Insulating gas recovery carbon emission data processing method and system

Abstract

The application discloses an insulating gas recovery carbon emission data processing method and system, relates to the technical field of greenhouse gas emission data processing of electric power equipment, and comprises a data acquisition module, a dynamic volume inversion module, a double-path mass fusion module, a reliability calculation module, a CO2 equivalent value calculation module, a leakage diagnosis and grade determination module and a carbon data generation and transmission module. The application establishes a multi-stage feedback relationship among a joint state observer, a volume dynamic inversion model, a double-path mass reliable fusion model, a time-varying reliability recursive model, an abnormal leakage diagnosis model and a historical error database, takes the deviation between the theoretical mass and the real-time recovery mass as a target function, and adopts a self-adaptive iterative optimization mode to correct the gas chamber volume in real time, so that continuous closed-loop data processing of the whole insulating gas recovery process is realized. Therefore, the volume estimation, mass fusion and carbon emission calculation are adaptively corrected according to the real-time operation state, and the carbon emission calculation precision is improved.

Description

Technical Field

[0001] This invention relates to the field of greenhouse gas emission data processing technology for power equipment, and specifically to a method and system for processing carbon emission data from insulation gas recovery. Background Technology

[0002] With the continuous advancement of the national "dual-carbon" strategic goals, the power industry is gradually establishing a carbon emission monitoring and management system covering the entire lifecycle of equipment. Insulating gases, especially SF6 (sulfur hexafluoride), are widely used in high-voltage electrical equipment such as GIS switchgear, circuit breakers, and gas-insulated switchgear due to their excellent insulation and arc-quenching properties. However, SF6 is a typical greenhouse gas with high global warming potential; any leakage will have a significant impact on the greenhouse effect. Therefore, accurately measuring the recovery, leakage, and carbon emissions of insulating gases during equipment overhaul, decommissioning, and operation and maintenance has become a crucial aspect of carbon emission management in the power industry.

[0003] In existing technologies, the calculation of the recovered gas mass either relies on the weighing method (mass difference before and after the recovery tank) or the flow integral method (cumulative flow rate of the pipeline flow meter). The weighing method is affected by factors such as the accuracy of the weighing sensor, environmental vibration, and changes in the weight of the recovery tank; the flow integral method is affected by factors such as zero-point drift of the flow meter and inaccurate temperature and pressure compensation. Both methods have systematic errors and lack effective error identification and correction mechanisms.

[0004] Furthermore, existing insulating gas recovery systems employ a one-way data processing flow, namely, data acquisition, quality calculation, and emission output. Each module operates independently, without considering the correlation between data and the dynamic reliability. They lack bidirectional feedback between state estimation, quality fusion, volume correction, and emission calculation, and cannot correct the calculation model in real time based on the operating status. This leads to the continuous accumulation of errors during the long-term operation of the recovery system, affecting the final emission calculation results. Summary of the Invention

[0005] To address the problems of existing technologies, this invention provides a method for processing carbon emission data from insulating gas recovery, the method comprising the following steps: S1: Collect multi-source operational data throughout the entire process of insulating gas recovery and establish a multi-source state dataset under a unified time reference. S2: Based on the multi-source state dataset, establish a multi-physics coupled state-space model of the gas chamber, construct a joint state observer, and simultaneously recursively estimate the gas chamber volume, temperature, pressure, and gas mass to obtain predicted values; based on the predicted values, establish a dynamic volume inversion model, using the deviation between the theoretical mass and the real-time recovered mass as the objective function, and adopt an adaptive iterative optimization method to correct the gas chamber volume in real time, forming a continuous closed-loop self-correction mechanism for the volume. S3: Calculate the weighing recovery mass and flow recovery mass based on the corrected air chamber volume, establish a dual-path reliable mass fusion model and output the fusion residual; then feed the fusion residual back to the volume dynamic inversion model as the correction term of the objective function in the next cycle, so as to realize the bidirectional coupling between mass fusion and volume inversion. S4: Establish a time-varying reliability recursive model, using the state residual, volume inversion error, fusion residual, pressure stability, temperature stability, weighing stability, flow stability and time synchronization error output by the joint state observer as reliability observation variables, recursively calculate the system reliability, obtain the real-time reliability index and output the reliability change rate. S5: Based on the corrected actual gas chamber volume, fusion mass value and credibility index, establish an insulating gas mass inversion model, calculate the real-time insulating gas emissions, and combine the global warming potential of insulating gas to calculate the real-time CO2 equivalent value. The credibility index is used to dynamically compensate for emission calculation errors. S6: Establish a theoretical recovery quality prediction model and perform deviation analysis between the theoretical recovery quality and the fusion quality; when the deviation exceeds the preset threshold, initiate abnormal leakage diagnosis; determine the leakage type and leakage level based on the abnormal leakage diagnosis; feed the leakage type and leakage level back to the credibility recursive model to correct the credibility index; S7: Generate carbon emission records based on actual gas chamber volume, fusion quality, credibility index, CO2 equivalent value and leak diagnosis results. At the same time, construct carbon asset data including recovery quality, leak status, credibility level and carbon emission amount, and upload it to the carbon emission management platform.

[0006] A further approach is to establish the following state vector using the joint state observer described in S2: X =[ V , T , P , m ]; in, V In the state of air chamber volume, T This refers to the temperature state of the air chamber. P This refers to the pressure state of the air chamber. m It is in gaseous mass state; The joint state observer performs synchronous recursive estimation of each state variable based on the collected pressure, temperature and recovery quality data, and outputs volume prediction, pressure prediction, temperature prediction and state residual; the state residual serves as the input parameter of the volume dynamic inversion model and the time-varying reliability recursive model.

[0007] A further approach is to construct an objective function based on the deviation between the theoretical gas mass and the fusion mass using the volume dynamic inversion model, and to iteratively correct the gas chamber volume parameters based on the objective function. In each iteration, the state residual output by the joint state observer is introduced into the objective function as a constraint term to dynamically adjust the direction and magnitude of volume correction. When the rate of change of volume is less than a preset threshold for multiple consecutive sampling periods, the current volume is determined as the actual air chamber volume and fed back to the joint state observer as the initial state for the next sampling period.

[0008] A further approach is that the establishment of the dual-path quality-trusted fusion model described in S3 includes: S30: Calculate the degree of local anomaly based on the weighing mass data and the flow integral data respectively; S31: Eliminate abnormal sampling points based on the degree of local anomaly to obtain two effective quality sequences; S32: Calculate the basic fusion weights based on the deviation, stability, and historical correction error between the two effective quality sequences; S33: Based on the state residuals and volumetric dynamic inversion errors output by the joint state observer, the basic fusion weights are corrected in real time to obtain the final fusion weights; S34: Calculate the fusion quality value based on the final fusion weights and output the fusion residual.

[0009] A further approach is that the real-time reliability index mentioned in S4 needs to be fed back to the volume dynamic inversion model, the dual-path quality reliability fusion model, and the carbon emission calculation model simultaneously. Among them: the real-time credibility index is used to adjust the volume dynamic inversion step size, adjust the dual-path quality fusion weight, and correct the carbon emission calculation compensation coefficient; The confidence change rate of the output described in S4 is used to adjust the state estimation gain for the next sampling period, thereby achieving closed-loop adaptive correction for the entire carbon emission calculation process.

[0010] A further approach is that the abnormal leakage diagnosis described in S6 includes: Spectral features were extracted based on infrared absorption spectroscopy during the recycling process; A multi-source leakage feature set was established based on pressure change characteristics, mass change characteristics, volume correction change characteristics, spectral characteristics, and confidence change trajectory. The leakage type and leakage level are determined based on the aforementioned multi-source leakage feature set; Specifically, when the same leakage type is detected in multiple consecutive sampling cycles, the reliability of the corresponding sampling cycle is reduced, the suppression weight of abnormal data in the dual-path quality fusion process is increased, and the volume dynamic inversion is re-executed to update the actual gas chamber volume.

[0011] A further solution is that, after receiving carbon emission records, leak diagnosis results and reliability indicators, the carbon emission management platform described in S7 will statistically analyze historical operating data from each recycling process and establish a historical error database. The historical error database is used to update the objective function parameters in the volume dynamic inversion model, the fusion weight correction parameters in the dual-path quality reliable fusion model, and the recursive parameters in the time-varying reliability recursive model. The updated parameters are then sent to the field data processing system to achieve self-learning optimization of the model parameters.

[0012] The present invention also provides an insulating gas recovery carbon emission data processing system, the processing system comprising: The data acquisition module is used to collect multi-source operational data throughout the entire process of insulating gas recovery and to establish a multi-source status dataset under a unified time reference. The dynamic volume inversion module is used to establish a multi-physics coupled state-space model of the gas chamber based on a multi-source state dataset, construct a joint state observer, and simultaneously recursively estimate the gas chamber volume, temperature, pressure, and gas mass to obtain predicted values. Based on the predicted values, a dynamic volume inversion model is established, using the deviation between the theoretical mass and the real-time recovered mass as the objective function. An adaptive iterative optimization method is used to correct the gas chamber volume in real time, and a continuous closed-loop self-correction mechanism for volume is formed. The dual-path mass fusion module is used to calculate the weighing recovery mass and flow recovery mass based on the corrected gas chamber volume, and to establish a reliable dual-path mass fusion model and output the fusion residual. The fusion residual is then fed back to the volume dynamic inversion model as the correction term of the objective function in the next cycle, realizing the bidirectional coupling between mass fusion and volume inversion. The credibility calculation module is used to establish a time-varying credibility recursive model. It uses the state residual, volume inversion error, fusion residual, pressure stability, temperature stability, weighing stability, flow stability and time synchronization error output by the joint state observer as credibility observation variables to recursively calculate the system credibility, obtain the real-time credibility index and output the credibility change rate. The CO2 equivalent value calculation module is used to establish an insulating gas quality inversion model based on the corrected actual gas chamber volume, fusion mass value and credibility index, calculate the real-time insulating gas emissions, and calculate the real-time CO2 equivalent value in combination with the global warming potential of insulating gas. The Leakage Diagnosis and Level Determination module is used to establish a theoretical recovery quality prediction model and perform deviation analysis between the theoretical recovery quality and the fusion quality. When the deviation exceeds a preset threshold, abnormal leakage diagnosis is initiated. Based on the abnormal leakage diagnosis, the leakage type and leakage level are determined. The leakage type and leakage level are fed back to the credibility recursive model to correct the credibility index. The carbon data generation and transmission module is used to generate carbon emission records based on the actual gas chamber volume, fusion quality, credibility index, CO2 equivalent value and leakage diagnosis results. At the same time, it constructs carbon asset data including recovery quality, leakage status, credibility level and carbon emission amount, and uploads it to the carbon emission management platform.

[0013] The present invention also provides an electronic device, comprising: A processor and a memory for storing a computer program capable of running on the processor, wherein the processor, when running the computer program, implements the steps of the processing method.

[0014] The present invention also provides a computer storage medium storing an executable program, wherein when the executable program is executed by a processor, it implements the steps of the processing method.

[0015] The advantages of this invention compared to the prior art are as follows: This invention establishes a multi-level feedback relationship between a joint state observer, a volume dynamic inversion model, a dual-path quality reliable fusion model, a time-varying reliability recursive model, an abnormal leakage diagnosis model, and a historical error database. Using the deviation between theoretical quality and real-time recovered quality as the objective function, it employs an adaptive iterative optimization method to correct the gas chamber volume in real time, feeding the corrected volume back to the observer as a state constraint. This achieves continuous closed-loop data processing for the entire insulating gas recovery process. The models not only transfer data but also form bidirectional coupling and collaborative correction through state residuals, fusion residuals, reliability indices, and leakage diagnosis results. This enables adaptive correction of volume estimation, quality fusion, and carbon emission calculation based on real-time operating conditions, improving the accuracy of carbon emission calculations. It avoids the cumulative errors caused by the fixed volume assumption in existing technologies, improves the utilization efficiency of multi-source measurement data and the measurement stability under abnormal operating conditions, and effectively enhances the accuracy of insulating gas recovery quality calculations, carbon emission accounting accuracy, and the reliability of carbon asset data. This provides a data processing method with high engineering application value for insulating gas recovery and carbon emission management in the power industry. Attached Figure Description

[0016] Figure 1 A flowchart of a carbon emission data processing method for insulating gas recovery provided in an embodiment of the present invention; Figure 2 A schematic diagram of the structure of an insulating gas recovery carbon emission data system provided in an embodiment of the present invention; Figure 3 A schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present invention; Figure captions: 10-Data acquisition module; 11-Dynamic volume inversion module; 12-Dual-channel quality fusion module; 13-Confidence calculation module; 14-CO2 equivalent value calculation module; 15-Leakage diagnosis and level determination module; 16-Carbon data generation and transmission module; 20-Memory; 21-Processor. Detailed Implementation

[0017] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0018] It should be noted that this embodiment takes the SF6 insulating gas recovery process of GIS equipment as an example to illustrate the carbon emission data processing method for insulating gas recovery of the present invention.

[0019] Example 1 like Figure 1 As shown, an embodiment 1 of the present invention discloses a method for processing carbon emission data from insulating gas recovery. The processing method includes the following steps: S1: Collect multi-source operational data throughout the entire process of insulating gas recovery and establish a multi-source state dataset under a unified time reference. It should be noted that the following data were collected during the SF6 insulating gas recovery process: Gas chamber pressure data: acquired through an absolute pressure sensor installed on the pressure measurement interface of the gas chamber, used to collect real-time gas pressure data inside the gas chamber.

[0020] Gas chamber temperature data: collected by a temperature sensor installed on a dedicated temperature measurement interface of the gas chamber or attached to the outer wall of the gas chamber, used to collect real-time gas temperature data inside the gas chamber.

[0021] Gas flow rate in the recovery pipeline: This is collected by a flow meter connected in series to the gas pipeline at the front end of the recovery cylinder, for real-time collection of gas flow rate in the recovery pipeline.

[0022] The weight of the recycled cylinders is collected by a weighing sensor installed below the support base of the recycled cylinders.

[0023] Insulating gas (SF6) concentration: Collected by an infrared spectral gas concentration sensor, which is installed in the recovery pipeline and the surrounding environment of the equipment to collect the concentration of insulating gas (SF6) in real time.

[0024] A high-precision time synchronization module is also included, employing GPS / BeiDou dual-mode time synchronization or the IEEE 1588 Precision Time Protocol (PTP) to ensure that the time synchronization accuracy of the data acquired by the aforementioned sensors is better than 1ms. Furthermore, all acquisition devices use a uniform sampling period. Synchronous sampling is performed. Multi-source state data are stored in a real-time database according to time series, forming a multi-source state dataset with a unified time reference. After data acquisition is completed, the joint state observation phase begins.

[0025] S2: Based on the multi-source state dataset, establish a multi-physics coupled state-space model of the gas chamber, construct a joint state observer, and simultaneously recursively estimate the gas chamber volume, temperature, pressure, and gas mass to obtain predicted values; based on the predicted values, establish a dynamic volume inversion model, using the deviation between the theoretical mass and the real-time recovered mass as the objective function, and adopt an adaptive iterative optimization method to correct the gas chamber volume in real time, forming a continuous closed-loop self-correction mechanism for the volume. It should be noted that, in order to avoid the continuous accumulation of quality errors caused by traditional fixed-volume calculations, this embodiment establishes a joint state observer, specifically as follows: First, establish the first i The system state vector for each sampling period. X ( i ),Right now: X ( i )=[ V ( i ), T ( i ), P ( i ), m ( i )]; in, V ( i ) is the first i The actual volume of the air chamber in each sampling period. T ( i ) is the first i Average temperature of the air chamber over one sampling period P ( i ) is the first i Average pressure of the air chamber over one sampling period m ( i ) is the first i Mass of insulating gas per sampling period; The state transition model is represented as: X ( i +1)= f ( X ( i ),U ( i ))+ w ( i ); in, X ( +1) indicates the first i+ The system state vector for one sampling period; f (·) represents the state transition function; U ( i ) indicates the first i The system input vector for each sampling period is used to describe the external inputs or control quantities that affect the changes in the state of the gas chamber, such as gas supply / leakage, ambient temperature, valve status, pressure regulation status, etc. w ( i ) indicates the first i System noise per sampling period; The observation model is: ; in, Z ( ) indicates the first The observations, such as pressure, temperature, weight, flow rate, and concentration, are collected in each sampling cycle. h (·) represents the observation function; v ( ) indicates observation noise.

[0026] The joint state observer uses current sampled data and the state prediction value from the previous period to perform synchronous recursive estimation of state variables. Joint state observation can be achieved using extended Kalman filtering, unscented Kalman filtering, particle filtering, or other nonlinear state estimation algorithms.

[0027] This embodiment uses a joint state observer to update the state, resulting in: Predicted volume Temperature forecast Predicted pressure value Predicted quality value And obtain the updated state vector ,Right now: ; Simultaneously output the state residual e x ( ),Right now: ; in, Z ( i This indicates the collected measurement data such as pressure, temperature, weight, flow rate, and concentration. HThis represents the observation matrix. The state residuals reflect the degree of agreement between the current predicted state and the actual measured values.

[0028] The state residuals are then input into the volumetric dynamic inversion model and the credibility recursive model.

[0029] This embodiment does not calculate pressure, temperature and mass independently, but establishes a multi-physics coupling model of the air chamber, specifically including pressure field, temperature field, volume field and mass field.

[0030] According to the equation of state for insulating gases: PV = ZnRT; in, P This represents the observed pressure value in the air chamber. V This represents the current estimated volume of the air chamber. Z Indicates the compression factor. R Represents the gas constant. T This represents the observed temperature of the air chamber. n Indicates the number of moles of gas.

[0031] Combined with the mass conservation equation Establish a joint state-space model.

[0032] in, m For the quality of the insulating gas in the gas chamber, dm / dt Q is the rate of change of the mass of insulating gas in the gas chamber per unit time. in Q represents the mass of insulating gas entering the gas chamber per unit time. out Q represents the mass of insulating gas entering the recovery equipment from the gas chamber per unit time. leak The mass of the leaked insulating gas.

[0033] Since volume, pressure, temperature, and mass are all involved in state estimation, a change in each state variable will affect the others. Compared to existing methods that calculate mass separately based on pressure, this invention provides a multi-physics joint state estimation.

[0034] This invention further establishes a volumetric dynamic inversion model. The volumetric dynamic inversion model constructs an objective function based on the deviation between the theoretical gas mass and the real-time recovered mass.

[0035] Theoretical gas mass m th The equation of state for insulating gases is derived by back-calculation based on the current volume estimate, pressure observation, and temperature observation: That is: ; in Z is the compressibility factor of SF6 gas, which is a function of pressure and temperature.

[0036] In addition, the real-time recovery quality is obtained by combining weighing and flow rates.m real The objective function is then: ; in, m th For the theoretical gas mass, m real To ensure real-time quality recovery.

[0037] To avoid deviations in the volume correction direction due to state estimation errors, this embodiment further introduces the state residuals output by the joint state observer as a constraint term. The objective function is then corrected as follows: ; in, a 1 represents the quality error weighting coefficient. a 2 represents the weighting coefficient of the state residual. e x For state residuals, m th For the theoretical gas mass, m real To ensure real-time quality recovery.

[0038] The volume is then updated using an adaptive iterative approach, namely: ; in, V ( i+ 1) is the first i+ The estimated actual volume of the air chamber after one sampling period, i.e. the volume parameter after one iteration correction; V ( i ) is the first i The estimated air chamber volume at the current moment of each sampling period is the volume state value obtained in the previous period; η is the adaptive iteration step size (learning rate), which is used to adjust the volume correction speed, and its value determines the adjustment range of the volume parameters in each iteration. The corrected objective function with respect to the air chamber volume variable V The gradient, i.e., the sensitivity of the corrected objective function to changes in volume, is used to determine the direction of volume correction.

[0039] In this embodiment, the step size is not a fixed value, but is automatically adjusted based on the current state residual; when the state residual is small, the correction speed is increased; when the state residual is large, the correction speed is decreased. This improves the stability of volume correction. After correction, the result is... V realThe corrected volume is then fed back to the joint state observer as the initial value for the state prediction in the next sampling period. This forms a continuous closed loop of state estimation, volume inversion, and state estimation. Through this closed loop, this embodiment can continuously correct the actual volume of the gas chamber, thereby effectively reducing systematic errors caused by equipment aging, structural deformation, and changes in sealing performance during long-term operation.

[0040] S3: Calculate the weighing recovery mass and flow recovery mass based on the corrected air chamber volume, establish a dual-path reliable mass fusion model and output the fusion residual; then feed the fusion residual back to the volume dynamic inversion model as the correction term of the objective function in the next cycle, so as to realize the bidirectional coupling between mass fusion and volume inversion. It should be noted that, in order to reduce the measurement error caused by factors such as sensor drift, local abnormal sampling, equipment vibration and environmental changes caused by a single measurement method, this embodiment establishes a dual-path quality reliable fusion model to dynamically fuse the re-recovery quality and the flow recovery quality.

[0041] Weighing and recycling mass calculation The mass of the recycled cylinders is calculated based on the change in their weight. The original weight of the recycled cylinders was... W (Before recycling) The weight of the recycled cylinders after recycling is W (After recovery), the theoretical value of the recovered gas mass is: m ( C )= W (Before recycling) W (After recycling); Considering the impact of changes in gas chamber volume during the recycling process on the weighing of the recycled cylinder (such as changes in buoyancy caused by changes in gas temperature inside the cylinder), a volume correction is required to obtain the corrected weighing and recycling mass. m '( C ),Right now: ; in β Δ is the preset correction factor. V This represents the change in volume. V This refers to the volume of the air chamber.

[0042] Flow recovery quality calculation.

[0043] Flow recovery quality is calculated based on the cumulative flow of the flow meter. Let a sampling period Δ... T Internal flow meter measurement value q ( t ),but i The cumulative recovered mass within each sampling period is: ; Considering the effects of temperature and pressure on gas density, temperature and pressure compensation is required to obtain the corrected flow recovery mass. m '( L ),Right now: ; in, P r , T r The pressure and temperature are for reference conditions.

[0044] based on m '( C )and m '( L Establishing a dual-path quality-assured fusion model includes: S30: Calculate the degree of local anomaly based on the weighing mass data and the flow integral data respectively; Weighing mass sequence ; Flow quality sequence ; in i For the first i One sampling period; For each of the two sequences, calculate the local rate of change for each sequence, i.e.: The local rate of change of the weighing mass sequence is: ; The local variation rate of the flow quality sequence is: ; The degree of local anomaly in the weighing mass series and flow rate mass series is calculated based on the changing trends over multiple consecutive sampling periods. The degree of local anomaly can be calculated by comprehensively considering the following indicators: deviation of rate of change, degree of continuous fluctuation, degree of historical drift, and consistency of state observation residuals. For example, for the weighing mass sequence described above, sampling is performed every 0.5 seconds, and the most recent value is taken. i= Ten sampling periods are used, meaning the data from the most recent 5 seconds is used to determine if the current sampling point is abnormal. For the current sampling point, the degree of deviation from the rate of change is used as a metric. A t Degree of continuous fluctuation B t Degree of historical drift C t Consistency of state observation residuals D t The degree of local anomaly in the weighing sequence was obtained. L ( C ) t ,Right now: ; in, c 1. c 2. c 3. c 4 correspond to the degree of deviation of the rate of change. A t Degree of continuous fluctuation B t Degree of historical drift C t Consistency of state observation residuals D t The weighting coefficients, and c 1+ c 2+ c 3+ c 4=1 If the degree of local abnormality L ( C ) t If the value exceeds a preset threshold, the sampling point is determined to be an abnormal sampling point and is removed. All abnormal sampling points in the weighing mass sequence are then removed using the method described above.

[0045] Similarly, the above method is used to remove all abnormal sampling points from the aforementioned flow quality sequence.

[0046] S31: Eliminate abnormal sampling points based on the degree of local anomaly to obtain two effective quality sequences; After removing all outlier sampling points, the two valid quality sequences obtained are: Effective weighing mass sequence W'm '( C and effective flow quality sequence W'm '( L ); For continuous anomalous data, compensation can be achieved using neighborhood interpolation, trend prediction, or state observation prediction.

[0047] S32: Calculate the basic fusion weights based on the deviation, stability, and historical correction error between the two effective quality sequences; Let the variance of the weighing path be... The variance of the traffic path is Then the weighted path-based fusion weight w( C ) and traffic path basic fusion weight w( L They are respectively: ; ; Among them, w( C )+w( L=1, and the higher the weighing stability, the greater the corresponding weight; the higher the traffic stability, the greater the corresponding weight.

[0048] S33: Based on the state residuals output by the joint state observer e x and volume dynamic inversion error e v Symmetric re-path basic fusion weight w( C ) and traffic path basic fusion weight w( L Real-time corrections are performed separately to obtain the final weighing path base fusion weight w. ' ( C ) and traffic path-based fusion weight w ' ( L ),Right now: ; ; in, α , β These are the preset correction coefficients. P ref This is a reference value for air pressure. V ref This is the volume reference value. When the state residual increases, it indicates that the observer estimation accuracy decreases, and the fusion weight of the corresponding measurement path is reduced accordingly.

[0049] S34: Calculate the fusion quality value based on the final fusion weights and output the fusion residual. Specifically: Calculate the fusion quality value based on the final fusion weights: ; in, m' ( C ) represents the corrected weighing recovery mass, w ' ( C The final weighing path is based on the fusion weights. m' ( L ) represents the corrected flow recovery quality, w ' ( L ) is the basic fusion weight for traffic paths.

[0050] Output fusion residual e f And merged residuals e f For the sake of quality integration m f and theoretical gas mass m th The difference.

[0051] Then merge the residuals ef The feedback is fed back into the corrected objective function of the volume dynamic inversion model as a correction term for the next cycle, thus achieving bidirectional coupling between mass fusion and volume inversion.

[0052] This embodiment establishes a dual-path quality fusion model for weighing and flow, eliminates abnormal sampling points based on the degree of local anomalies, and adjusts the fusion weights in real time according to the state residuals and volume inversion errors, thus achieving adaptive weighted fusion of the two quality measurements. Simultaneously, the fusion residuals are fed back to the volume inversion model, forming a bidirectional coupling between quality fusion and volume inversion, overcoming the problem of insufficient reliability of a single measurement path.

[0053] S4: Establish a time-varying reliability recursive model, using the state residual, volume inversion error, fusion residual, pressure stability, temperature stability, weighing stability, flow stability and time synchronization error output by the joint state observer as reliability observation variables, recursively calculate the system reliability, obtain the real-time reliability index and output the reliability change rate. It should be noted that in this embodiment, the eight confidence observation variables mentioned above are respectively e x The state residuals output by the joint state observer reflect the degree of convergence of the state estimation. e v Volume inversion error reflects the stability of volume correction. e f The fusion residual reflects the consistency of the dual-path quality fusion. P Pressure stability is the standard deviation of the pressure measurement value within the sliding window, reflecting the stability of the pressure conditions. T Temperature stability is the standard deviation of temperature measurements within a sliding window, reflecting the degree of stability of temperature operating conditions. w Weighing stability is the standard deviation of the weighing measurement value within the sliding window, reflecting the measurement stability of the weighing sensor. f Flow stability is the standard deviation of the flow measurement value within the sliding window, reflecting the measurement stability of the flow meter. t Time synchronization error reflects the time registration accuracy of data from each sensor.

[0054] Establish a credibility state observation vector: C =[e x , e v , e f , P , T , w , f , t ] T .

[0055] This embodiment then employs the Recursive Least Squares (RLS) algorithm to recursively estimate the system credibility. The system credibility is defined as... R c ,but t The system reliability at time is R c ( t The calculation process is as follows: Establish a credibility parameter model: ; in, for t The predicted confidence level of the observed values ​​at each time point. Let be the confidence state feature vector at time t, which is composed of the confidence state observation vector. C ( t This is obtained after normalization, feature mapping, or feature combination; for Transpose of; for t The confidence parameter vector obtained at time -1; Calculate the confidence observation residuals, which are based on... t The deviation between the credibility obtained from actual observation at any given time and the credibility predicted based on historical credibility parameters.

[0056] ; Among them, e R ( t )for t The reliability of the observation residuals at any given time. for t The predicted confidence level of the observed values ​​at each time point. y ( t )for t The confidence observations obtained at each time point can be used to calculate confidence sub-indices based on the eight confidence observation variables mentioned above, and then fused together. The specific calculation formula is as follows: ; in, For the first i The weights corresponding to the credibility scores of each credibility observation variable, and , For the first i Each credibility observation variable corresponds to a credibility score, and 0 ≤ ≤1; 8 represents the number of confidence level observation variables, and 0 ≤ y ( t )≤1.

[0057] When e R ( t When ) > 0, it means t The fact that the actual reliability at any given moment is higher than the predicted reliability indicates that the current state of the system is more stable than the prediction results of historical models.

[0058] When e R ( t When ) < 0, it means t If the actual reliability of a time point is lower than the predicted reliability, it indicates that there is a certain degree of anomaly or inconsistency in the current observation state.

[0059] When e R ( t When )≈0, it means that the current observation results are basically consistent with the historical model prediction results.

[0060] Calculate the recursive gain. Further calculation is performed using the Recursive Least Squares (RLS) algorithm: ; in, K ( t )for t The recursive gain at time step, F ( t -1) is t The parameter covariance matrix at time -1; for t The credibility state feature vector at time step is composed of the credibility state observation vector. C ( t This is obtained after normalization, feature mapping, or feature combination; for transpose, m It is the forgetting factor; and the forgetting factor generally satisfies 0 < m ≤1, its function is to adjust the degree of influence of historical data and current data on model parameter updates.

[0061] when m When the data is smaller, the model pays more attention to changes in the current data; when mWhen the value is close to 1, the model retains more of the influence of historical data.

[0062] Credibility parameter update: based on t Time-based reliability observation residuals, for t The credibility parameter at time -1 is corrected online, allowing the credibility model to be dynamically updated as the system's operating state changes. The credibility parameter update expression is: ; in, for t The credibility parameter vector updated at each time step. for t The confidence parameter vector obtained at time -1 K ( t )for t The recursive gain at time step, e R ( t )for t The credibility of the observation residuals is determined. K ( t )e R ( t The number indicates the amount of parameter update in this instance.

[0063] In order for the Recursive Least Squares (RLS) algorithm to run continuously, the parameter covariance matrix also needs to be updated: ; in, F ( t )for t The parameter covariance matrix at time t, F ( t -1) is t The parameter covariance matrix at time -1; I It is the identity matrix. K ( t )for t The recursive gain at time step, for t The credibility state feature vector at time step. m It is a forgetting factor.

[0064] The updated parameter covariance matrix is ​​mainly used to reflect t The uncertainty of the time-based reliability parameter estimation and the correlation between different parameters are considered. This is then used to calculate the next sampling period. t The recursive gain at time +1 provides the basis.

[0065] After obtaining the updated parameter covariance matrix, calculate t The system reliability at time is R c( t )for: ; in, The original confidence level is calculated based on the features at time t and the updated confidence parameters. This is a limiting function used to restrict the calculation results to the range [0,1], and: ; That is, ensure 0≤ R c ( t )≤1, when R c ( t The closer the value is to 1, the more reliable the current operating state of the system; when... R c ( t The closer the value is to 0, the higher the degree of inconsistency between the current observed data and the expected operating state, meaning that the current operating state is unreliable.

[0066] In the current output t Simultaneously outputting the system's reliability at any given time, it also outputs the reliability change rate. ,Right now: ; in, R c ( t )for t System reliability at any given moment R c ( t- Δ T )for t -Δ T System reliability at any given time; Δ T The sampling period.

[0067] The confidence rate of change is further fed back to the joint state observer to adjust the state estimation gain. This forms a confidence-state estimation-volume inversion-quality fusion-confidence-closed-loop feedback.

[0068] Credibility Indicators R c After the calculation is completed, the results are simultaneously fed back to the following three models.

[0069] The first feedback is to the objective function of the volumetric dynamic inversion model. A confidence index is used to adjust the iteration step size of the volumetric inversion. When R c When the value is high, a larger step size is used to accelerate convergence; when R c When the threshold is low, a smaller step size is used to ensure convergence stability.

[0070] The second feedback is sent to the dual-path quality-trusted fusion model. A trustworthiness index is used to adjust the fusion weights. When R c When the weight is low, increase the weight dependence on historical reliable measurement paths and decrease the weight of current abnormal measurement paths.

[0071] The third feedback is to the carbon emission calculation model: the credibility index is used to correct the emission calculation compensation coefficient. When R c Increase the compensation factor when the level is low to cover unmeasured emission risks.

[0072] Credibility change rate The Kalman gain of the joint state observer is used to adjust the following parameters for the next sampling period: when the confidence level drops rapidly, the covariance matrix of the observation noise is increased to reduce the impact of the observation on the state estimation; when the confidence level stabilizes, the preset parameter settings are restored.

[0073] This embodiment constructs a time-varying reliability recursive model, based on eight observed variables (state residual, volume inversion error, fusion residual, pressure stability, temperature stability, weighing stability, flow stability, and time synchronization error), to recursively calculate real-time reliability indicators and change rates, thereby realizing online quantitative evaluation of the reliability of the operating status of the recycling system and providing a quantitative basis for adaptive correction of each calculation link.

[0074] S5: Based on the corrected actual gas chamber volume, fusion mass value and reliability index, establish an insulating gas mass inversion model, calculate the real-time insulating gas emissions, and combine the global warming potential of insulating gas to calculate the real-time CO2 equivalent value. It should be noted that the actual emission quality m loss The calculation formula is as follows: m loss =m initial - m f - m remain ; in, m initia Indicates the theoretical mass before recycling. m f Indicates the quality of fusion. m remain This indicates the residual mass after the recycling process is completed.

[0075] Theoretical mass before recycling m initia Correction value based on air chamber volume V ( k +1) Initial pressureP 0. Initial temperature T 0. Calculations based on the gas law yielded the following: ; During the recycling process, t Residual mass after recovery in the gas chamber m remain : ; in P ( t ), T ( t )for t Pressure and temperature observations at any given time.

[0076] m f Indicates the quality of fusion. m remain This indicates the residual mass after the recycling process is completed.

[0077] Then, based on the global warming potential (GWP) of the insulating gas, the CO2 equivalent (CO2eq) is calculated, i.e.: CO 2eq= m loss GWP; Further dynamic compensation based on credibility: ; in K(T) This represents a reliable compensation function; the lower the reliability, the higher the compensation level, thus improving the accuracy of the final carbon emission calculation.

[0078] For SF6 gas, the GWP is taken as 23900 (based on the 100-year timescale global warming potential in the IPCC Fifth Assessment Report).

[0079] S6: Establish a theoretical recovery quality prediction model and perform deviation analysis between the theoretical recovery quality and the fusion quality; when the deviation exceeds the preset threshold, initiate abnormal leakage diagnosis; determine the leakage type and leakage level based on the abnormal leakage diagnosis; feed the leakage type and leakage level back to the credibility recursive model to correct the credibility index; It should be noted that the theoretical recovery mass prediction model in this embodiment is based on the gas chamber volume correction value V(k+1), pressure change trend, and temperature change trend to predict the theoretical recovery mass under the current operating conditions. m th ( t ),Right now: ; That is, the theoretical recovered mass is equal to the difference between the theoretical mass before recovery in the gas chamber and the residual mass after recovery.

[0080] Calculate theoretical recovery mass m th ( t ) and fusion quality m f ( t relative deviation d ( t ),Right now: ; Preset deviation threshold d th Typically, the value is taken as 0.05~0.10, which is 5%~10%. When d ( t )> d th If an abnormal leak is detected, the abnormal leak diagnosis process is initiated.

[0081] Abnormal leak diagnosis includes: Spectral features were extracted based on infrared absorption spectroscopy during the recycling process; A multi-source leakage feature set was established based on pressure change characteristics, mass change characteristics, volume correction change characteristics, spectral characteristics, and confidence change trajectory. The leakage type and leakage level are determined based on the multi-source leakage feature set. Specifically, when the same leakage type is detected in multiple consecutive sampling cycles, the reliability of the corresponding sampling cycle is reduced, the suppression weight of abnormal data in the dual-path quality fusion process is increased, and the volume dynamic inversion is re-executed to update the actual gas chamber volume.

[0082] It should be noted that, in order to further improve the reliability of emission calculations, this embodiment establishes an abnormal leakage diagnosis model.

[0083] First, infrared absorption spectra are collected, and characteristic absorption peaks of insulating gases (SF6 has a strong absorption peak near 10.6 μm) are obtained from infrared spectral sensors. Spectral characteristic parameters such as absorption peak intensity, peak position shift, and peak width variation are then extracted.

[0084] Subsequently, a multi-source leakage feature set was established, including: Pressure change characteristics, including the rate of pressure drop d P / d t Pressure fluctuation amplitude Δ P The degree of nonlinearity of the pressure change curve, etc.

[0085] Characteristics of quality change, including the rate of quality loss d m / d tCharacteristics of the inflection point of the mass loss curve, etc.

[0086] Volume correction variations include the time-varying pattern of volume correction amount and volume correction frequency.

[0087] Credibility change trajectory, including credibility indicators R c ( t The abnormal fluctuations in the trajectory of changes before and after the leak, and the rate of change of credibility.

[0088] By fusing the above multi-source features, a multi-source leakage feature set is constructed. F A classifier (such as a support vector machine, decision tree, or neural network) is used to determine the leakage type based on a multi-source leakage feature set. Leakage types include, but are not limited to: Micro-leakage at the sealing surface: characterized by a slow decrease in pressure, weak changes in the spectral signal, and a slow decrease in reliability; Pipe joint leakage is characterized by pressure fluctuations, moderate mass loss rate, and intermittent enhancement of spectral signal. Internal leakage of valves: characterized by abnormal pressure fluctuations, rapid mass loss rate, and abnormal volume correction. Gas chamber body damage and leakage: characterized by rapid pressure drop, fast mass loss rate, significant enhancement of spectral signal, and sharp decrease in reliability.

[0089] Leakage levels are classified according to leakage rate as follows: Level I minor leak: Leakage rate ≤ 0.1% / h; Level II ordinary leakage: 0.1% / h < leakage rate < 1% / h; Level III severe leak: Leakage rate ≥1% / h.

[0090] The leakage type and leakage level are fed back into the aforementioned recursive model for system reliability calculation. The specific feedback method is as follows: ; in, For the revised version t System reliability at any given moment for t The system reliability at time t, η1 is the correction coefficient, and its value ranges from (0,1]; leak type represents the leakage type, and leak level represents the leakage level. For a Level III severe leak, η1 takes a small value (e.g., 0.25~0.45), which significantly reduces the credibility; for a Level I minor leak, η1 takes a value close to 1 (e.g., 0.90~0.95), which moderately reduces the credibility.

[0091] If multiple consecutive sampling periods identify the same leak type, the reliability of the corresponding sampling period is further reduced, and the suppression weight of abnormal data during dual-path mass fusion is increased (i.e., the weight of abnormal measurement paths is reduced to below 50% of the original weight). Volume dynamic inversion, mass fusion, and carbon emissions are then recalculated to update the actual gas chamber volume. This achieves leak diagnosis, reliability adjustment, volume correction, emission calculation, and continuous feedback.

[0092] This embodiment uses the deviation analysis between theoretical recovery quality and fusion quality as the trigger condition for leak diagnosis. Based on the multi-source feature set, the leak type and level are determined, and the diagnosis results are fed back to the reliability recursive model and the data fusion module. This realizes the proactive intervention and correction of the data processing flow by the leak diagnosis results, and improves the data reliability under abnormal operating conditions.

[0093] S7: Generate carbon emission records based on actual gas chamber volume, fusion quality, credibility index, CO2 equivalent value and leak diagnosis results. At the same time, construct carbon asset data including recovery quality, leak status, credibility level and carbon emission amount, and upload it to the carbon emission management platform.

[0094] It should be noted that after each sampling period, carbon emission records are generated based on the following data: Timestamp: A unified time stamp for the current sampling period; Equipment number: A unique identifier for the monitored gas chamber; Air chamber volume correction value: V ( k +1); Fusion quality value: m f ; Credibility Indicators: R c ; CO2 equivalent value: CO2 eq ; Leak type: (If there is no leak, it is "normal"); Leak level: (None if there is no leak) Carbon emission records are encapsulated in a preset data format (such as JSON or XML) and stored in a local database.

[0095] Carbon asset data includes the following information dimensions: Recovery mass: Total mass of insulating gas recovered cumulatively; Leakage status: Leakage type, leakage level, leakage duration; Credibility level: based on R c Value classification (e.g.)R c ≥0.8 indicates high confidence, 0.6≤ R c <0.8 is considered moderately reliable. R c <0.6 indicates low confidence); Carbon emissions: Total cumulative CO2 equivalent emissions.

[0096] Carbon asset data is encapsulated according to the communication protocol agreed upon with the carbon emission management platform (such as MQTT, Modbus TCP or IEC61850) and uploaded to the carbon emission management platform via wired Ethernet or 4G / 5G wireless communication.

[0097] After receiving carbon emission records, leak diagnosis results, and reliability indicators uploaded by each field terminal, the carbon emission management platform performs the following operations: Statistical analysis: Historical operational data for each recovery process are statistically analyzed, including volume inversion convergence curves, dual-path quality fusion deviation distribution, reliability index change trajectory, and leak diagnosis accuracy.

[0098] Establish a historical error database: Store data such as model prediction errors, fusion biases, and reliability fluctuations from each recycling process into the historical error database to form an error sample library.

[0099] Model parameter optimization: Based on the historical error database, machine learning or statistical learning methods are used to periodically optimize and update the following parameters: Mass error weighting coefficients of the corrected objective function in the volumetric dynamic inversion model a 1 and state residual weighting coefficient a 2; The basic fusion weight coefficient w(C) of the weighing path and the basic fusion weight coefficient w(L) of the flow path in the dual-path quality reliable fusion model; The weighting coefficients and recursive gain corresponding to the credibility scores of the credibility observation variables in the credibility recursive model. K ( t The initial value of ).

[0100] Parameter distribution: The updated model parameters are distributed to the field data processing system. After receiving the data, the field system updates its local model parameters, thereby achieving self-learning optimization of the model parameters.

[0101] Through the above methods, the recycling system continuously evolves during long-term operation, constantly adapting to the impact of factors such as sensor drift, changes in operating conditions, and equipment aging, thus maintaining the long-term accuracy and reliability of carbon emission calculations.

[0102] This embodiment establishes a historical error database through a carbon emission management platform, continuously optimizes and updates the parameters of the volume dynamic inversion model, the dual-path mass fusion model, and the reliability recursive model, enabling the recycling system to continuously evolve during long-term operation and adapt to the effects of factors such as sensor drift, changes in operating conditions, and equipment aging, thus maintaining the long-term accuracy and reliability of carbon emission calculations.

[0103] Scenario setting: SF6 gas is recovered in a gas chamber of a 500kV GIS device. The nominal volume of the gas chamber is 5.0m³, but the actual volume has a certain deviation due to years of operation.

[0104] Initial Phase: After the recovery system is started, the joint state observer uses a nominal volume of 5.0 m³ as the initial state for recursive estimation. Simultaneously, the volume dynamic inversion model begins iteratively correcting the volume, using the deviation between the theoretical mass and the weighed / flow fusion mass as the objective function.

[0105] During the first 12 sampling periods, the volume inversion model found a deviation of approximately 3% between the theoretical quality and the fused quality. Through iterative optimization, the volume was corrected to 5.12 m³. The corrected volume was fed back to the joint state observer, which used it as a constraint to re-estimate the temperature and pressure distributions, reducing the state residual from 0.8% to 0.2%.

[0106] In the 18th sampling period, the weighing sensor was subjected to on-site vibration interference, resulting in abnormal fluctuations in the weighing mass. The local anomaly detection module of the dual-path mass fusion model identified this anomaly, marked the sampling point as abnormal, and removed it. Simultaneously, due to the increased variance of the weighing path, the basic fusion weights were adjusted from 0.5 / 0.5 to 0.3 / 0.7 (increased weights for flow paths). This increased fusion residual was fed back to the volumetric inversion model, which added a fusion residual constraint term to the objective function to slow down the volume correction speed and prevent erroneous volume corrections due to abnormal data.

[0107] During the 25th-35th sampling periods, due to significant fluctuations in ambient temperature (temperature stability σ) T Increase), the credibility index calculated by the time-varying credibility recursive model R c The confidence level dropped from 0.95 to 0.82. The confidence index was simultaneously fed back to all three models: ① The volume inversion step size was reduced from 0.05 to 0.03 to ensure convergence stability; ② In the dual-path fusion weighting, the weight of the weighing path, which is less sensitive to temperature, is increased from 0.3 to 0.4; ③ The compensation coefficient for carbon emission calculations has been increased from 0.08 to 0.12 to cover the risk of unmeasured emissions due to temperature fluctuations.

[0108] In the 60th sampling cycle, the relative deviation between the theoretical recovered quality and the fused quality reached 7.2%, exceeding the preset threshold of 5%, triggering an abnormal leak diagnosis. The diagnostic module, based on infrared spectral characteristics (SF6 absorption peak intensity decreased by 5%), pressure change characteristics (pressure decrease rate increased from 0.02% / h to 0.15% / h), quality change characteristics (increased mass loss rate), and the reliability trajectory (decreasing from 0.88 to 0.71 over 10 cycles), determined a "general leak at the sealing surface." The diagnostic result was fed back to the reliability recursive model, and the reliability index was corrected to 0.71 × 0.75 = 0.53. Simultaneously, the suppression weight of abnormal data during the dual-path quality fusion process was increased to 1.5 times, and the volumetric dynamic inversion was re-executed to update the volume.

[0109] After the recovery was completed, carbon emission records and all process data were uploaded to the carbon emission management platform. The platform stored the volumetric inversion convergence curve (finally converging to 5.18 m³), ​​fusion bias distribution, and confidence change trajectory during this recovery process in a historical error database. After accumulating historical data from multiple devices and multiple recovery operations, the platform optimized the iteration termination threshold of the volumetric inversion (adjusted from 0.001 to 0.0008) and the weights of the observation variables in the confidence recursion through statistical analysis, and sent the updated parameters to the field system. The volumetric convergence speed of subsequent recovery operations improved by approximately 20%, and the accuracy of confidence assessment improved by approximately 15%.

[0110] Through the coordinated operation of the above-mentioned multiple closed loops, this invention achieves full-process self-adaptation, self-correction, and self-optimization in the processing of carbon emission data for insulating gases, significantly improving the accuracy and reliability of carbon emission calculation.

[0111] Therefore, this invention establishes a multi-level feedback relationship between a joint state observer, a volume dynamic inversion model, a dual-path quality reliable fusion model, a time-varying reliability recursive model, an abnormal leakage diagnosis model, and a historical error database, achieving continuous closed-loop data processing for the entire process of insulating gas recovery. The models not only transfer data but also form bidirectional coupling and collaborative correction through state residuals, fusion residuals, reliability indices, and leakage diagnosis results. This enables the recovery system to continuously optimize volume estimation, quality fusion, and carbon emission calculation results based on real-time operating conditions. This avoids the cumulative errors caused by the fixed volume assumption in existing technologies, improves the utilization efficiency of multi-source measurement data and the metering stability under abnormal operating conditions, and effectively enhances the accuracy of insulating gas recovery quality calculation, carbon emission accounting accuracy, and the reliability of carbon asset data. It provides a data processing method with high engineering application value for insulating gas recovery and carbon emission management in the power industry.

[0112] Example 2 like Figure 2As shown, an embodiment 2 of the present invention discloses a processing system based on the processing method described in embodiment 1, and the processing system includes: Data acquisition module 10 is used to collect multi-source operation data throughout the entire process of insulating gas recovery and establish a multi-source status dataset under a unified time reference. The dynamic volume inversion module 11 is used to establish a multi-physics coupled state-space model of the gas chamber based on a multi-source state dataset, construct a joint state observer, and simultaneously recursively estimate the gas chamber volume, temperature, pressure, and gas mass to obtain predicted values. Based on the predicted values, a dynamic volume inversion model is established, with the deviation between the theoretical mass and the real-time recovered mass as the objective function. An adaptive iterative optimization method is used to correct the gas chamber volume in real time, and a continuous closed-loop self-correction mechanism for volume is formed. The dual-path mass fusion module 12 is used to calculate the weighing recovery mass and flow recovery mass based on the corrected air chamber volume, establish a reliable dual-path mass fusion model and output the fusion residual; then the fusion residual is fed back to the volume dynamic inversion model as the correction term of the objective function in the next cycle, realizing the bidirectional coupling between mass fusion and volume inversion. The credibility calculation module 13 is used to establish a time-varying credibility recursive model. It uses the state residual, volume inversion error, fusion residual, pressure stability, temperature stability, weighing stability, flow stability and time synchronization error output by the joint state observer as credibility observation variables to recursively calculate the system credibility, obtain the real-time credibility index and output the credibility change rate. The CO2 equivalent value calculation module 14 is used to establish an insulating gas quality inversion model based on the corrected actual gas chamber volume, fusion mass value and credibility index, calculate the real-time insulating gas emissions, and calculate the real-time CO2 equivalent value in combination with the global warming potential of insulating gas. Leakage diagnosis and level determination module 15 is used to establish a theoretical recovery quality prediction model and perform deviation analysis between the theoretical recovery quality and the fusion quality; when the deviation exceeds a preset threshold, abnormal leakage diagnosis is initiated; based on the abnormal leakage diagnosis, the leakage type and leakage level are determined; the leakage type and leakage level are fed back to the credibility recursive model to correct the credibility index. The carbon data generation and transmission module 16 is used to generate carbon emission records based on the actual gas chamber volume, fusion quality, credibility index, CO2 equivalent value and leakage diagnosis results. At the same time, it constructs carbon asset data including recovery quality, leakage status, credibility level and carbon emission amount, and uploads it to the carbon emission management platform.

[0113] like Figure 3As shown, this embodiment of the invention also provides an electronic device, which includes a memory 20, a processor 21, and computer instructions stored in the memory 20 and executable on the processor 21; when the processor 21 executes the instructions, it implements steps applied to the processing method.

[0114] In some embodiments of the present invention, the memory 20 may be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory may be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory may be random access memory (RAM), which serves as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchlink DRAM (SLDRAM), and Direct Rambus RAM (DRRAM). The memory 20 of the systems and methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0115] The processor 21 may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by the integrated logic circuitry in the hardware of the processor 21 or by software instructions. The processor 21 can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this invention. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this invention can be directly manifested as execution by a hardware decoding processor, or execution by a combination of hardware and software modules in the decoding processor. The software modules can be located in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory 20, and the processor 21 reads the information in memory 20 and, in conjunction with its hardware, completes the steps of the above processing method.

[0116] In some embodiments, the embodiments described herein can be implemented using hardware, software, firmware, middleware, microcode, or a combination thereof. For hardware implementation, the processing unit can be implemented in one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), general-purpose processors, controllers, microcontrollers, microprocessors, other electronic units for performing the functions described herein, or combinations thereof.

[0117] For software implementation, the techniques described herein can be achieved through modules (e.g., procedures, functions, etc.) that perform the functions described herein. The software code can be stored in memory and executed by a processor. The memory can be implemented within the processor or externally.

[0118] Another embodiment of the present invention provides a computer storage medium storing an executable program, which, when executed by a processor 21, can implement steps applied to a processing method. For example, such as... Figure 1 One or more of the methods shown.

[0119] In some embodiments, the computer storage medium may include various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0120] It should be noted that the technical solutions described in the embodiments of the present invention can be combined arbitrarily without conflict.

[0121] Finally, it should be noted that the above description is only for specific embodiments of the present invention. However, the present invention is not limited to the specific embodiments described above. Equivalent modifications and substitutions made to the present invention by those skilled in the art are also within the scope of the present invention. Therefore, all equivalent changes and modifications made without departing from the spirit and scope of the present invention are covered within the scope of the present invention.

Claims

1. A method for processing carbon emission data from insulating gas recovery, characterized in that: The processing method includes the following steps: S1: Collect multi-source operational data throughout the entire process of insulating gas recovery and establish a multi-source state dataset under a unified time reference. S2: Based on the multi-source state dataset, establish a multi-physics coupled state-space model of the gas chamber, construct a joint state observer, and simultaneously recursively estimate the gas chamber volume, temperature, pressure, and gas mass to obtain predicted values; based on the predicted values, establish a dynamic volume inversion model, using the deviation between the theoretical mass and the real-time recovered mass as the objective function, and adopt an adaptive iterative optimization method to correct the gas chamber volume in real time, forming a continuous closed-loop self-correction mechanism for the volume. S3: Calculate the weighing recovery mass and flow recovery mass based on the corrected air chamber volume, establish a dual-path reliable mass fusion model and output the fusion residual; then feed the fusion residual back to the volume dynamic inversion model as the correction term of the objective function in the next cycle, so as to realize the bidirectional coupling between mass fusion and volume inversion. S4: Establish a time-varying reliability recursive model, using the state residual, volume inversion error, fusion residual, pressure stability, temperature stability, weighing stability, flow stability and time synchronization error output by the joint state observer as reliability observation variables, recursively calculate the system reliability, obtain the real-time reliability index and output the reliability change rate. S5: Based on the corrected actual gas chamber volume, fusion mass value and credibility index, establish an insulating gas mass inversion model, calculate the real-time insulating gas emissions, and combine the global warming potential of insulating gas to calculate the real-time CO2 equivalent value. The reliability index is used to dynamically compensate for emission calculation errors. S6: Establish a theoretical recovery quality prediction model and perform deviation analysis between the theoretical recovery quality and the fusion quality; when the deviation exceeds the preset threshold, initiate abnormal leakage diagnosis; determine the leakage type and leakage level based on the abnormal leakage diagnosis; feed the leakage type and leakage level back to the credibility recursive model to correct the credibility index; S7: Generate carbon emission records based on actual gas chamber volume, fusion quality, credibility index, CO2 equivalent value and leak diagnosis results. At the same time, construct carbon asset data including recovery quality, leak status, credibility level and carbon emission amount, and upload it to the carbon emission management platform.

2. The method for processing carbon emission data from insulating gas recovery according to claim 1, characterized in that: The joint state observer establishes the following state vector: X =[ V ， T ， P ， m ]； in, V In the state of air chamber volume, T This refers to the temperature state of the air chamber. P This refers to the pressure state of the air chamber. m It is in gaseous mass state; The joint state observer performs synchronous recursive estimation of each state variable based on the collected pressure, temperature and recovery quality data, and outputs volume prediction, pressure prediction, temperature prediction and state residual; the state residual serves as the input parameter of the volume dynamic inversion model and the time-varying reliability recursive model.

3. The method for processing carbon emission data from insulating gas recovery according to claim 1, characterized in that: The volume dynamic inversion model constructs an objective function based on the deviation between the theoretical gas mass and the fusion mass, and iteratively corrects the gas chamber volume parameters based on the objective function. In each iteration, the state residual output by the joint state observer is introduced as a constraint term into the objective function to dynamically adjust the direction and magnitude of the volume correction. When the rate of change of volume is less than a preset threshold for multiple consecutive sampling periods, the current volume is determined as the actual air chamber volume and fed back to the joint state observer as the initial state for the next sampling period.

4. The method for processing carbon emission data from insulating gas recovery according to claim 1, characterized in that: The establishment of a dual-path quality-reliable fusion model as described in S3 includes: S30: Calculate the degree of local anomaly based on the weighing mass data and the flow integral data respectively; S31: Eliminate abnormal sampling points based on the degree of local anomaly to obtain two effective quality sequences; S32: Calculate the basic fusion weights based on the deviation, stability, and historical correction error between the two effective quality sequences; S33: Based on the state residuals and volumetric dynamic inversion errors output by the joint state observer, the basic fusion weights are corrected in real time to obtain the final fusion weights; S34: Calculate the fusion quality value based on the final fusion weights and output the fusion residual.

5. The method for processing carbon emission data from insulating gas recovery according to claim 1, characterized in that: The real-time reliability index mentioned in S4 needs to be fed back to the volume dynamic inversion model, the dual-path quality reliability fusion model and the carbon emission calculation model simultaneously. Among them: the real-time reliability index is used to adjust the volume dynamic inversion step size, adjust the dual-path quality fusion weight, and correct the carbon emission calculation compensation coefficient; The confidence change rate of the output described in S4 is used to adjust the state estimation gain for the next sampling period, thereby achieving closed-loop adaptive correction for the entire carbon emission calculation process.

6. The method for processing carbon emission data from insulating gas recovery according to claim 1, characterized in that: The abnormal leakage diagnosis described in S6 includes: Spectral features were extracted based on infrared absorption spectroscopy during the recycling process; A multi-source leakage feature set was established based on pressure change characteristics, mass change characteristics, volume correction change characteristics, spectral characteristics, and confidence change trajectory. The leakage type and leakage level are determined based on the aforementioned multi-source leakage feature set; Specifically, when the same leakage type is detected in multiple consecutive sampling cycles, the reliability of the corresponding sampling cycle is reduced, the suppression weight of abnormal data in the dual-path quality fusion process is increased, and the volume dynamic inversion is re-executed to update the actual gas chamber volume.

7. The method for processing carbon emission data from insulating gas recovery according to claim 1, characterized in that: After receiving carbon emission records, leak diagnosis results and reliability indicators, the carbon emission management platform described in S7 statistically analyzes historical operating data of each recycling process and establishes a historical error database. The historical error database is used to update the objective function parameters in the volume dynamic inversion model, the fusion weight correction parameters in the dual-path quality reliable fusion model, and the recursive parameters in the time-varying reliability recursive model. The updated parameters are then sent to the field data processing system to achieve self-learning optimization of the model parameters.

8. A data processing system for recovering carbon emissions from insulating gases, characterized in that: The processing system includes: The data acquisition module is used to collect multi-source operational data throughout the entire process of insulating gas recovery and to establish a multi-source status dataset under a unified time reference. The dynamic volume inversion module is used to establish a multi-physics coupled state-space model of the gas chamber based on a multi-source state dataset, construct a joint state observer, and simultaneously recursively estimate the gas chamber volume, temperature, pressure, and gas mass to obtain predicted values. Based on the predicted values, a dynamic volume inversion model is established, using the deviation between the theoretical mass and the real-time recovered mass as the objective function. An adaptive iterative optimization method is used to correct the gas chamber volume in real time, and a continuous closed-loop self-correction mechanism for volume is formed. The dual-path mass fusion module is used to calculate the weighing recovery mass and flow recovery mass based on the corrected gas chamber volume, and to establish a reliable dual-path mass fusion model and output the fusion residual. The fusion residual is then fed back to the volume dynamic inversion model as the correction term of the objective function in the next cycle, realizing the bidirectional coupling between mass fusion and volume inversion. The credibility calculation module is used to establish a time-varying credibility recursive model. It uses the state residual, volume inversion error, fusion residual, pressure stability, temperature stability, weighing stability, flow stability and time synchronization error output by the joint state observer as credibility observation variables to recursively calculate the system credibility, obtain the real-time credibility index and output the credibility change rate. The CO2 equivalent value calculation module is used to establish an insulating gas quality inversion model based on the corrected actual gas chamber volume, fusion mass value and credibility index, calculate the real-time insulating gas emissions, and calculate the real-time CO2 equivalent value in combination with the global warming potential of insulating gas. The Leakage Diagnosis and Level Determination module is used to establish a theoretical recovery quality prediction model and perform deviation analysis between the theoretical recovery quality and the fusion quality. When the deviation exceeds a preset threshold, abnormal leakage diagnosis is initiated. Based on the abnormal leakage diagnosis, the leakage type and leakage level are determined. The leakage type and leakage level are fed back to the credibility recursive model to correct the credibility index. The carbon data generation and transmission module is used to generate carbon emission records based on the actual gas chamber volume, fusion quality, credibility index, CO2 equivalent value and leakage diagnosis results. At the same time, it constructs carbon asset data including recovery quality, leakage status, credibility level and carbon emission amount, and uploads it to the carbon emission management platform.

9. An electronic device, characterized in that, include: A processor and a memory for storing a computer program capable of running on the processor, wherein the processor, when running the computer program, implements the steps of the processing method according to any one of claims 1 to 7.

10. A computer storage medium, characterized in that: The system contains an executable program that, when executed by a processor, implements the steps of the processing method as described in any one of claims 1 to 7.

Images