive safety envelope, and generate source tracing and location results based on the comparison results and the numerical comparison results of the disorder penalty parameter and the attenuation parameter. In this step, the adaptive safety envelope refers to the carbon emission monitoring safety boundary that is dynamically expanded and contracted according to the actual working conditions and operational risks, generated based on the nonlinear collaborative calculation of multi-dimensional risk vector and real-time overall production load rate. It actively sinks when the plant is operating at low load to prevent ineffective energy loss, rises when the plant is operating at high load and without operational risks to ensure production capacity, and dynamically accelerates and tightens when the dual-dimensional risks of spatial disorder and time loss are superimposed, breaking the technical limitations of fixed thresholds and linear addition and subtraction.

[0028] In this embodiment, it should be noted that in S1, the physical space of the plant system is discretized into multiple micro-environment control bodies, and the real-time airflow vector of the indoor HVAC system is introduced to perform asymmetric directional spatial interpolation mapping on the data collected by the physical monitoring nodes. In a specific application, it is assumed that the rated maximum wind speed is 10 m / s and the real-time wind speed is 6 m / s. When the original carbon emission data of monitoring node 1 is 125 kg / h and it is located 10 m upwind of a certain control body (airflow angle is 0 degrees), and the data of node 2 is 45 kg / h and it is located 20 m downwind (airflow angle is 180 degrees), the straight-line distance is reduced and adjusted using the proportion of airflow direction and wind speed. The effective aerodynamic distance of node 1 is reduced to 4 m, while that of node 2 remains at 20 m, which gives the upwind node a higher weight in the interpolation. After reconstruction and calculation, the carbon emission data for the four control bodies across the entire area reached 120 kg / h, 20 kg / h, 30 kg / h, and 30 kg / h respectively, totaling 200 kg / h. This technique improves the existing isotropic spatial diffusion model into an aerodynamic diffusion mapping within a confined space, overcoming data reconstruction bias caused by indoor airflow interference and improving the monitoring accuracy of concentration distribution in the plant's underlying environment.

[0029] In S2, based on the real-time carbon emission data of each microenvironment control body reconstructed in the previous step, the polarization and imbalance of spatial distribution are further quantitatively assessed. First, the spatial distribution probability of each control body in the entire domain is calculated. Taking the reconstructed data as an example, the spatial distribution probabilities of the four control bodies are 0.6, 0.1, 0.15, and 0.15, respectively. Then, logarithmic logic is used to solve the discreteness of this probability distribution, extracting the negative discrete term representing the carbon emission aggregation phenomenon, and comparing it with the maximum theoretical distribution state corresponding to the total number of microenvironment control bodies in the entire domain. Through this comparison mechanism, the current disorder penalty parameter is finally calculated to be 0.202. This technique differs from the conventional simple algorithm that relies on comparing maximum and minimum values; instead, it measures the spatial abnormal accumulation caused by sudden leaks or partial shutdowns from the perspective of the overall spatial information distribution. This processing method can effectively filter out local numerical disturbances caused by single-point sensor anomalies, objectively reflecting the current plant operations exhibiting a certain degree of directional risk of local spatial distribution imbalance.

[0030] In S3, to isolate the baseline interference caused by macro-load changes and independently assess the physical losses accumulated by power equipment over operating time, a longitudinal comparison is made between the real-time reconstructed data of each micro-environment control body and the data at preset historical time points. Following the aforementioned spatial reconstruction and probability calculation scenario, the current carbon emission reconstructed data for the first micro-environment control body is 120 kg / h, while the extracted data at the historical baseline time point is 84 kg / h. The positive drift increment is calculated as 36 kg / h by subtraction. Further dividing this increment by the current reconstructed data yields a single-unit degradation increment ratio of 0.30 for this micro-environment control body. Assuming the degradation ratios of the other three control bodies are all 0, the maximum value of 0.30 across the entire domain is extracted as the overall degradation parameter characterizing equipment aging. This technique no longer relies on fixed alarm tolerances preset by human experience, but directly calculates and extracts the progressive degradation ratio contributed by mechanical wear or pipeline aging from the historical operating trajectory of the plant equipment itself. Its beneficial effect is that it can adaptively quantify the chronic operational hazards of core power equipment, improving the sensitivity and accuracy of medium- and long-term operation and maintenance diagnosis.

[0031] In S4, the carbon emission early warning standard is intersected with the factory's actual production plan, and the aforementioned multi-dimensional risk parameters are integrated for dynamic threshold adjustment and source tracing. The current overall production load rate of the plant system is 0.90, and the carbon emission baseline quota is 150 kg / h. These two values ​​combine to form a reasonable emission baseline of 135 kg / h constrained by the load. Subsequently, the disorder penalty parameter of 0.202 and the attenuation parameter of 0.30 are coupled using a risk vector to calculate the comprehensive risk penalty weight, which is then used to safely shrink the 135 kg / h baseline, generating an adaptive safety envelope of 100.48 kg / h. When the data of the first control unit (120 kg / h) exceeds 100.48 kg / h, a location alarm is triggered. Simultaneously, a comparison reveals that the attenuation parameter of 0.30 is greater than the disorder penalty parameter of 0.202, indicating that the anomaly is caused by aging core equipment rather than a sudden pipeline failure. This technology enables the control boundary to dynamically fluctuate with the production load, preventing false alarms and missed alarms caused by fixed thresholds. Furthermore, by comparing the magnitude of risk parameters, it decouples fault types and improves the targeted nature of on-site maintenance and troubleshooting.

[0032] In summary, the overall plant carbon emission dynamic monitoring and tracing method firstly, by introducing real-time HVAC airflow vectors to construct an asymmetric directed spatial interpolation mapping, the data from physical monitoring nodes is transformed into reconstructed data for the micro-environment control body. This overcomes the limitations of existing omnidirectional symmetric interpolation in confined indoor spaces. Higher weights are assigned to upwind nodes based on the airflow direction, ensuring the data flow reconstruction conforms to aerodynamic diffusion laws and reducing spatial data reconstruction errors. Secondly, spatial distribution probability and logarithmic operations are used to extract disorder penalty parameters, measuring the polarization state of carbon emission spatial distribution from the information entropy dimension, enabling the identification of spatial imbalances caused by sudden leaks, etc. Thirdly, the attenuation parameter is extracted using the drift increment ratio between current and historical data, without the need for preset tolerances. Under this premise, the progressive aging degree of equipment caused by mechanical wear is quantified, reducing the interference of environmental background noise and normal production fluctuations. Furthermore, this solution binds the carbon emission control boundary with the comprehensive production load rate and uses a risk vector coupling mechanism to fuse the disorder penalty parameter and the attenuation parameter to generate an adaptive safety envelope. This achieves boundary sinking under low load and boundary rising under high load, and quantitative adjustments are made when risks are detected, reducing the false alarm and false negative rates caused by fixed thresholds. Finally, by comparing the results of exceeding limits and the numerical values ​​of two-dimensional risk factors, it is possible to distinguish whether the anomaly originates from a sudden failure of the supporting environment or physical loss of the core unit, providing executable traceability and location instructions for the industrial site and improving the low-carbon operation and maintenance efficiency of the plant.

[0033] like Figure 2As shown, in one specific embodiment, S1 includes: S11, determining the position vector pointing from the original data acquisition location to the geometric center of the microenvironment control body, and obtaining the spatial angle between the position vector and the real-time HVAC airflow vector; S12. Extract the positive downwind influence factor based on the spatial angle, and calculate the wind speed ratio between the real-time wind speed scalar and the rated maximum wind speed scalar. S13. Based on the wind speed ratio and downwind influence factor, the straight-line distance between the original data acquisition location and the microenvironment control body is reduced to obtain the effective aerodynamic distance. S14. Using the negative exponent of the effective aerodynamic distance as a weight, perform a weighted summation calculation on the raw real-time carbon emission data to obtain the reconstructed real-time carbon emission data for each microenvironment control body. Specifically, the calculation logic for obtaining the reconstructed real-time carbon emission data for each microenvironment control body can be expressed as follows:

[0034] Among them, D k (t): Real-time reconstructed carbon emission data of the k-th microenvironment control system at the current detection time t; N: Total number of monitoring nodes; E j (t): Real-time raw carbon emission data collected by the j-th monitoring node at the current monitoring time t; d kj : The straight-line distance between the original data acquisition location and the microenvironment control volume, referring to the Euclidean straight-line distance between the geometric center of the k-th microenvironment control volume and the j-th monitoring node; v(t): The real-time wind speed scalar at the current detection time t; v max : Rated maximum wind speed scalar value for HVAC systems in a plant; θ jk : The spatial angle between the position vector pointing from the j-th monitoring node to the geometric center of the k-th microenvironment control volume and the real-time HVAC airflow vector; max: Take the maximum value.

[0035] In this embodiment, it should be noted that in S11, the main operation is to determine the position vector pointing from the original data acquisition location to the geometric center of the microenvironment control body, and to obtain the spatial angle between this position vector and the real-time HVAC airflow vector in the plant. In the scenario of a large semiconductor plant's power air compressor station, the spatial relationship between the microenvironment control body and two high-precision monitoring nodes is established through physical coordinate mapping. Specifically, the straight-line length of node 1 from the geometric center of the control body is 10m, and the distance of node 2 is 20m. At the same time, by accessing the environmental management system to extract the airflow direction, it is found that node 1 is located directly upwind of the control body, and the spatial angle between its position vector and the airflow vector is 0 degrees; node 2 is located on the leeward side, with a corresponding spatial angle of 180 degrees. Through this step, a homogeneous and undirected physical space is endowed with vector attributes based on fluid dynamics characteristics. This operation overcomes the technical deficiency of conventional monitoring that relies solely on pure geometric straight-line distances, and uses the aerodynamic directionality of the forced convection environment at the bottom of the plant as a prerequisite for data processing, providing basic parameters for accurately reconstructing the pollutant diffusion pattern, making the model more closely resemble the real state of gas transmission within a closed plant.

[0036] In S12, the positive downwind influence factor is extracted based on the previously obtained spatial angle, and the wind speed ratio between the real-time wind speed scalar and the rated maximum wind speed scalar is calculated. Considering the aforementioned air compressor station application scenario, the current rated maximum wind speed scalar of the HVAC main duct is read from the underlying control platform as 10 m / s, and the real-time wind speed scalar is 6 m / s. A division operation yields a wind speed ratio of 0.6. This rated maximum wind speed scalar value is not arbitrarily set, but is obtained by analyzing the nameplate parameters of the core fan equipment in the plant's HVAC system and the factory design specifications. Specifically, the system extracts the historical measured average steady-state flow velocity at the fan outlet under 100% variable frequency full-load conditions, and performs fluid dynamics calculations based on the cross-sectional area and friction coefficient of the main ventilation duct. Finally, the maximum theoretical aerodynamic wind speed under the physical limits of this area is calibrated and sent to the underlying control platform. For example, by reviewing the performance curve of the No. 1 main exhaust fan and combining it with the actual measured data of the anemometer during the full-load acceptance test, the system accurately determined the rated maximum wind speed of the pipeline network to be 10 m / s.

[0037] Furthermore, for node 1, since its spatial angle is 0 degrees and its cosine value is 1, the extracted downwind influence factor is 1. For node 2, its spatial angle is 180 degrees and its cosine value is -1; the maximum value between this and 0 is taken, resulting in a downwind influence factor of 0. The core of this calculation logic lies in the asymmetric nature of carbon emission gas diffusion in indoor environments due to airflow pull. Upwind airflow blows the gas towards the target area, while downwind nodes contribute very little to the concentration in that area. By extracting the positive cosine value and combining it with the wind speed ratio, the specific degree of intervention of airflow intensity on gas transport can be quantified. This technique eliminates interference from leeward side data, ensuring that the extraction of aerodynamic influence factors is based on actual operating conditions, providing dynamic parameters that conform to physical principles for subsequent distance reduction.

[0038] In S13, based on the calculated wind speed percentage and downwind influence factor, the straight-line distance between the original data acquisition location and the microenvironment control body is reduced to obtain the aerodynamic effective distance. Continuing the aforementioned numerical calculation, for node 1, with a straight-line distance of 10m, a wind speed percentage of 0.6, and a downwind influence factor of 1, its aerodynamic effective distance is calculated to be 10×(1-0.6×1)=4m; for node 2, with a straight-line distance of 20m, a wind speed percentage of 0.6, and a downwind influence factor of 0, its aerodynamic effective distance is calculated to be 20×(1-0.6×0)=20m. The physical background of this calculation logic is that directional airflow is equivalent to shortening the transmission distance between the pollution source and the receptor in aerodynamics, and the greater the wind speed, the stronger the shortening effect. In contrast, the leeward node maintains its original physical straight-line distance because there is no forward airflow. This technique transforms a static geometric space model into a dynamic aerodynamic space model, solving the distance deviation caused by existing isotropic monitoring models in complex indoor confined spaces, and making the reconstructed distance conform to the actual transport law of gas under forced ventilation conditions.

[0039] In S14, the negative exponent of the calculated aerodynamic effective distance is used as a weight to perform a weighted summation of the raw real-time carbon emission data, ultimately obtaining the reconstructed real-time carbon emission data for each microenvironment control body. Specifically, the calculation logic for obtaining the reconstructed real-time carbon emission data for each microenvironment control body is as follows: .

[0040] In existing industrial environmental monitoring, isotropic spatial distance attenuation algorithms (such as the inverse distance weighting method) are commonly used. These algorithms assume that the diffusion rate and concentration attenuation of gas in space are completely equal in all directions. However, in the actual dynamic environment of a plant, space is usually limited, and there is forced convection dominated by HVAC systems. If the traction effect of this mechanical airflow is ignored, simply applying physical straight-line distances for data reconstruction will lead to a large calculation deviation in the concentration in the downwind region, while the upwind region will be overestimated, resulting in aerodynamic data distortion.

[0041] To address this technical issue, the above expression introduces the real-time wind speed scalar v(t) and the rated maximum wind speed scalar v0. max The ratio, which constructs a dimensionless wind speed proportion parameter between 0 and 1, directly couples the algorithm model with the physical operating state of the plant.

[0042] At the same time, the expression introduces θ jk , which is the spatial angle between the position vector from the j-th monitoring node to the geometric center of the k-th microenvironment control volume and the real-time HVAC airflow vector. This is calculated by... jk The projection of the airflow direction onto the line connecting the node to the control volume was extracted.

[0043] More importantly, the expression sets max(0,cosθ) jk This filtering function serves to implement directional filtering: airflow will only exert a positive transport effect when the node is located upwind of the control volume (i.e., the angle is less than 90 degrees and the cosine value is positive). The expression uses a constant minus the product of the wind speed percentage and the cosine value to mathematically represent the actual physical straight-line distance d. kj The aerodynamic effective distance was shortened proportionally to obtain a smaller effective distance. When the node is located downwind of the control volume (i.e., the angle is greater than 90 degrees and the cosine value is negative), the filtering function is directly set to 0, which means that the airflow in the upwind direction will not accelerate the gas to diffuse in the opposite direction to the control volume, and the aerodynamic effective distance in this direction remains the original physical straight-line distance.

[0044] Calculations and verifications were performed using data from specific application scenarios: assuming the rated maximum wind speed scalar v of the plant's HVAC ducts... max The wind speed is 10 m / s, and the currently extracted real-time wind speed scalar v(t) is 6 m / s, so the wind speed percentage is 0.6. There are two physical monitoring nodes (N=2). For a specific first microenvironment control body, the real-time raw carbon emission data E1(t) collected by node 1 is 125 kg / h, and the straight-line distance d between node 1 and the geometric center of the control body is... k1 The distance is 10m, and node 1 is located directly upwind, with a spatial angle θ. k1The temperature is 0 degrees; the real-time raw carbon emission data E2(t) collected by node 2 is 45 kg / h, and its straight-line distance d k2 It is 20m, located directly downwind, with a spatial angle θ. k1 The angle is 180 degrees. Substituting into the expression, the cosine value of node 1 is 1, which remains 1 after maximum value filtering. Its straight-line distance reduction factor is 1 - 0.6 × 1 = 0.4. Therefore, the effective aerodynamic distance from node 1 to the control volume is reduced to 10m × 0.4 = 4m. The cosine value of node 2 is -1, which becomes 0 after maximum value filtering. Its reduction factor is 1 - 0.6 × 0 = 1. Therefore, the effective aerodynamic distance of node 2 remains unchanged at 20m. Subsequently, the expression uses the negative square of the effective aerodynamic distance as the distance attenuation weight, and the weight of node 1 is calculated to be 4. -2 =0.0625, the weight of node 2 is 20. -2 =0.0025. Finally, the numerator is the weighted emission amount: 125 kg / h × 0.0625 + 45 kg / h × 0.0025 = 7.8125 + 0.1125 = 7.925 kg / h. The denominator is the sum of the weights: 0.0625 + 0.0025 = 0.065. Performing division yields the real-time carbon emission reconstruction data D for this microenvironment control system. k (t) is approximately 121.92 kg / h (rounded to 120 kg / h for ease of subsequent calculations).

[0045] Through the detailed mathematical calculations described above, it can be seen that this expression successfully quantifies the influence of the HVAC flow field on carbon gas transport, assigning a higher weight to the upwind node than the downwind node. Thus, without the need to increase the number of physical sensors, it accurately restores the true aerodynamic distribution of carbon emissions in a closed space.

[0046] like Figure 3 As shown, in one specific embodiment, S2 includes: S21, summing the real-time carbon emission reconstruction data of all microenvironment control bodies in the entire domain at the current detection time point to obtain the total carbon emission data of the entire domain; S22. Divide the real-time carbon emission reconstruction data of each micro-environment control body by the total carbon emission data of the entire domain to obtain the spatial distribution probability of each micro-environment control body. S23. Perform logarithmic operations on the spatial distribution probabilities of each microenvironment control body, multiply the spatial distribution probabilities of each microenvironment control body by their corresponding logarithmic operation results and sum them to obtain the negative spatial distribution discrete term. S24. Divide the spatial distribution discrete term by the logarithmic result of the total number of micro-environment control volumes to obtain the dispersion ratio; then add the constant 1 to the dispersion ratio to obtain the disorder penalty parameter. Specifically, the calculation logic for obtaining the disorder penalty parameter can be expressed as follows:

[0047] Where C1(t): the disorder penalty parameter generated at the current detection time t; M: the total number of microenvironment control volumes; p k (t): Spatial distribution probability of each microenvironment control body, referring to the spatial distribution probability of the k-th microenvironment control body at the current detection time point t; D k (t): Real-time reconstructed carbon emission data of the k-th microenvironment control body at the current detection time t; D v (t): Real-time carbon emission reconstruction data of the vth microenvironment control body at the current detection time point t.

[0048] In this embodiment, it should be noted that in S21, the main operation is to sum the real-time carbon emission reconstruction data of all micro-environment control bodies in the entire region at the current detection time point, thereby obtaining the total carbon emission data of the entire region. Considering the application scenario of the air compressor station, the previous steps have already obtained the real-time carbon emission reconstruction data of four micro-environment control bodies at the current moment, which are 120 kg / h, 20 kg / h, 30 kg / h, and 30 kg / h respectively. These four specific values ​​are arithmetically added together, i.e., 120 + 20 + 30 + 30, resulting in a total carbon emission data of 200 kg / h for the entire region. This calculation step represents the macroscopic physical total amount of carbon emissions in the entire target plant area at the current detection time point, and is a necessary prerequisite for converting the absolute emission values ​​of each local micro-environment control body into relative spatial distribution probabilities. By acquiring total carbon emissions data across the entire region, a baseline denominator can be established for normalization processing. This allows subsequent data analysis to move beyond the fluctuations in absolute concentration values ​​in a single region and focus on the relative distribution within the overall mass conservation framework of the plant, providing reliable data support for further measuring the balance of spatial structure.

[0049] In step S22, the real-time carbon emission reconstruction data of each microenvironment control body is divided by the previously obtained total carbon emission data for the entire region to obtain the spatial distribution probability of each microenvironment control body. Substituting specific scenario values, the reconstruction data of the first microenvironment control body is 120 kg / h, divided by the total data of 200 kg / h, yielding a spatial distribution probability of 0.6; similarly, the data of the second control body is 20 kg / h, yielding a probability of 0.1; the data of the third and fourth control bodies are both 30 kg / h, yielding probabilities of 0.15 each. The calculation logic of this step is that directly comparing the absolute values ​​of different parameters can easily introduce scale bias. By using division operations, the reconstructed data in mass flow units is converted into dimensionless probability data, mapping the specific physical emissions to the weighted share of each local area in the overall emission space. This technique eliminates background interference caused by fluctuations in overall production load, enabling the examination of carbon emission distribution patterns from a spatial probability structure perspective across different operating conditions. It provides standardized input parameters for assessing the overall spatial discrete state using statistical thermodynamic tools.

[0050] In S23, the spatial distribution probabilities of each microenvironment control entity are logarithmically calculated. The spatial distribution probability of each microenvironment control entity is multiplied by its corresponding logarithmic result and summed to obtain a negative spatial distribution discrete term. According to the aforementioned data, the probabilities of the four control entities are 0.6, 0.1, 0.15, and 0.15, respectively. Their natural logarithms are calculated, and the product of probability and logarithm is summed: 0.6 multiplied by -0.5108, plus 0.1 multiplied by -2.3026, plus two 0.15 multiplied by -1.8971. The summation yields a spatial distribution discrete term of -1.106. This calculation logic aligns with the underlying principle of Shannon's information entropy. In information theory and thermodynamics, the sum of the product of probability and logarithm can quantify the degree of disorder in a microstate. When carbon emissions exhibit a polarized distribution, some probability values ​​are larger while others are smaller, resulting in a smaller absolute value of the negative sum of these products; conversely, in equilibrium, the absolute value of this value reaches its maximum. By extracting this spatial distribution discrete term, the spatial distribution of carbon emissions within the plant system is mathematically characterized, laying the computational foundation for subsequent objective normalization assessments of disorder.

[0051] In S24, the spatial distribution discrete term of the aforementioned negative value is divided by the logarithmic result of the total number of micro-environment control volumes to obtain the dispersion ratio. This ratio is then added to the constant to obtain the disorder penalty parameter. Specifically, the calculation logic for obtaining the disorder penalty parameter characterizing the spatial imbalance state of the system is as follows: The spatial distribution probability .

[0052] In industrial carbon emission monitoring, directly using the variance or range of absolute concentration data to assess spatial anomalies can lead to significant instability in algorithms. For example, when a factory ramps up its overall capacity from low to full based on production orders, the absolute emission values ​​of all controlled entities will increase synchronously. This results in a larger variance mathematically, making existing algorithms prone to misinterpreting this normal production load ramp-up as a spatial anomaly. On the other hand, if a sensor experiences a disconnection and outputs a fixed abnormal value, the numerator of the range algorithm will undergo a sudden change, triggering false alarms.

[0053] To address this technical challenge, firstly, the above expression reconstructs the real-time carbon emission data D of the k-th microenvironment control body at the current detection time t. k (t), divided by the sum of the reconstructed data of all M control bodies in the entire domain. Mass flow rate data with specific physical dimensions (such as kg / h) is transformed into a dimensionless spatial distribution probability p within the interval of 0 to 1. k (t). This division operation physically removes the background interference caused by fluctuations in the total factory capacity, allowing the spatial structure and proportion of emissions to be examined on the same normalized scale, regardless of whether the total discharge is 200 kg / h or 2000 kg / h.

[0054] Next, the expression performs a natural logarithm operation ln(p) on the spatial distribution probability of each control volume. k (t)) and multiply the probability by its logarithm and sum to extract the spatial distribution discrete term with negative values. When in a stable equilibrium state (i.e., the emissions of each control body are equal), the absolute value of this discrete term is the largest; however, when a local pipeline rupture and leakage occurs, causing gas to accumulate in a few spatial polarizations, the probability of some control bodies approaches 1 (its logarithm tends to 0), and the probability of the remaining control bodies approaches 0 (in the limit, plnp also tends to 0). At this time, the absolute value of the entire discrete term will decrease sharply.

[0055] Furthermore, in order to transform this negative discrete term into a positive penalty coefficient that conforms to engineering understanding, the expression divides it by the logarithm ln(M) of the total number of micro-environment control volumes M. ln(M) represents the maximum information entropy that the space can theoretically achieve. Through this division operation, the relative ratio between the current dispersion and the maximum theoretical dispersion is obtained, which is a negative number between -1 and 0.

[0056] Finally, the expression adds the constant 1 to the ratio of the dispersion, inverting the numerical scale so that the more uniform the spatial distribution, the closer the ratio is to -1, and the closer C1(t) is to 0; the more polarized and disordered the spatial distribution, the closer the ratio is to 0, and the closer C1(t) is to 1.

[0057] Based on the aforementioned scenario data, the calculations are as follows: There are four micro-environment control bodies (M=4) in the entire region. Due to local fluctuations at a specific moment, the reconstructed real-time carbon emission data for the four control bodies are 120 kg / h, 20 kg / h, 30 kg / h, and 30 kg / h, respectively. The total carbon emission data for the entire region is the sum of these four values, which is 200 kg / h. The calculated spatial distribution probabilities for each control body are: p1=120 / 200=0.6, p2=20 / 200=0.1, p3=30 / 200=0.15, p4=30 / 200=0.15. Then, substituting into the logarithmic summation unit, the discrete term is calculated: 0.6×ln(0.6)+0.1×ln(0.1)+2×[0.15×ln(0.15)]=0.6×(-0.5108)+0.1×(-2.3026)+2×0.15×(-1.8971)=-0.3065-0.2303-0.5691=-1.1059. The maximum information entropy ln(4)≈1.3863 is obtained. Dividing the two, the dispersion ratio is -1.1059 / 1.3863≈-0.7977. Finally, it is added to the constant to obtain the disorder penalty parameter C1(t)=1-0.7977=0.2023 (in engineering applications, it is usually retained to three decimal places, i.e., 0.202).

[0058] This calculation process proves that the expression can objectively and quantitatively extract the 20.2% spatial imbalance penalty ratio of the current plant operation through pure probability distribution structure changes, without relying on manually preset alarm upper limits, effectively solving the identification interference problem caused by background load changes.

[0059] like Figure 4 As shown, in one specific embodiment, S3 includes: S31, subtracting the historical data at a preset historical time point from the real-time carbon emission reconstruction data of each microenvironment control body at the current detection time point to obtain the drift increment; S32. If the drift increment is positive, divide the drift increment by the real-time carbon emission reconstruction data at the current detection time point to obtain the percentage of individual degradation increment of each microenvironment control body; if the drift increment does not exceed zero, the corresponding percentage of individual degradation increment is set to zero. S33. Obtain the maximum value of the percentage of individual degradation increment of each microenvironment control body in the entire domain, and use the maximum value as the attenuation parameter.

[0060] In this embodiment, it should be noted that in S31, the real-time carbon emission reconstruction data of each microenvironment control body at the current detection time point is subtracted from its historical data at a preset historical time point to obtain the drift increment. The method for determining the preset historical time point is as follows: extract continuous operating logs of the core power equipment over multiple complete maintenance cycles, use a clustering algorithm to aggregate features of the steady-state operating conditions, calculate the time distance decay weight of each steady-state cluster, and finally select the historical timestamp that represents the equipment just completed maintenance and is in its optimal performance baseline state as the preset historical time point. Taking the first microenvironment control body of the air compressor station as an example, the system extracts its maintenance data from the past three years and finds that its operating energy efficiency ratio is most stable on the 7th day after each deep maintenance. Therefore, the current detection time point is pushed back to the 7th day after the most recent maintenance as the preset historical time point, and the steady-state historical data at that moment is accurately extracted from the historical cache database as 84 kg / h. The real-time carbon emission reconstruction data at the current detection time point is 120 kg / h. By subtracting 84 kg / h from 120 kg / h, the positive drift increment is calculated to be 36 kg / h. The physical background of this calculation step is that industrial power equipment, after experiencing mechanical wear and filter clogging, often experiences an increase in equivalent carbon emissions while performing the same amount of work. By introducing historical benchmark data and performing difference calculations, the net increase in carbon emissions during this operating cycle can be directly extracted. This operation filters out the inherent emissions of the equipment and background energy consumption under normal operating conditions, focusing on the state shift of the equipment over time, providing direct physical deviation data for subsequent quantification of the equipment's aging degree.

[0061] In S32, assuming the drift increment is positive, it is divided by the real-time carbon emission reconstruction data at the current detection time to obtain the percentage of individual degradation increment for each microenvironment control body. Combining the data for the first microenvironment control body, its positive drift increment is 36 kg / h, and the current real-time carbon emission reconstruction data is 120 kg / h. Performing a division operation (36 divided by 120) yields a percentage of 0.30 for the individual degradation increment of this microenvironment control body. This calculation logic aims to achieve dimensionless data processing. Directly using absolute drift increments to assess aging risk suffers from inconsistent scales. By dividing it by the total amount of reconstruction data at the current moment, the percentage contributed by recent equipment degradation accumulation is calculated. This method eliminates the interference of overall environmental load fluctuations, reflecting the relative severity of equipment performance degradation within a specific area. This technique allows for horizontal comparison and comprehensive assessment of degradation risk for plant power facilities with different power ratings on the same dimensionless scale.

[0062] In step S33, the degradation increment percentage of each individual microenvironment control body across the entire domain is scanned, and the maximum value is extracted and used as the overall degradation parameter. In a specific air compressor station application scenario, it is assumed that, except for the first control body with a degradation increment percentage of 0.30, the historical data of the other three microenvironment control bodies are basically the same as the current data, and their increment percentages are all set to 0. This data set is traversed, and the maximum value of 0.30 is extracted as the current degradation parameter. The reason for using the maximum value extraction logic in this step is that, as an energy supply network, the reliability of plant operations is often constrained by the equipment with the most severe performance degradation. By locking the value with the highest degradation increment percentage across the entire domain as the representative degradation parameter, the monitoring sensitivity to the hidden chronic degradation risks of the entire plant is ensured. This technique, without the need for manual intervention to set differentiated alarm tolerances for various equipment, utilizes the maximum degradation ratio generated by real-time data evolution to construct a unified macro-control index reflecting the long-term physical aging state of the plant's equipment group.

[0063] like Figure 5 As shown, in one specific embodiment, S4 includes: S41, multiplying the carbon emission baseline allowance by the overall production load rate to obtain a reasonable emission baseline constrained by the production load; S42. Calculate the sum of squares of the disorder penalty parameter and the decay parameter, and calculate the root mean square value based on the sum of squares to obtain the overall risk penalty weight. S43. Subtract the overall risk penalty weight from the constant to obtain the safe contraction ratio, and multiply the reasonable emission baseline by the safe contraction ratio to generate an adaptive safety envelope. S44. Compare the real-time carbon emission reconstruction data of each micro-environment control body with the adaptive safety envelope. When the real-time carbon emission reconstruction data of a certain micro-environment control body is greater than the adaptive safety envelope, trigger an alarm and lock the physical location corresponding to the micro-environment control body. S45. Compare the values ​​of the disorder penalty parameter and the attenuation parameter at the current moment; if the disorder penalty parameter is greater than the attenuation parameter, output the source tracing result representing the sudden failure of the supporting environment; if the attenuation parameter is greater than the disorder penalty parameter, output the source tracing result representing the physical aging of the core power equipment. Specifically, the specific calculation logic for generating the adaptive safety envelope can be expressed as follows:

[0064] Among them, T env (t): The adaptive safety envelope generated at detection time t; T limit : The carbon emission baseline quota preset by the plant; ρ(t): The comprehensive production load rate extracted at the current detection time point t; C1(t): The disorder penalty parameter at the current detection time point t; C2(t): The attenuation parameter determined at the current detection time point t.

[0065] In this embodiment, it should be noted that in S41, the carbon emission baseline allowance is multiplied by the comprehensive production load rate to obtain a reasonable emission baseline constrained by the production load. Regarding the acquisition and quantification method of the comprehensive production load rate, it is not a manually fixed input value. Instead, the system connects in real-time with the Manufacturing Execution System and the Enterprise Resource Planning System through industrial protocols to extract the number of currently active production lines, real-time material input, and operating power of core process equipment from the plant management system. This data is then calculated by comparing it with the plant's designed maximum full-load capacity parameters using a multiple linear regression model. For example, if the system reads in real-time that 90 core pieces of equipment are currently operating in the workshop and the main transformer's active power has reached 88% of its full-load design, and inputs indicators such as the comprehensive production schedule into the normalized model, the calculated comprehensive production load rate of the plant management system at the current moment is 0.90.

[0066] Meanwhile, the carbon emission baseline allowance set according to management requirements is 150 kg / h. The method for determining this preset carbon emission baseline allowance is as follows: based on the factory's historical monitoring logs over the past three to five years at a medium baseline capacity, Monte Carlo simulation is used to remove outliers caused by extreme weather or maintenance shutdowns, calculating the average carbon emissions within the standard confidence interval. This average is then adjusted downwards based on the company's carbon reduction constraints for the current year. For example, historical data from the past three years with a load rate between 80% and 85% is retrieved. After outlier removal, the average carbon emissions per unit time are calculated to be 158 kg / h. Considering the company's requirement to lower the emission reduction target by approximately 5% this year, after rounding, the allowance is finally strictly established as 150 kg / h. A multiplication operation is then performed, multiplying 150 kg / h by 0.90, to calculate a reasonable emission baseline of 135 kg / h.

[0067] The underlying principle of this calculation is that carbon emissions from industrial plant systems are not static figures; they are positively correlated with the plant's actual capacity planning. Using a static baseline quota for control can lead to underreporting during low-load operation and false alarms during continuous full-load production. By linearly multiplying the absolute quota by the real-time load rate, a flexible monitoring benchmark that fluctuates with production pace is established. This technique ensures the alignment of carbon emission monitoring indicators with the plant's core production tasks, providing a reasonable starting point for subsequent risk envelope trimming that aligns with actual process logic.

[0068] In S42, the sum of squares of the disorder penalty parameter and the decay parameter is calculated, and the root mean square value is calculated based on this sum to obtain the overall risk penalty weight. In the current scenario, the disorder penalty parameter is 0.202, and the decay parameter is 0.30. First, the sum of their squares is calculated, i.e., 0.202 squared (0.0408) plus 0.30 squared (0.09), resulting in 0.1308. Then, the root mean square of this value is calculated, i.e., 0.1308 divided by 2 and then taking the square root, yielding an overall risk penalty weight of approximately 0.2557. The reason for using the root mean square calculation logic is that the disorder penalty parameter and the decay parameter represent two different dimensions of risk: spatial abrupt imbalance and temporal gradual aging. If a simple scalar addition is used for fusion, it is easy to exaggerate the composite risk or cause numerical overflow. By using vector fusion based on the Euclidean distance norm, the superposition effect of the two-dimensional risks can be smoothly and comprehensively evaluated. This technique ensures that the penalty weights after fusion are strictly controlled within the dimensionless convergence interval, providing a mathematical model for the objective normalization measurement of multimodal risk and avoiding model bias caused by direct linear addition.

[0069] In step S43, the overall risk penalty weight is subtracted from a constant to obtain the safe contraction ratio. This safe contraction ratio is then multiplied by the reasonable emission baseline to generate the adaptive safety envelope. Based on the previous calculations, the overall risk penalty weight is 0.2557. Subtracting this value from the constant yields a safe contraction ratio of 0.7443. Subsequently, the reasonable emission baseline constrained by production load (135 kg / h) is multiplied by 0.7443 to calculate the current adaptive safety envelope of 100.48 kg / h. This step's calculation logic constructs an inward-pressuring defense mechanism. When no spatial or temporal risks are detected, the penalty weight approaches zero, and the envelope equals the reasonable emission baseline, ensuring normal production space. However, once a leak or equipment aging risk is detected, the penalty weight increases, and the safe contraction ratio decreases accordingly, thus proportionally narrowing the permissible emission boundary. This technique overcomes the lag of passive interception with fixed thresholds, enabling the monitoring defense line to proactively quantify and contract thresholds based on fluctuations in underlying risk factors.

[0070] In step S44, the real-time carbon emission reconstruction data of each microenvironment control body is compared with the generated adaptive safety envelope. When the data of a control body exceeds the envelope, an alarm is triggered and its corresponding physical location is locked. During the monitoring cycle of this air compressor station, the generated adaptive safety envelope is 100.48 kg / h. By polling the data of each control body, it was found that the real-time carbon emission reconstruction data of the first microenvironment control body reached 120 kg / h. Through numerical comparison, 120 kg / h is greater than the envelope boundary value of 100.48 kg / h, thereby triggering an over-limit alarm. Based on the spatial coordinates of the microenvironment control body, the No. 1 air compressor unit under its jurisdiction and the surrounding pipeline network are locked. This step performs the final anomaly interception judgment. Since the envelope is a dynamic boundary generated after load constraints and risk trimming, this comparison mechanism can not only intercept obvious faults that seriously exceed the limit, but also capture hidden defects that have deviated from the healthy state but have not exceeded the standard under the conventional static threshold. This technology enables a closed-loop process from data reconstruction and risk quantification to over-limit determination, providing highly timely fault location capabilities.

[0071] In S45, based on the comparison results of real-time carbon emission reconstruction data of each microenvironment control body with the adaptive safety envelope, and the numerical comparison results of the disorder penalty parameter and the attenuation parameter, a source location result including fault attribute determination is generated. The decision-making basis of this step is established on the calculation foundation of the preceding adaptive safety envelope, namely… .

[0072] In existing carbon emission management, the alarm threshold is usually a fixed scalar value. When companies increase production load to meet delivery deadlines, the natural increase in carbon emissions will frequently exceed the fixed threshold, causing false alarms. At the same time, when complex risks such as spatial dispersion or temporal degradation are detected, if a simple arithmetic addition is used to lower the threshold, the direct addition of the two risk parameters can easily lead to excessive penalties, or even cause the threshold to shrink into a negative number, which lacks mathematical rigor.

[0073] To address the aforementioned technical issues, the safety envelope expression in this method first incorporates the plant-mandated carbon emission baseline allowance T. limit The emission baseline, constrained by production load, is obtained by linearly multiplying it with the current comprehensive production load rate ρ(t). This multiplication operation establishes a flexible benchmark for emissions that fluctuate with production capacity, logically enabling the monitoring system to adapt to production capacity.

[0074] Subsequently, the expression adopts the Euclidean distance norm, i.e. A multidimensional risk vector coupling is performed between the disorder penalty parameter C1(t) and the decay parameter C2(t) characterizing aging. Since both C1 and C2 are dimensionless parameters that fluctuate between 0 and 1 after normalization, the operation of summing the squares and then taking the root mean square not only achieves the orthogonal fusion of different physical dimensions, but also ensures that the calculated overall risk penalty weight is always constrained within the safe closed interval of 0 to 1.

[0075] Furthermore, the expression then subtracts the penalty weight from the constant to obtain the safe contraction ratio, and multiplies it by the reasonable emission baseline, thereby dynamically and proportionally reducing the allowable emission boundary downwards based on the underlying two-dimensional risk index.

[0076] Substituting specific scenario data for calculation: Assuming the statutory carbon emission baseline allowance T limit Given a current production load factor of ρ(t) of 0.90, the reasonable emission baseline constrained by the load is 150 × 0.90 = 135 kg / h. The disorder penalty parameter C1(t) obtained in the previous steps is 0.202, and the monomer degradation attenuation parameter C2(t) extracted based on historical data is 0.30. The sum of their squares is calculated as: 0.202. 2 +0.30 2 =0.0408 + 0.09 = 0.1308; Calculating the root mean square yields the overall risk penalty weight as follows: Subtracting this weight from the constant yields a safe shrinkage ratio of 1 - 0.2557 = 0.7443. The final generated adaptive safe envelope T... env (t) = 135 kg / h × 0.7443 = 100.48 kg / h. Since the reconstructed data of the first microenvironment control body, 120 kg / h, is numerically greater than the current envelope after nonlinear trimming and compression, 100.48 kg / h, an abnormal alarm is immediately triggered. More importantly, S45 then executes the core source tracing decision logic: comparing the two current driving parameters, it is found that the decay parameter C2 (t) (0.30) representing time-progressive aging is numerically greater than the disorder penalty parameter C1 (t) (0.202) representing sudden spatial imbalance.

[0077] Based on this numerical comparison, the underlying logic determines that the primary driver of this carbon emission exceeding the safety threshold stems from long-term, chronic degradation of the equipment, rather than a sudden pipe burst or air conditioning malfunction. Therefore, the final output identifies the root cause of the physical aging of core power equipment (such as long-term wear of the air compressor rotor leading to reduced energy efficiency) and assigns targeted equipment maintenance work orders. This logical judgment process, by separating and comparing the early warning threshold calculation with multi-dimensional characteristic parameters, achieves decoupled identification of complex and cross-functional fault modes, effectively guiding troubleshooting in industrial settings.

[0078] A dynamic monitoring and traceability system for carbon emissions in a plant system is also provided, the system including: The spatial reconstruction module is used to discretize the space within the plant system into multiple micro-environment control bodies based on the plant's physical boundaries; it acquires the real-time raw carbon emission data collected by the monitoring nodes and the real-time HVAC airflow vector of the plant system; it performs asymmetric directed spatial interpolation mapping on the real-time raw carbon emission data based on the real-time HVAC airflow vector to acquire the real-time carbon emission reconstruction data of each micro-environment control body. The disorder assessment module is used to obtain the spatial distribution probability of each microenvironment control body in the whole domain based on the real-time carbon emission reconstruction data of each microenvironment control body, and obtain the disorder penalty parameter characterizing the spatial imbalance state of the system based on the spatial distribution probability and its logarithmic operation result. The attenuation analysis module is used to compare the real-time carbon emission reconstruction data of each microenvironment control body with its corresponding historical data to extract the percentage of individual degradation increment, and obtain attenuation parameters characterizing equipment aging based on the percentage of individual degradation increment of each microenvironment control body in the whole domain. The dynamic envelope decision module is used to obtain the overall production load rate and carbon emission baseline quota of the plant system. Based on the constraint of the overall production load rate, the disorder penalty parameter and the attenuation parameter are coupled by risk vector to generate an adaptive safety envelope. The real-time carbon emission reconstruction data of each micro-environment control body are compared with the adaptive safety envelope, and the source location result is generated based on the comparison result and the numerical comparison result of the disorder penalty parameter and the attenuation parameter.

[0079] In one specific implementation, the dynamic envelope decision module is specifically used to multiply the carbon emission baseline quota by the comprehensive production load rate to obtain a reasonable emission baseline constrained by the production load, calculate the sum of squares of the disorder penalty parameter and the attenuation parameter and obtain the root mean square value to obtain the overall risk penalty weight, subtract the overall risk penalty weight from the constant to obtain the safe contraction ratio, and multiply the reasonable emission baseline by the safe contraction ratio to generate an adaptive safe envelope.

[0080] To enable those skilled in the art to fully understand and implement the technical solutions described in this specification, the following section, in conjunction with a specific application scenario, provides a detailed deduction and data analysis of the entire process of a method and system for dynamic monitoring and tracing of carbon emissions in a plant system.

[0081] In a scenario of a power air compressor station in a large semiconductor manufacturing plant, the air compressor station is responsible for providing clean compressed air to the cleanroom. Its internal physical space is discretized into four micro-environment control volumes. Two high-precision carbon emission equivalent physical monitoring nodes are deployed within the station. When the system enters S1, the environmental management and control system extracts the rated maximum wind speed of the current plant's HVAC main duct as 10 m / s, while the real-time wind speed scalar is 6 m / s. At a certain detection moment, the raw carbon emission reading for node 1 is 125 kg / h, and the reading for node 2 is 45 kg / h. Taking the first micro-environment control volume as an example, its distance from node 1 is 10 m, and its distance from node 2 is 20 m. Since node 1 is located upwind, the angle between the airflow vector and the position vector is 0 degrees, and the cosine value is 1; node 2 is located downwind on the return air side, with an angle of 180 degrees and a cosine value of -1. Substituting these values ​​into the asymmetric directed space interpolation mapping expression... The calculation was performed. The effective aerodynamic distance of node 1 was reduced to 10 × (1 - 0.6 × 1) = 4 m, while the aerodynamic distance of node 2, being on the leeward side, remained at 20 m. The corresponding interpolation weights were 0.0625 and 0.0025, respectively, and the final reconstructed data for this control volume was approximately 122 kg / h. This demonstrates from a physical perspective that HVAC airflow pulls upwind monitoring nodes closer, thereby accurately reconstructing the aerodynamic diffusion pattern of pollutants. For ease of subsequent global simulation, it was assumed that the real-time carbon emission reconstruction data for the four microenvironment control volumes after matrix operations were 120 kg / h, 20 kg / h, 30 kg / h, and 30 kg / h, respectively, with a total carbon emission of 200 kg / h for the entire area.

[0082] After the data flow reconstruction is complete, the system enters S2, beginning to quantify the spatial imbalance state of the plant management system based on information entropy theory. The system first calculates the spatial distribution probability of each control entity in the entire domain, with corresponding probability matrices of 0.60, 0.10, 0.15, and 0.15. Then, the probability values ​​are substituted into the disorder penalty expression. The system performs calculations. It sums the logarithmic products of the discrete terms one by one: 0.60×(-0.5108)+0.10×(-2.3026)+2×(0.15×(-1.8971)), calculating that the spatial distribution of the negative discrete terms is approximately -1.106. The maximum theoretical logarithmic value corresponding to the four control volumes is 1.386, and dividing the two yields a dispersion ratio of -0.798. Finally, this is added to a constant to obtain the current disorder penalty parameter of 0.202. From the perspective of practical application, this value accurately reflects that the carbon emissions within the air compressor station exhibit a certain degree of local aggregation rather than uniform distribution, which the system quantifies as a spatial imbalance risk of 20.2%.

[0083] Next, the system enters S3 to perform a stripping analysis of the long-term health status of the equipment. This step retrieves historical baseline data from 60 days ago for each microenvironment control unit. Taking the first microenvironment control unit as an example, its historical carbon emission data was 84 kg / h, while the current reconstructed data is 120 kg / h. Subtracting the two yields a positive drift increment of 36 kg / h. The system divides this increment by the current reconstructed data of 120 kg / h, obtaining a single-unit degradation increment percentage of 0.30. Assuming that the historical data of the other three control units are the same as or decrease from the current data, the increment percentage is 0. After scanning the entire domain, the system extracts the maximum value of 0.30 as the degradation parameter characterizing equipment aging. The core principle of this step is to adaptively calculate and extract the 30% absolute aging loss percentage accumulated by the equipment over time, based on the historical operating footprint of the plant system itself, without relying on manual experience settings.

[0084] Subsequently, the control center enters S4, where it integrates multi-dimensional risk parameters with the factory's actual production plan to make a decision. The system connects with the manufacturing execution system, extracting the current overall production load rate of the entire plant at 0.90, indicating that the air compressor station is operating at high capacity. Simultaneously, it reads the baseline carbon emission allowance for environmental compliance as 150 kg / h. The system first multiplies the baseline allowance by the production load rate, establishing a reasonable emission baseline constrained by the production load at 135 kg / h. This also means that under ideal conditions without any anomalies, the system will allow emissions to rise moderately in line with high production loads.

[0085] However, since the preceding steps had detected underlying operational risks, the system began executing the nonlinear safety envelope expression. The system performs a safety margin trimming process. It extracts the disorder penalty parameter (0.202) and the attenuation parameter (0.30), calculating the root mean square (RMS) of their sum of squares to be 0.2557. This RMS value represents the overall risk penalty weight after incorporating spatial imbalance and time aging. The system subtracts this weight from a constant to obtain a safe contraction ratio of 0.7443, which is then multiplied by the reasonable emission baseline of 135 kg / h, ultimately generating a strictly downward-pressured adaptive safety envelope of 100.48 kg / h.

[0086] In the final logic comparison and source tracing diagnosis phase, the system discovered that the real-time reconstructed data of 120 kg / h for the first microenvironment control unit was greater than the adaptive safety envelope of 100.48 kg / h after dynamic pressure reduction. The system immediately triggered a hardware alarm and locked the core air compressor unit under that control unit. During root cause analysis, the decision matrix compared the values ​​of two core risk parameters: since the attenuation parameter of 0.30 was significantly greater than the disorder penalty parameter of 0.202, the system determined that the dominant cause of this carbon emission exceedance was not a sudden pipeline rupture or leak, but rather irreversible physical wear and tear on the core power equipment under long-term high load (such as severe energy efficiency degradation caused by increased rotor clearance). The system then automatically generated a targeted work order characterizing equipment aging, thus completing the entire technical closed loop from data capture and nonlinear intervention to fault attribute determination in the industrial field.

[0087] The preferred embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the specific details of the above embodiments. Within the scope of the technical concept of the present invention, various simple modifications can be made to the technical solution of the present invention, and these simple modifications all fall within the protection scope of the present invention.

[0088] It should also be noted that the various specific technical features described in the above embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, the present invention will not describe the various possible combinations separately.

[0089] Furthermore, various different embodiments of the present invention can be combined in any way, as long as they do not violate the spirit of the present invention, they should also be regarded as the content disclosed by the present invention.

[0090] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A method for dynamic monitoring and source tracing of carbon emissions in a plant system, characterized in that the method... include: Based on the physical boundary of the plant, the plant space is discretized to obtain multiple micro-environment control volumes; Based on real-time HVAC airflow vectors, asymmetric directed spatial interpolation mapping is performed on the raw real-time carbon emission data collected by monitoring nodes to obtain real-time carbon emission reconstructed data for each microenvironment control body. The spatial distribution probability is obtained from real-time carbon emission reconstruction data; the disorder penalty parameter is obtained from the spatial distribution probability and its logarithmic operation result; the percentage of individual cell degradation increment is obtained from real-time carbon emission reconstruction data and corresponding historical data; and the attenuation parameter is obtained from the percentage of degradation increment of each individual cell in the entire domain. A reasonable emission baseline is obtained based on the carbon emission baseline quota and the overall production load rate; the sum of squares is obtained based on the disorder penalty parameter and the attenuation parameter; and the root mean square value is obtained based on the sum of squares. The safe shrinkage ratio is obtained by subtracting the root mean square value from the constant. An adaptive safety envelope is generated based on a reasonable emission baseline and a safe shrinkage ratio; When the real-time carbon emission reconstruction data is greater than the adaptive safety envelope, the corresponding physical location is locked; the source tracing result is generated based on the numerical comparison of the disorder penalty parameter and the attenuation parameter.

2. The method for dynamic monitoring and source tracing of carbon emissions in a plant system according to claim 1, characterized in that, Based on real-time HVAC airflow vectors, asymmetric directed spatial interpolation mapping is performed on the raw real-time carbon emission data collected by monitoring nodes, including: Determine the position vector pointing from the original data acquisition location to the geometric center of the microenvironment control volume; obtain the spatial angle based on the position vector and the real-time HVAC airflow vector; obtain the downwind influence factor based on the spatial angle. The wind speed percentage is obtained from the real-time wind speed scalar and the rated maximum wind speed scalar; the distance reduction factor is obtained from the wind speed percentage and the downwind influence factor; and the aerodynamic effective distance is obtained from the distance reduction factor and the straight-line distance between the original data acquisition location and the microenvironment control volume.

3. The method for dynamic monitoring and source tracing of carbon emissions in a plant system according to claim 2, characterized in that, Asymmetric directed space interpolation mappings include: Perform cosine calculation on the spatial angle to obtain the cosine value; take the maximum value between the cosine value and the constant zero to obtain the downwind influence factor; divide the real-time wind speed scalar by the rated maximum wind speed scalar to obtain the wind speed ratio; Subtract the product of wind speed percentage and downwind influence factor from constant 1 to obtain the distance reduction factor; multiply the straight-line distance by the distance reduction factor to obtain the aerodynamic effective distance; calculate the negative square of the aerodynamic effective distance to obtain the distance attenuation weight. The weighted emissions are obtained by multiplying the raw real-time carbon emission data of each monitoring node by the corresponding distance attenuation weight and summing the results; the weighted emissions are then summed by summing the distance attenuation weights of each monitoring node to obtain the weighted sum; and the reconstructed real-time carbon emission data is obtained by dividing the weighted emissions by the weighted sum.

4. The method for dynamic monitoring and source tracing of carbon emissions in a plant system according to claim 1, characterized in that, The disorder penalty parameters are obtained based on the spatial distribution probability and its logarithmic operation, including: Perform natural logarithm calculations on the probabilities of each spatial distribution to obtain the natural logarithm results; based on the probabilities of each spatial distribution and the corresponding natural logarithm results, obtain the spatial distribution discrete terms with negative values; The dispersion ratio is obtained by calculating the natural logarithm of the spatial distribution discrete terms and the total number of micro-environment control volumes; the disorder penalty parameter is obtained by adding the constant to the dispersion ratio.

5. The method for dynamic monitoring and source tracing of carbon emissions in a plant system according to claim 1, characterized in that, The percentage of incremental degradation of individual cells is obtained based on real-time carbon emission reconstructed data and corresponding historical data, including: Subtract the historical data at the preset historical time point from the real-time carbon emission reconstruction data at the current detection time point to obtain the drift increment; If the drift increment is positive, divide the drift increment by the real-time carbon emission reconstruction data at the current detection time point to obtain the percentage of single-unit degradation increment; if the drift increment does not exceed zero, set the percentage of single-unit degradation increment to zero.

6. The method for dynamic monitoring and source tracing of carbon emissions in a plant system according to claim 1, characterized in that, The attenuation parameters are obtained based on the percentage of degradation increment for each individual unit across the entire domain, including: Perform maximum value calculation on the percentage of individual degradation increment of all microenvironment control bodies in the entire domain to obtain the attenuation parameter.

7. The method for dynamic monitoring and source tracing of carbon emissions in a plant system according to claim 1, characterized in that, A reasonable emission baseline is obtained based on carbon emission baseline allowances and overall production load rate, including: The carbon emission baseline allowance is multiplied by the overall production load rate to obtain a reasonable emission baseline constrained by the production load.

8. The method for dynamic monitoring and source tracing of carbon emissions in a plant system according to claim 1, characterized in that, Generate an adaptive safe envelope, including: The disorder penalty parameter and the decay parameter are squared and summed to obtain the sum of squares; the sum of squares is divided by the constant two and then the square root is obtained to obtain the root mean square value. Subtract the root mean square value from the constant to obtain the safe shrinkage ratio; multiply the reasonable emission baseline by the safe shrinkage ratio to generate the adaptive safety envelope.

9. The method for dynamic monitoring and source tracing of carbon emissions in a plant system according to claim 1, characterized in that, The corresponding physical location is located, and source tracing results are generated based on a comparison of the disorder penalty parameter and the decay parameter, including: When the real-time carbon emission reconstruction data of any micro-environment control body is greater than the adaptive safety envelope, the physical location corresponding to that micro-environment control body is locked. If the disorder penalty parameter is greater than the attenuation parameter, the output will be the source tracing result representing the sudden failure of the supporting environment; if the attenuation parameter is greater than the disorder penalty parameter, the output will be the source tracing result representing the physical aging of the core power equipment.

10. A dynamic monitoring and traceability system for carbon emissions in a plant system, characterized in that, The system is used to implement the dynamic monitoring and source tracing method for carbon emissions in a plant system as described in any one of claims 1 to 9, and the system includes: The spatial reconstruction module is used to acquire real-time carbon emission reconstruction data; the disorder assessment module is used to acquire disorder penalty parameters; and the decay analysis module is used to acquire decay parameters. The dynamic envelope decision module is used to generate an adaptive safety envelope; the dynamic envelope decision module is also used to generate source tracing and localization results.

Images