Method for accounting for carbon dioxide emissions based on a genset and readable storage medium

Abstract

The application relates to the technical field of carbon emission management, in particular to a carbon dioxide emission accounting method based on a generator set and a readable storage medium. The method comprises the following steps: obtaining a set of operation parameters of the generator set in a current time period; the set of operation parameters comprises operation data of a plurality of operation parameters in the current time period; based on the set of operation parameters, a parameter disturbance tensor sequence and a disturbance speed sequence are constructed; the parameter disturbance tensor sequence is used for representing disturbance conditions of the generator set at different time points, and the disturbance speed sequence is used for representing dynamic trends of a disturbance trajectory; based on the parameter disturbance tensor sequence, the disturbance speed sequence, a historical trajectory embedding library and a reference carbon emission trajectory corresponding to a historical disturbance trajectory, a carbon emission trajectory of the generator set is predicted to generate a carbon emission prediction trajectory. Since the parameter disturbance tensor sequence and the disturbance speed sequence accurately reflect the operation state of the generator set and dynamic changes under complex disturbance, the accuracy of carbon emission accounting can be improved.

Description

Technical Field

[0001] This application relates to the field of carbon emission management technology, and in particular to a carbon dioxide emission accounting method based on generator sets and a readable storage medium. Background Technology

[0002] Traditional carbon emission accounting methods mainly rely on simple fuel consumption and emission factors to estimate carbon dioxide emissions.

[0003] However, these methods often lack precision and cannot accurately reflect the real-time operating status of generator sets and complex disturbance factors, resulting in low accuracy in carbon emission accounting. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, apparatus, computer equipment, computer-readable storage medium, and computer program product for calculating carbon dioxide emissions based on generator sets, which can improve the accuracy of carbon emission accounting, in response to the above-mentioned technical problems.

[0005] Firstly, this application provides a method for calculating carbon dioxide emissions based on generator sets, including:

[0006] Obtain the set of operating parameters for the generator set within the current time period; the set of operating parameters includes the operating data of multiple operating parameters within the current time period;

[0007] Based on the set of operating parameters, a parameter perturbation tensor sequence and a perturbation velocity sequence are constructed. The parameter perturbation tensor sequence is used to characterize the perturbation of the generator set at different time points, and the perturbation velocity sequence is used to characterize the dynamic trend of the perturbation trajectory.

[0008] Based on the parameter perturbation tensor sequence, perturbation velocity sequence, historical trajectory embedding library, and reference carbon emission trajectory corresponding to historical perturbation trajectory, the carbon emission trajectory of generator set is predicted, and a carbon emission prediction trajectory is generated.

[0009] In one embodiment, based on the set of operating parameters, a parameter perturbation tensor sequence and a perturbation velocity sequence are constructed, including:

[0010] Using a preset sliding time window and a preset moving step size, parameter change rate analysis is performed on each operating parameter in the operating parameter set to generate a parameter perturbation tensor sequence. The parameter perturbation tensor sequence includes parameter perturbation tensors for multiple time windows within the current time period, and the parameter perturbation tensor includes the change rate of multiple operating parameters within the corresponding time window.

[0011] The rate of change of adjacent parameter perturbation tensors in the parameter perturbation tensor sequence is analyzed to generate a perturbation velocity sequence.

[0012] In one embodiment, the carbon emission trajectory of the generator set is predicted based on the parameter perturbation tensor sequence, the perturbation velocity sequence, the historical trajectory embedding library, and the reference carbon emission trajectory corresponding to the historical perturbation trajectory, generating a carbon emission prediction trajectory, including:

[0013] Based on the parameter perturbation tensor sequence and perturbation velocity sequence, a target perturbation trajectory map of the generator set in the current time period is constructed;

[0014] The target disturbance trajectory map and the historical trajectory embedding library are matched to determine the search matching set; the search matching set includes at least one historical disturbance trajectory with a similarity greater than or equal to the target disturbance trajectory map.

[0015] Based on the reference carbon emission trajectories corresponding to each historical disturbance trajectory in the search matching set, the carbon emission trajectory of the generator set is predicted, and a carbon emission prediction trajectory is generated.

[0016] In one embodiment, based on the reference carbon emission trajectories corresponding to each historical disturbance trajectory in the search matching set, the carbon emission trajectory of the generator set is predicted to generate a carbon emission prediction trajectory, including:

[0017] Determine the fusion weights corresponding to each historical disturbance trajectory in the search matching set;

[0018] Based on the fusion weights corresponding to each historical disturbance trajectory in the search matching set, the reference carbon emission trajectories corresponding to each historical disturbance trajectory in the search matching set are weighted and fused to generate carbon emission prediction trajectories.

[0019] In one embodiment, determining the fusion weights corresponding to each historical disturbance trajectory in the search matching set includes:

[0020] Multiple evaluation metrics are used to quantitatively evaluate the similarity between each historical disturbance trajectory and the target disturbance trajectory map in the search matching set, and a similarity score vector is generated; the similarity score vector includes the similarity score value between each historical disturbance trajectory and the target disturbance trajectory map;

[0021] Based on the similarity score vector, the fusion weight corresponding to each historical disturbance trajectory in the search matching set is determined; the fusion weight is positively proportional to the corresponding similarity score value.

[0022] In one embodiment, multiple evaluation metrics are used to quantitatively evaluate the similarity between each historical perturbation trajectory in the search matching set and the target perturbation trajectory map, generating a similarity score vector, including:

[0023] The target perturbation trajectory map is segmented at multiple scales to obtain the segmentation results at each scale; the segmentation results include multiple perturbation trajectory sub-maps corresponding to the target perturbation trajectory map at the corresponding scale.

[0024] For each historical perturbation trajectory in the search matching set, various evaluation indicators are used to determine the fusion matching score between the historical perturbation trajectory and the segmentation results at each segmentation scale.

[0025] Based on the fusion matching scores corresponding to each evaluation indicator, the similarity score between the historical disturbance trajectory and the target disturbance trajectory map is determined.

[0026] A similarity score vector is generated based on the similarity scores between each historical disturbance trajectory and the target disturbance trajectory map.

[0027] In one embodiment, the target disturbance trajectory map is segmented at multiple scales to obtain the segmentation results at each scale, including:

[0028] Obtain multi-scale segmentation schemes;

[0029] Based on the segmentation schemes at each scale, the target disturbance trajectory map is segmented to obtain the segmentation results at each segmentation scale.

[0030] In one embodiment, obtaining a multi-scale segmentation scheme includes:

[0031] Based on the parameter perturbation tensor sequence and the perturbation velocity sequence, a joint perturbation gradient sequence is established;

[0032] Based on a multi-scale smoothing filter, a multi-resolution view is extracted from the perturbation joint gradient sequence to obtain multiple perturbation joint gradient views at different resolutions.

[0033] Under the joint gradient view of perturbation at various resolutions, the joint gradient sequence of perturbation is segmented based on the local peak of perturbation intensity to obtain the segmented structure of multiple perturbation trajectory subgraphs under the joint gradient view of different resolutions.

[0034] Based on the segmented structure of multiple perturbation trajectory subgraphs under the joint gradient view of the resolution, the scale segmentation scheme corresponding to the resolution is determined.

[0035] In one embodiment, the multiple evaluation metrics include at least two of the following: DTW matching distance metric, perturbation direction consistency metric, and anomaly point overlap metric.

[0036] In one embodiment, various evaluation metrics are used to determine the fusion matching score between historical perturbation trajectories and segmentation results at each segmentation scale, including:

[0037] Activate the DTW matching distance evaluation channel and use the DTW matching distance index to determine the first fusion matching score between the historical perturbation trajectory and the segmentation results at each segmentation scale;

[0038] Activate the parameter perturbation direction evaluation channel, use the perturbation direction consistency index, and determine the second fusion matching score between the historical perturbation trajectory and the segmentation results at each segmentation scale;

[0039] Activate the anomaly overlap channel and use the anomaly overlap index to determine the third fusion matching score between the historical disturbance trajectory and the segmentation results at each segmentation scale.

[0040] Accordingly, based on the fusion matching scores corresponding to each evaluation indicator, the similarity score between the historical perturbation trajectory and the target perturbation trajectory map is determined, including:

[0041] Based on the first fusion matching score, the second fusion matching score, and the third fusion matching score, the similarity score between the historical disturbance trajectory and the target disturbance trajectory map is determined.

[0042] In one embodiment, the process of obtaining the historical trajectory embedding library includes:

[0043] Acquire multi-source datasets; multi-source datasets include at least two of the following: internal unit data, similar consortium data, and literature data, and multi-source datasets have a defined trust range;

[0044] Based on the defined trust range, data scores are performed on each data source in the multi-source dataset to obtain the data score results for each data source.

[0045] Cross-authentication is performed on each data source in the multi-source dataset to obtain the cross-authentication results of each data source; cross-authentication includes at least one of similarity comparison verification, response pattern consistency verification, multi-source redundancy support verification, and outlier reverse verification.

[0046] Based on the data scoring results and cross-certification results of each data source, a comprehensive score for each data source is determined;

[0047] Based on the comprehensive scores from various data sources, the data sources in the multi-source dataset are cleaned and updated to establish a historical trajectory embedding library.

[0048] In one embodiment, based on a defined trust range, data scores are performed on each data source in the multi-source dataset to obtain data score results for each data source, including:

[0049] Completeness evaluation is performed on each data source in the multi-source dataset, and the first scoring feature of each data source is generated.

[0050] Evaluate the time freshness of each data source in the multi-source dataset and generate a second scoring feature for each data source.

[0051] The calibrated trust range is used as the third scoring feature for each data source;

[0052] Based on the first, second, and third scoring features of each data source, data scores are assigned to each data source to obtain the data score results for each data source.

[0053] In one embodiment, the method further includes:

[0054] Monitor the actual carbon emissions of generator sets and generate actual carbon emission trajectories;

[0055] A deviation analysis was performed between the actual carbon emission trajectory and the predicted carbon emission trajectory to determine the deviation residual.

[0056] The carbon emission prediction trajectory after the current time is corrected based on the deviation residual, and the corrected carbon emission prediction trajectory is generated.

[0057] In one embodiment, the method further includes:

[0058] Based on the set of operating parameters and the predicted carbon emission trajectory, a joint optimization objective function is constructed;

[0059] The balance control parameters of the generator set are determined based on the joint optimization objective function;

[0060] The generator set is operated and controlled based on the balance control parameters.

[0061] In one embodiment, the method further includes:

[0062] Data mapping is performed on the set of operating parameters and the carbon emission prediction trajectory, and the correlation between the data-mapped set of operating parameters and the carbon emission prediction trajectory is stored in a preset data recording space;

[0063] Based on the correlations stored in the preset data recording space, the working status of the generator set is continuously monitored;

[0064] If an abnormal operating status of the generator set is detected, an abnormal data warning will be generated based on the abnormal data.

[0065] Secondly, this application also provides a carbon dioxide emission accounting device based on a generator set, comprising:

[0066] The parameter acquisition module is used to acquire the set of operating parameters of the generator set within the current time period; the set of operating parameters includes the operating data of multiple operating parameters within the current time period;

[0067] The disturbance construction module is used to construct parameter disturbance tensor sequences and disturbance velocity sequences based on the set of operating parameters. The parameter disturbance tensor sequences are used to characterize the disturbance situation of the generator set at different time points, and the disturbance velocity sequences are used to characterize the dynamic trend of the disturbance trajectory.

[0068] The carbon emission prediction module is used to predict the carbon emission trajectory of generator sets based on parameter perturbation tensor sequences, perturbation velocity sequences, historical trajectory embedding libraries, and reference carbon emission trajectories corresponding to historical perturbation trajectories, and generate carbon emission prediction trajectories.

[0069] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the carbon dioxide emission accounting method based on generator sets described in the first aspect above.

[0070] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the carbon dioxide emission accounting method based on generator sets described in the first aspect above.

[0071] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the carbon dioxide emission accounting method based on generator sets described in the first aspect above.

[0072] The aforementioned method, apparatus, computer equipment, storage medium, and computer program products for calculating carbon dioxide emissions from generator sets acquire a set of operating parameters for the generator set within the current time period. This set includes operating data for multiple operating parameters within the current time period. Based on this set, a parameter perturbation tensor sequence and a perturbation velocity sequence are constructed. The parameter perturbation tensor sequence characterizes the perturbation of the generator set at different points in time, while the perturbation velocity sequence characterizes the dynamic trend of the perturbation trajectory. Furthermore, based on the parameter perturbation tensor sequence, the perturbation velocity sequence, a historical trajectory embedding library, and reference carbon emission trajectories corresponding to historical perturbation trajectories, the carbon emission trajectory of the generator set is predicted, generating a predicted carbon emission trajectory. In other words, the carbon emission accounting method proposed in this application constructs a parameter perturbation tensor sequence and a perturbation velocity sequence to characterize the perturbation status and dynamic change trend of the perturbation trajectory of the generator set in the current time period, and matches them with a historical trajectory embedding library. The reference carbon emission trajectory corresponding to the historical perturbation trajectory is used as a reference trajectory for carbon emission prediction. Since the parameter perturbation tensor sequence and the perturbation velocity sequence can accurately reflect the real-time operating status of the generator set and the dynamic change of the operating status under complex perturbation factors, this method can achieve accurate accounting of the carbon emissions of the generator set and improve the accuracy of carbon emission accounting. Attached Figure Description

[0073] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0074] Figure 1 This is a diagram illustrating the application environment of a carbon dioxide emission accounting method based on generator sets in one embodiment.

[0075] Figure 2 This is a flowchart illustrating a method for calculating carbon dioxide emissions based on generator sets in one embodiment.

[0076] Figure 3 This is a flowchart illustrating a method for calculating carbon dioxide emissions based on generator sets in another embodiment;

[0077] Figure 4 This is a flowchart illustrating a method for calculating carbon dioxide emissions based on generator sets in another embodiment;

[0078] Figure 5 This is a flowchart illustrating a method for calculating carbon dioxide emissions based on generator sets in another embodiment;

[0079] Figure 6 This is a flowchart illustrating a method for calculating carbon dioxide emissions based on generator sets in another embodiment;

[0080] Figure 7 This is a flowchart illustrating a method for calculating carbon dioxide emissions based on generator sets in another embodiment;

[0081] Figure 8 This is a flowchart illustrating a method for calculating carbon dioxide emissions based on generator sets in another embodiment;

[0082] Figure 9 This is a flowchart illustrating a method for calculating carbon dioxide emissions based on generator sets in another embodiment;

[0083] Figure 10 This is a flowchart illustrating a method for calculating carbon dioxide emissions based on generator sets in another embodiment;

[0084] Figure 11 This is a schematic diagram of the complete process of a carbon dioxide emission accounting method based on generator sets in one embodiment;

[0085] Figure 12 This is a structural block diagram of a carbon dioxide emission accounting device based on a generator set in one embodiment.

[0086] Figure 13 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0087] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0088] To address the technical problem that existing carbon emission accounting methods cannot accurately reflect the dynamic changes and complex disturbances in the operating status of generator sets, resulting in insufficient accuracy and real-time performance, this application provides a carbon dioxide emission accounting method based on generator sets. This method constructs parameter perturbation tensor sequences and perturbation velocity sequences to characterize the actual disturbance conditions and dynamic trends of the disturbance trajectory of the generator set. It then combines these with a historical trajectory embedding library for trajectory similarity matching and searching, and uses reference carbon emission trajectories corresponding to historical disturbance trajectories for carbon emission trajectory prediction. This method accurately calculates the carbon dioxide emission trajectory of the generator set, solving the technical problem that existing carbon emission accounting methods cannot accurately reflect the dynamic changes in the operating status of generator sets, resulting in insufficient accuracy and real-time performance. It achieves the technical effect of improving the accuracy and real-time performance of carbon emission accounting through parameter perturbation analysis of generator sets.

[0089] The carbon dioxide emission accounting method based on generator sets provided in this application embodiment can be applied to, for example... Figure 1 In the application environment shown, computer device 101 can be used to calculate and predict the carbon dioxide emissions of generator sets and generate carbon emission prediction trajectories. Computer device 101 can be a terminal or a server. The terminal can be, but is not limited to, various personal computers, laptops, smartphones, tablets, etc., and the server can be implemented using a standalone server or a server cluster composed of multiple servers, including but not limited to local servers, remote servers, cloud servers, etc.

[0090] In one exemplary embodiment, such as Figure 2 As shown, a method for calculating carbon dioxide emissions based on generator sets is provided, and this method is applied to... Figure 1 The following steps, 201 to 203, are used as an example of computer equipment.

[0091] Step 201: Obtain the set of operating parameters of the generator set in the current time period.

[0092] The set of operating parameters can include the operating data of multiple operating parameters within the current time period.

[0093] For example, within the current time period, the generator set's operating parameters can be collected at fixed time intervals to establish a set of operating parameters for that time period. This means periodically collecting the generator set's operating parameters at fixed time intervals. These operating parameters can include, but are not limited to, key indicators such as power output, load, speed, temperature, and pressure, which directly reflect the generator set's operating status and performance. To ensure data representativeness and completeness, parameter collection should be performed at fixed time intervals, such as once per minute or hour, to obtain a parameter dataset with time-series characteristics—the operating parameter set. The collected operating parameters cover various key indicators during generator set operation, such as power output, fuel consumption, temperature, and pressure. These parameters are integrated into an operating parameter set, providing a comprehensive data foundation for subsequent analysis.

[0094] For example, the generator set may be equipped with acquisition devices corresponding to different parameter types. These acquisition devices are used to periodically acquire multiple operating parameters of the generator set. These acquisition devices can communicate with computer devices and send the acquired operating data of the corresponding operating parameters to the computer devices so that the computer devices can construct an operating parameter set based on the operating data of multiple operating parameters.

[0095] For example, the acquisition device can also send the acquired operating data of the corresponding operating parameters to a preset data storage space. The computer device obtains the operating data of multiple operating parameters in the current time period from the preset data storage space and constructs the corresponding set of operating parameters.

[0096] Step 202: Based on the set of operating parameters, construct the parameter perturbation tensor sequence and the perturbation velocity sequence.

[0097] Among them, the parameter perturbation tensor sequence is used to characterize the perturbation of the generator set at different time points, and the perturbation velocity sequence is used to characterize the dynamic trend of the perturbation trajectory.

[0098] For example, the computer device can perform parameter change rate analysis on the set of operating parameters and generate a parameter perturbation tensor sequence based on the parameter change rate at multiple time points; further, the computer device can also perform tensor change rate analysis on the parameter perturbation tensor sequence and generate a perturbation velocity sequence based on the tensor change rate at multiple time points.

[0099] In one optional implementation, the computer device can employ a preset sliding time window and a preset moving step size to perform parameter change rate analysis on each operating parameter in the operating parameter set, generating a parameter perturbation tensor sequence. This parameter perturbation tensor sequence includes parameter perturbation tensors for multiple time windows within the current time period, and each parameter perturbation tensor includes the change rate of multiple operating parameters within the corresponding time window. Specifically, the computer device can perform parameter change rate analysis on the operating parameter set within a predetermined sliding time window. This predetermined sliding time window refers to selecting a fixed-length time window within the time series of operating parameters and gradually moving this time window along the time axis based on a preset moving step size, thereby analyzing the parameter changes within each time window. Optionally, the length of the sliding time window can be flexibly set according to the operating characteristics of the generator set and the frequency of parameter changes, for example, it can range from a few minutes to several hours. Through sliding time window analysis, the changing trends of operating parameters at different time scales can be captured, thus more accurately identifying the perturbations of operating parameters.

[0100] By using a sliding time window approach, the rate of change of each operating parameter within each time window can be calculated. The rate of change refers to the speed at which a given operating parameter changes within a certain time window, and can be calculated by comparing the initial and final values ​​of the operating parameter within the time window. For example, the formula for calculating the rate of change can be the difference between the initial and final parameter values ​​within the time window, divided by the initial value. If the time window includes two time points, the formula can be the difference between the current parameter value and the parameter value at the previous time point, divided by the parameter value at the previous time point.

[0101] Next, having obtained the rate of change of each operating parameter in each time window, a parameter perturbation tensor sequence can be constructed based on this rate of change. This parameter perturbation tensor sequence refers to organizing the calculated rate of change of parameters within each sliding time window into an ordered sequence. A tensor is a multidimensional array that can simultaneously contain the rate of change information of multiple operating parameters, thus comprehensively reflecting the multidimensional perturbation of the generator set's operating parameters. For example, if the generator set's power, temperature, and pressure are considered, the rate of change of parameters within each time window can be represented as a three-dimensional tensor. This three-dimensional tensor includes the rate of change of power, temperature, and pressure within that time window, and the rate of change of parameters across all time windows during the entire operation constitutes a parameter perturbation tensor sequence. This sequence provides crucial data support for the construction of perturbation trajectories and carbon emission prediction in subsequent steps, enabling the entire carbon dioxide emission accounting method to achieve accurate carbon emission accounting and prediction based on the actual perturbation of the generator set's operating parameters.

[0102] Furthermore, the computer equipment can also perform rate-of-change analysis on adjacent parameter perturbation tensors in the parameter perturbation tensor sequence to generate a perturbation velocity sequence. That is, it performs further rate-of-change analysis on the parameter perturbation tensor sequence obtained in the previous stage, calculates the rate of change between adjacent parameter perturbation tensors, and constructs a perturbation velocity sequence based on these rates of change, thereby characterizing the dynamic trend of the perturbation trajectory during generator unit operation.

[0103] The parameter perturbation tensor sequence obtained from the above analysis reflects the perturbation of the generator unit at different time points. To further understand the evolution of the perturbation, the rate of change of perturbation between adjacent time points can be calculated. In other words, by comparing the parameter perturbation tensors at two adjacent time points, the degree of change of the perturbation during that period can be determined. For example, the formula for calculating the rate of change can be expressed as: Perturbation rate of change = (Parameter perturbation tensor at the current time point - Parameter perturbation tensor at the previous time point) / Parameter perturbation tensor at the previous time point.

[0104] Through the above calculations, the rate of change of the disturbance at each time point can be obtained. These rates of change are further organized into a disturbance velocity sequence. The disturbance velocity sequence is a time series data that reflects the rate and direction of change of the parameter disturbance over time. For example, if the disturbance velocity is positive, it indicates that the parameter disturbance is increasing at that time point; if it is negative, it indicates that the parameter disturbance is decreasing at that time point.

[0105] Perturbation velocity sequences are used to characterize the dynamic trends of perturbation trajectories, revealing the development trends of perturbations in the system over different time periods. This helps identify potential operational risks and further provides a basis for carbon emission prediction. For example, if a perturbation velocity sequence shows an accelerating trend in a certain parameter's perturbation, it may indicate that the operating state of the generator unit is changing, thus affecting its carbon dioxide emissions. Furthermore, perturbation velocity sequences can provide important reference information for historical trajectory matching and multi-dimensional index scoring in subsequent steps. By analyzing perturbation velocity sequences, historical trajectories similar to the current perturbation trajectory can be identified more accurately, thereby improving the accuracy and reliability of carbon emission prediction.

[0106] Step 203: Based on the parameter perturbation tensor sequence, perturbation velocity sequence, historical trajectory embedding library, and reference carbon emission trajectory corresponding to historical perturbation trajectory, predict the carbon emission trajectory of the generator set and generate the carbon emission prediction trajectory.

[0107] Among them, the parameter perturbation tensor sequence and the perturbation velocity sequence can characterize the perturbation status and dynamic change trend of the perturbation trajectory of the generator set in the current time period. Based on this, the computer device can match the parameter perturbation tensor sequence and perturbation velocity sequence of the generator set in the current time period with a historical trajectory embedding library to obtain at least one historical perturbation trajectory that is similar to the perturbation status and dynamic change trend of the perturbation trajectory of the generator set in the current time period (similarity greater than or equal to a preset similarity). Furthermore, the computer device can predict the carbon emission trajectory of the generator set based on the reference carbon emission trajectory corresponding to the at least one historical perturbation trajectory to generate a carbon emission prediction trajectory.

[0108] For example, if there is only one similar historical disturbance trajectory, the reference carbon emission trajectory corresponding to that historical disturbance trajectory can be used as the predicted carbon emission trajectory of the generator set. Alternatively, based on the prediction correction coefficient, the reference carbon emission trajectory corresponding to that historical disturbance trajectory can be corrected to obtain the predicted carbon emission trajectory of the generator set. If there are multiple similar historical disturbance trajectories, the predicted carbon emission trajectory of the generator set can be obtained by fusing the reference carbon emission trajectories corresponding to multiple historical disturbance trajectories. For example, the predicted carbon emission trajectory of the generator set can be obtained by weighted fusing the reference carbon emission trajectories corresponding to multiple historical disturbance trajectories. Alternatively, the predicted carbon emission trajectory of the generator set can be obtained by inputting the reference carbon emission trajectories corresponding to multiple historical disturbance trajectories into the carbon emission trajectory prediction model for fusion prediction.

[0109] For example, regarding the above matching process, namely, matching the parameter perturbation tensor sequence and perturbation velocity sequence of the generator set in the current time period with the historical trajectory embedding library to obtain at least one historical perturbation trajectory similar to the perturbation situation and dynamic change trend of the perturbation trajectory of the generator set in the current time period, the implementation of this matching process may include: assuming that the historical trajectory embedding library includes multiple historical perturbation trajectories and the historical parameter perturbation tensor sequence and historical perturbation velocity sequence corresponding to each historical perturbation trajectory, based on this, the parameter perturbation tensor sequence and perturbation velocity sequence of the generator set in the current time period can be matched with the historical parameter perturbation tensor sequence and historical perturbation velocity sequence corresponding to each historical perturbation trajectory. Similarity matching is performed between the dynamic tensor sequence and the historical disturbance velocity sequence. For example, for each historical disturbance trajectory, the first similarity between the parameter disturbance tensor sequence of the current time period and the historical parameter disturbance tensor sequence corresponding to the historical disturbance trajectory is determined, and the second similarity between the disturbance velocity sequence of the current time period and the historical disturbance velocity sequence corresponding to the historical disturbance trajectory is determined. A comprehensive similarity is determined based on the first and second similarities. If the comprehensive similarity is greater than or equal to the preset similarity, the historical disturbance trajectory corresponding to the comprehensive similarity can be determined to be a historical disturbance trajectory similar to the disturbance situation and dynamic change trend of the engine unit in the current time period.

[0110] For example, the matching process can also be implemented by: assuming that the historical trajectory embedding library includes multiple historical disturbance trajectories, constructing a target disturbance trajectory map of the generator set in the current time period based on the parameter disturbance tensor sequence and the disturbance velocity sequence, and then matching the target disturbance trajectory map with each historical disturbance trajectory in the historical trajectory embedding library to obtain at least one historical disturbance trajectory that is similar to the disturbance situation and dynamic change trend of the disturbance trajectory of the generator set in the current time period.

[0111] In the aforementioned method for calculating carbon dioxide emissions from generator sets, the following steps are taken: First, a set of operating parameters for the generator set within the current time period is obtained. This set includes operating data for multiple operating parameters within the current time period. Second, based on this set, a parameter perturbation tensor sequence and a perturbation velocity sequence are constructed. The parameter perturbation tensor sequence characterizes the perturbation of the generator set at different time points, and the perturbation velocity sequence characterizes the dynamic trend of the perturbation trajectory. Third, based on the parameter perturbation tensor sequence, the perturbation velocity sequence, a historical trajectory embedding library, and the reference carbon emission trajectory corresponding to the historical perturbation trajectory, the carbon emission trajectory of the generator set is predicted, generating a predicted carbon emission trajectory. In other words, the carbon emission accounting method proposed in this application constructs a parameter perturbation tensor sequence and a perturbation velocity sequence to characterize the perturbation status and dynamic change trend of the perturbation trajectory of the generator set in the current time period, and matches them with a historical trajectory embedding library. The reference carbon emission trajectory corresponding to the historical perturbation trajectory is used as a reference trajectory for carbon emission prediction. Since the parameter perturbation tensor sequence and the perturbation velocity sequence can accurately reflect the real-time operating status of the generator set and the dynamic change of the operating status under complex perturbation factors, this method can achieve accurate accounting of the carbon emissions of the generator set and improve the accuracy of carbon emission accounting.

[0112] In one exemplary embodiment, an optional implementation scheme for predicting carbon emission trajectories based on parameter perturbation tensor sequences and perturbation velocity sequences is provided. For example... Figure 3 As shown, step 203 above may include steps 301 to 303. Wherein:

[0113] Step 301: Based on the parameter perturbation tensor sequence and the perturbation velocity sequence, construct the target perturbation trajectory map of the generator set in the current time period.

[0114] In this example, the parameter perturbation tensor sequence and perturbation velocity sequence obtained from the preceding process can be used to construct the target perturbation trajectory diagram. The perturbation trajectory diagram is a visualization tool that can intuitively show the dynamic changes of generator set operating parameter perturbations. The parameter perturbation tensor sequence reflects the parameter perturbation state at different time points, while the perturbation velocity sequence further reveals the rate and direction of perturbation change.

[0115] Step 302: Match the target disturbance trajectory map with the historical trajectory embedding library to determine the search matching set.

[0116] The search matching set includes at least one historical disturbance trajectory with a similarity to the target disturbance trajectory map that is greater than or equal to a preset similarity. The historical trajectory embedding library may include historical disturbance trajectories (such as historical disturbance trajectory maps) for multiple historical time periods, with each historical time period including one or more historical disturbance trajectories.

[0117] For example, a similar trajectory search and matching process can be performed on a historical trajectory embedding library based on the constructed target perturbation trajectory map for the current time period. The goal of this search and matching process is to select multiple historical perturbation trajectories from the historical trajectory embedding library that are most similar to the current target perturbation trajectory map, so as to obtain possible operating state patterns, potential failure modes, or carbon emission characteristics through matching. For example, the Euclidean distance between the target perturbation trajectory map and the historical perturbation trajectories can be determined, and the Euclidean distance can be input into a normalization function to obtain the similarity between the two; the closer the distance, the higher the similarity, and the farther the distance, the lower the similarity. If the similarity is greater than or equal to a preset similarity, the historical perturbation trajectory can be added to the search matching set to obtain a search matching set that includes at least one historical perturbation trajectory similar to the target perturbation trajectory map.

[0118] Step 303: Based on the reference carbon emission trajectories corresponding to each historical disturbance trajectory in the search matching set, predict the carbon emission trajectory of the generator set and generate a carbon emission prediction trajectory.

[0119] For example, when a computer device selects multiple historical disturbance trajectories that are similar to the target disturbance trajectory map, it can predict the carbon emission trajectory of the generator set by weighted fusion of the reference carbon emission trajectories corresponding to each historical disturbance trajectory in the search matching set, and obtain the carbon emission prediction trajectory.

[0120] In this embodiment, a target disturbance trajectory map of the generator set for the current time period is constructed based on a parameter disturbance tensor sequence and a disturbance velocity sequence. Next, the target disturbance trajectory map is matched with a historical trajectory embedding library to determine a search matching set. This search matching set includes at least one historical disturbance trajectory with a similarity greater than or equal to a preset similarity to the target disturbance trajectory map. Then, based on the reference carbon emission trajectories corresponding to each historical disturbance trajectory in the search matching set, the carbon emission trajectory of the generator set is predicted, generating a carbon emission prediction trajectory. In other words, this method first determines similar historical disturbance trajectories through trajectory matching, then predicts carbon emissions based on the reference carbon emission trajectories corresponding to these similar historical disturbance trajectories, ultimately obtaining the carbon emission prediction trajectory. This method improves the accuracy of carbon emission prediction.

[0121] In an exemplary embodiment, before performing the similar trajectory search and matching in step 302 above, a historical trajectory embedding library is first established to ensure the high reliability and accuracy of the data used. For example... Figure 4 As shown, the process of obtaining the historical trajectory embedding library may include:

[0122] Step 401: Obtain the multi-source dataset.

[0123] The multi-source dataset includes at least two of the following: internal unit data, similar consortium data, and literature data. The multi-source dataset has a defined trust range.

[0124] Step 402: Based on the calibrated trust range, score the data of each data source in the multi-source dataset to obtain the data score results of each data source.

[0125] Step 403: Perform cross-authentication on each data source in the multi-source dataset to obtain the cross-authentication results of each data source.

[0126] Cross-certification can include at least one of the following: similarity comparison verification, response pattern consistency verification, multi-source redundancy support verification, and anomaly point reverse verification.

[0127] Step 404: Determine the comprehensive score of each data source based on the data scoring results and cross-certification results of each data source.

[0128] Step 405: Based on the comprehensive score of each data source, clean and update each data source in the multi-source dataset to establish a historical trajectory embedding library.

[0129] For example, the first step is to acquire a multi-source dataset, which includes internal unit data, consortium data, and literature data. These data sources are diverse: internal unit data reflects the operating status of the generator unit itself; consortium data comes from other generator units with similar operating characteristics; and literature data originates from publicly available materials such as academic research and industry reports. To ensure data credibility, the multi-source dataset is assigned a defined trust range to measure data reliability.

[0130] Subsequently, data scoring based on a defined trust range is performed on the multi-source datasets, and cross-certification is conducted to generate cross-certification results. The cross-certification process includes at least one of the following: similarity comparison verification to ensure consistency between data from different sources; response pattern consistency verification to confirm whether the response patterns of different data match under the same conditions; multi-source redundancy support verification to enhance the credibility of the data through redundant information from multiple data sources; and outlier reverse verification to reverse-check outliers in the data to ensure their reasonableness.

[0131] Finally, based on the cross-certification results and the defined trust range, the multi-source dataset is cleaned and updated. This cleaning and updating process removes low-quality and inaccurate data, ensuring the integrity and accuracy of the data in the historical trajectory embedding library. The cleaned dataset becomes part of the historical trajectory embedding library. After these steps, the data in the historical trajectory embedding library is not only of high quality but also has high reliability, providing reliable data support for subsequent similar trajectory searches.

[0132] For example, step 402 above, which scores data from each data source in the multi-source dataset based on the calibrated trust range, to obtain the data scoring results for each data source, may include: evaluating the completeness of each data source in the multi-source dataset to generate a first scoring feature for each data source; evaluating the time freshness of each data source in the multi-source dataset to generate a second scoring feature for each data source; using the calibrated trust range as a third scoring feature for each data source; and scoring each data source based on its first, second, and third scoring features to obtain the data scoring results for each data source.

[0133] Specifically, data scoring based on a defined trust range can include three main scoring features: data completeness, time freshness, and the defined trust range. Each feature represents a different dimension of data quality, and a comprehensive evaluation of these scoring features can more fully reflect the credibility of the dataset.

[0134] First, a data completeness assessment is performed on the multi-source dataset. Data completeness assessment involves examining each data source in the multi-source dataset to evaluate the completeness of its data, including the presence of missing values ​​and the continuity of data records. During the data scoring process, the first step is to assess the extent of missing data in each data source and its impact on the overall analysis, generating a primary scoring feature. This score quantifies the effectiveness of each data source based on its completeness. For example, a data source with significant missing data receives a lower score, while a complete data source with no missing data receives a higher score.

[0135] Next, a time freshness evaluation is performed. Time freshness evaluation assesses the timeliness of the data, ensuring it is up-to-date and effectively reflects the current system status. In dynamic operating environments, older data has limited reference value, especially when generator operating parameters change frequently; using older data may not accurately reflect current operating trends. Therefore, a secondary scoring feature is generated based on the interval between the data acquisition time and the current time. The shorter the time interval, the higher the score; the longer the time interval, the lower the score.

[0136] Finally, the calibrated trust range is used as the third scoring feature. The calibrated trust range is a quantitative indicator of the data source's reliability. Each data source in the dataset has a calibrated trust range, representing its reliability within a certain range. A wider trust range for a data source results in a higher score. Conversely, a narrower trust range indicates potentially greater uncertainty and a lower score.

[0137] After generating these three scoring features, a comprehensive data scoring is performed based on the first, second, and third scoring features. This comprehensive score takes into account the data's completeness, freshness, and reliability within a trust range. For example, different weights are assigned to each scoring feature according to its importance, and then the scores are weighted to assign a comprehensive score to each data source. This score ultimately determines which data sources can be used as trustworthy inputs and which data sources need further verification or are removed.

[0138] This scoring process comprehensively assesses the quality and credibility of the data, ensuring that the data used is complete, fresh, and highly credible, thereby providing high-quality foundational data for subsequent historical trajectory embedding libraries and similar trajectory search and matching.

[0139] After establishing the historical trajectory embedding library, a target perturbation trajectory map is constructed using the parameter perturbation tensor sequence and perturbation velocity sequence. A similar trajectory search and matching operation is then performed using the historical trajectory embedding library. Based on the similarity comparison of historical perturbation trajectories, historical perturbation trajectories similar to the current target perturbation trajectory map can be identified, and a search matching set is established accordingly. This search matching set will provide important reference for subsequent carbon emission prediction, ensuring the accuracy and reliability of the entire carbon emission accounting method.

[0140] In one exemplary embodiment, an optional implementation scheme is provided for weighted fusion of reference carbon emission trajectories corresponding to each historical perturbation trajectory in the search matching set. For example... Figure 5 As shown, step 303 above may include steps 501 to 502. Wherein:

[0141] Step 501: Determine the fusion weights corresponding to each historical disturbance trajectory in the search matching set.

[0142] For example, based on the similarity between each historical disturbance trajectory and the target disturbance trajectory map in the search matching set, a corresponding fusion weight can be assigned to each historical disturbance trajectory. The higher the similarity, the larger the corresponding fusion weight, and the lower the similarity, the smaller the corresponding fusion weight.

[0143] For example, multiple evaluation metrics can be used to score each historical disturbance trajectory in the search matching set using multidimensional metrics. Based on the score, a corresponding fusion weight is assigned to each historical disturbance trajectory. The higher the score, the greater the corresponding fusion weight, and the lower the score, the smaller the corresponding fusion weight.

[0144] In one optional implementation, the process of allocating fusion weights using a multi-dimensional index scoring method may include: using multiple evaluation indicators to quantitatively evaluate the similarity between each historical perturbation trajectory in the search matching set and the target perturbation trajectory map, generating a similarity score vector, which includes the similarity score value between each historical perturbation trajectory and the target perturbation trajectory map; then, based on the similarity score vector, determining the fusion weight corresponding to each historical perturbation trajectory in the search matching set, where the fusion weight is directly proportional to the corresponding similarity score value, that is, the higher the similarity, the larger the corresponding fusion weight, and the lower the similarity, the smaller the corresponding fusion weight.

[0145] The evaluation indicators may include at least two of the following: DTW matching distance indicator, disturbance direction consistency indicator, and anomaly overlap indicator. Each indicator focuses on different aspects of the evaluation to ensure a comprehensive assessment of the disturbance trajectory.

[0146] For example, for each historical disturbance trajectory in the search matching set, the above-mentioned multiple evaluation indicators are used to quantitatively evaluate the similarity between the historical disturbance trajectory and the target disturbance trajectory map, and the evaluation value corresponding to each evaluation indicator is obtained. Then, the evaluation values ​​corresponding to each evaluation indicator are weighted and summed to obtain the comprehensive evaluation value. The comprehensive evaluation value is the similarity score value after quantitatively evaluating the similarity between the historical disturbance trajectory and the target disturbance trajectory map.

[0147] Furthermore, based on the similarity score between the historical disturbance trajectory and the target disturbance trajectory map, a corresponding fusion weight can be assigned to the historical disturbance trajectory. For example, the similarity score can be normalized, and the normalized value between 0 and 1 obtained after normalization can be used as the corresponding fusion weight assigned to the historical disturbance trajectory. The magnitude of the fusion weight is proportional to the similarity score.

[0148] Step 502: Based on the fusion weights corresponding to each historical disturbance trajectory in the search matching set, the reference carbon emission trajectories corresponding to each historical disturbance trajectory in the search matching set are weighted and fused to generate carbon emission prediction trajectories.

[0149] For example, for a historical trajectory embedding library, a set of historical disturbance-carbon emission trajectory pairs corresponding to the historical trajectory embedding library can be constructed. This set includes multiple historical trajectories, each carrying disturbance data and corresponding carbon emission data. These historical trajectories can reflect the carbon emission situation of the generator set under different disturbance conditions. If multiple historical disturbance trajectories similar to the target disturbance trajectory map are identified, a reference carbon emission trajectory corresponding to each similar historical disturbance trajectory can be further determined from the set of historical disturbance-carbon emission trajectory pairs. Based on this, the reference carbon emission trajectories corresponding to each historical disturbance trajectory are weighted and fused to generate the final carbon emission prediction trajectory.

[0150] It should be noted that the key to carbon emission prediction lies in how to generate a more accurate carbon emission prediction by combining the similarity and weight of historical data with the perturbations and carbon emission information in different historical trajectories.

[0151] For example, the reference carbon emission trajectories corresponding to each historical disturbance trajectory can be multiplied by their respective fusion weights, and then all weighted reference carbon emission trajectories can be aggregated and merged to obtain the final carbon emission prediction trajectory. This carbon emission prediction trajectory not only reflects the carbon emission trend under the current operating state of the generator unit, but also provides a scientific basis for subsequent carbon emission management and emission reduction strategy formulation.

[0152] In this embodiment, a corresponding fusion weight is assigned to each historical disturbance trajectory in the search matching set, and a weighted fusion process is performed on the reference carbon emission trajectories corresponding to each historical disturbance trajectory based on the fusion weight to generate a carbon emission prediction trajectory. This method can further clarify the important contribution of each historical disturbance trajectory in the carbon emission prediction process and improve the accuracy of carbon emission prediction.

[0153] In one exemplary embodiment, an optional implementation scheme for multidimensional index scoring is provided. For example... Figure 6 As shown, the above steps employ multiple evaluation metrics to quantitatively assess the similarity between each historical disturbance trajectory in the search matching set and the target disturbance trajectory map, generating a similarity score vector. This may include steps 601 to 604. Wherein:

[0154] Step 601: Perform multi-scale segmentation on the target disturbance trajectory map to obtain the segmentation results at each segmentation scale.

[0155] The segmentation results include multiple perturbation trajectory sub-maps corresponding to the target perturbation trajectory map at the corresponding segmentation scale.

[0156] For example, a multi-scale segmentation scheme can be obtained first, and then the target perturbation trajectory map can be segmented according to each scale segmentation scheme to obtain the segmentation result at each segmentation scale. For example, a multi-scale aggregation analysis of graph structure under parameter perturbation tensor sequence and perturbation velocity sequence can be performed on the target perturbation trajectory map to configure a multi-scale segmentation scheme; then, the target perturbation trajectory map can be segmented at multiple scales according to the multi-scale segmentation scheme to obtain the segmentation result at each segmentation scale.

[0157] To ensure the accuracy of the multi-dimensional index scoring, a multi-scale graph structure aggregation analysis was first performed on the target perturbation trajectory map under parametric perturbation tensor sequences and perturbation velocity sequences. The core objective of this analysis is to capture the features of the target perturbation trajectory map at different levels of detail through segmentation at different scales, thereby more comprehensively evaluating the similarity of the trajectories. It should be noted that the configuration of the multi-scale segmentation scheme needs to be based on the specific characteristics of the target perturbation trajectory map, reasonably dividing different scale levels so that the local and global features of the trajectory can be fully considered in the subsequent scoring process.

[0158] Step 602: For each historical disturbance trajectory in the search matching set, use each evaluation index to determine the fusion matching score between the historical disturbance trajectory and the segmentation results at each segmentation scale.

[0159] After completing the multi-scale aggregation analysis, a multi-dimensional index scoring process is performed on each segmentation result and the search matching set in the multi-scale segmentation scheme. Taking the multi-dimensional indices, including DTW matching distance, parameter perturbation direction consistency, and outlier overlap, as an example, the multi-dimensional index scoring process may include:

[0160] The DTW (Dynamic Time Warping) matching distance evaluation channel is activated, and the DTW matching distance metric is used to determine the first fusion matching score between historical perturbation trajectories and segmentation results at each segmentation scale. DTW is a commonly used time series matching algorithm that can calculate the similarity between two time series. Here, the DTW matching distance is used to evaluate the temporal matching degree between each segmentation result and historical perturbation trajectories in the search matching set. By performing multi-path fusion matching scoring on each segmentation result, a first fusion matching score is generated. The first fusion matching score can measure the temporal similarity between the target perturbation trajectory map and historical perturbation trajectories, helping to identify the historical perturbation trajectory that best matches the current perturbation behavior.

[0161] The perturbation direction evaluation channel is activated, and a perturbation direction consistency index is used to determine the second fusion matching score between historical perturbation trajectories and segmentation results at each segmentation scale. The perturbation direction consistency index is used to evaluate the perturbation direction similarity and average velocity consistency between each segmentation result and the search matching set. Perturbation direction consistency refers to whether the perturbation direction is consistent across different trajectories, while average velocity consistency measures whether the rates of perturbation change are similar. By scoring these parameters, a second fusion matching score can be generated. The second fusion matching score reflects the similarity between the target perturbation trajectory map and historical perturbation trajectories in terms of perturbation direction and velocity, and is crucial for evaluating the similarity of perturbation patterns.

[0162] The anomaly overlap channel is activated, and the anomaly overlap index is used to determine the third fusion matching score between historical disturbance trajectories and segmentation results at each segmentation scale. For example, the third fusion matching score is obtained by matching the local structure of anomalies in each segmentation result with the historical disturbance trajectory. The anomaly overlap index is used to identify the matching of anomalies in the trajectory. Through local structure matching calculation, the similarity of anomalies can be evaluated more accurately. Anomalies may be caused by system failures or other special circumstances; therefore, the similarity of anomalies can reveal the similarity between the two under fault or abnormal states.

[0163] Step 603: Based on the fusion matching scores corresponding to each evaluation index, determine the similarity score between the historical disturbance trajectory and the target disturbance trajectory map.

[0164] For example, the similarity score between the historical perturbation trajectory and the target perturbation trajectory map can be determined based on the first, second, and third fusion matching scores corresponding to the historical perturbation trajectory, thus establishing a multi-dimensional index score. By comprehensively considering multiple dimensions of indicators such as DTW matching distance, consistency of parameter perturbation direction, and overlap of outliers, the similarity between the perturbation trajectory map and the search matching set can be assessed more comprehensively and accurately. This multi-dimensional index score result will provide a reliable basis for establishing a set of historical perturbation-carbon emission trajectory pairs, thereby providing accurate support for subsequent carbon emission prediction and accounting.

[0165] For example, the similarity score between the historical disturbance trajectory and the target disturbance trajectory map can be obtained by weighted summing of the first fusion matching score, the second fusion matching score and the third fusion matching score corresponding to the historical disturbance trajectory.

[0166] Step 604: Generate a similarity score vector based on the similarity score values ​​of each historical disturbance trajectory and the target disturbance trajectory map.

[0167] For example, the similarity score values ​​of each historical disturbance trajectory and the target disturbance trajectory map can be inserted into the score vector in the order of their arrangement in the search trajectory set to obtain the similarity score vector.

[0168] In this embodiment, the target perturbation trajectory map is first segmented at multiple scales to obtain segmentation results at multiple scales. Then, each segmentation result is compared with each historical perturbation trajectory in the search matching set using multi-dimensional index scoring. Specifically, multi-dimensional index scoring (such as DTW matching distance, parameter perturbation direction consistency, and outlier overlap) is used to quantitatively evaluate the similarity between each historical perturbation trajectory in the search matching set and the current target perturbation trajectory, generating a similarity score vector. This similarity score vector reflects the degree of similarity between the current target perturbation trajectory and each historical perturbation trajectory; a higher value indicates greater similarity, and a lower value indicates less similarity. This method, through multi-dimensional index scoring, can evaluate the quality of each historical perturbation trajectory matched in the search, improving the accuracy of the similarity measurement results for the search matches, and thus improving the accuracy of carbon emission prediction.

[0169] In one exemplary embodiment, a process for constructing a multi-scale segmentation scheme is provided. For example... Figure 7 As shown, the steps for obtaining the multi-scale segmentation scheme described above may include steps 701 to 704. Wherein:

[0170] Step 701: Establish a joint perturbation gradient sequence based on the parameter perturbation tensor sequence and the perturbation velocity sequence.

[0171] For example, the joint perturbation gradient sequence can be established by jointly analyzing the parameter perturbation tensor sequence and the perturbation velocity sequence to calculate the gradient information at each time point. This gradient information reflects the rate and direction of change of the perturbation in time and space, and is the basis for subsequent multi-scale segmentation. For instance, the joint perturbation gradient sequence can be obtained by performing difference calculations on the parameter perturbation tensor sequence and the perturbation velocity sequence, and by calculating the gradient change at each time point, the joint perturbation gradient sequence is obtained.

[0172] Step 702: Based on the multi-scale smoothing filter, extract the perturbation joint gradient sequence using a multi-resolution view to obtain multiple perturbation joint gradient views at different resolutions.

[0173] For example, a multi-scale smoothing filter can be used to extract multi-resolution views of the perturbation joint gradient sequence based on multiple smoothing kernel widths. A multi-scale smoothing filter is a tool capable of handling features at different scales. By selecting different smoothing kernel widths, views of the perturbation joint gradient sequence at different resolutions can be extracted. The choice of smoothing kernel width can be adjusted according to the characteristics of the perturbation trajectory plot and the analysis requirements; for example, a smaller kernel width can be selected to capture high-frequency features, and a larger kernel width can be selected to capture low-frequency features. Through smoothing, multiple perturbation joint gradient views at different resolutions can be obtained.

[0174] Step 703: Under the joint gradient view of perturbation at each resolution, the joint gradient sequence of perturbation is segmented based on the local peak value of perturbation intensity to obtain a segmented structure of multiple perturbation trajectory subgraphs under the joint gradient view of each resolution.

[0175] For example, at each resolution of the joint gradient view of perturbations, trajectory submaps can be segmented based on local peak clustering of perturbation intensity. Local peaks refer to points where the perturbation intensity reaches its maximum value within a certain range; these points typically correspond to key change points in the trajectory. Through local peak clustering, the joint gradient sequence of perturbations can be divided into multiple segments, each corresponding to a local peak and its surrounding region. Based on the results of local peak clustering, the joint gradient sequence of perturbations is segmented to construct multiple segmented structures of perturbation trajectory submaps; each segmented structure contains perturbation information from one local peak to the next, and these segmented structures can reflect the local features of the target perturbation trajectory map in greater detail.

[0176] Step 704: Based on the segmentation structure of multiple perturbation trajectory subgraphs under the joint gradient view of the resolution, determine the scale segmentation scheme corresponding to the resolution.

[0177] For example, based on the segmentation structure of multiple perturbation trajectory subgraphs under a joint gradient view of perturbations at a given resolution, a multi-scale segmentation scheme can be configured, namely, a scale segmentation scheme corresponding to that resolution. This scale segmentation scheme can automatically segment the perturbation trajectory according to its characteristics, enabling more accurate analysis of the behavioral characteristics of each perturbation segment and providing a flexible and sophisticated tool for subsequent trajectory matching, analysis, and evaluation.

[0178] In this embodiment, a multi-scale segmentation scheme is configured by performing multi-scale aggregation analysis of the graph structure under parameter perturbation tensor sequence and perturbation velocity sequence on the target perturbation trajectory map. The multi-scale aggregation analysis is based on multi-scale smoothing filter to extract multi-resolution view of the perturbation joint gradient sequence, and further performs local peak clustering and sequence splitting on the perturbation joint gradient sequence based on the multi-resolution view. Using this method, multiple segmentation schemes with different segmentation scales can be configured. Each scale segmentation scheme can more accurately segment and characterize the behavioral characteristics of each perturbation segment at different resolution scales, providing more accurate data support for subsequent trajectory matching, analysis and evaluation, and further improving the accuracy of carbon emission trajectory prediction.

[0179] In one exemplary embodiment, an optional implementation scheme for backtracking and correcting carbon emission prediction trajectories is provided. For example... Figure 8 As shown, the above method may further include steps 801 to 803. Wherein:

[0180] Step 801: Monitor the actual carbon emissions of the generator set and generate the actual carbon emission trajectory.

[0181] For example, computer equipment can perform actual emission monitoring of generator sets and establish an actual emission dataset based on the monitoring results. Then, based on the actual emission dataset, a carbon emission trajectory can be constructed. Actual emission data is collected by monitoring the carbon emissions of generator sets in real time. This data can accurately reflect the carbon emission levels of generator sets under different operating conditions, providing verification of the predicted trajectory based on actual conditions.

[0182] It should be noted that the carbon emission prediction trajectory can include carbon emission predictions for a preset period following the current time period. Within this preset period, the actual emissions of the generator units can be continuously monitored, and the carbon emission prediction trajectory can be corrected in real time based on the actual carbon emission trajectory. The aforementioned actual emission dataset should contain a sufficient time span and data points within the preset period to ensure a comprehensive reflection of the generator units' true emissions.

[0183] Step 802: Perform a deviation analysis on the actual carbon emission trajectory and the predicted carbon emission trajectory to determine the deviation residual.

[0184] For example, a bias analysis can be performed on the actual emissions dataset and the carbon emission prediction trajectory, and a bias residual can be established based on the analysis results. For instance, the actual emissions values ​​in the actual carbon emission trajectory (i.e., the actual emissions dataset) can be compared with the predicted emissions values ​​in the carbon emission prediction trajectory based on historical perturbation data, and the difference or bias between the two can be calculated. For example, this can be calculated by subtracting the predicted emissions value from the actual emissions value; a positive bias indicates that the actual emissions are higher than the predicted value, and a negative bias indicates that the actual emissions are lower than the predicted value. By analyzing the residuals, potential sources of error or bias in carbon emission prediction can be identified, helping to determine whether the bias in the prediction model is systematic or due to random errors caused by special factors (such as equipment failure, environmental changes, etc.). The results of the residual analysis can provide a basis for subsequent backtracking corrections.

[0185] Step 803: Correct the portion of the carbon emission prediction trajectory after the current time based on the deviation residual to generate the corrected carbon emission prediction trajectory.

[0186] For example, computer equipment can use the deviation residuals as correction data to perform retrospective correction of carbon emission predictions. Through retrospective correction, the model's prediction parameters or the prediction results can be adjusted to reduce the deviation between actual and predicted emissions. For example, the deviation residuals can be used to adjust weights in the historical trajectory matching process, correct the correlation pattern between disturbances and carbon emissions, or optimize other parameters of the prediction model. Furthermore, the carbon emission prediction trajectory can be directly corrected by adding the deviation residuals back into the carbon emission prediction trajectory, thereby generating a corrected carbon emission prediction trajectory that more closely approximates actual emissions. The corrected carbon emission prediction trajectory can more accurately reflect the actual operating conditions of the generator set, improving the reliability and practicality of the prediction.

[0187] For example, the preset duration can be divided into a first duration and a second duration. The first duration represents the actual operating duration, i.e., the duration corresponding to the actual carbon emission trajectory, while the second duration represents the non-operating duration, i.e., the remaining duration excluding the first duration. During backtracking correction, the portion of the predicted trajectory corresponding to the second duration in the carbon emission prediction trajectory can be corrected based on the deviation residual to generate a corrected carbon emission prediction trajectory. The corrected carbon emission prediction trajectory may include only the corrected prediction trajectory corresponding to the second duration, or it may include both the uncorrected prediction trajectory corresponding to the first duration and the corrected prediction trajectory corresponding to the second duration.

[0188] In this embodiment, carbon emissions from the generator set are monitored in real time, and the predicted carbon emission trajectory is corrected based on actual emission data to achieve retrospective correction of carbon emission prediction. Since the corrected carbon emission prediction trajectory more accurately reflects the actual operating conditions of the generator set, the reliability and practicality of carbon emission prediction are improved. In other words, by introducing actual emission monitoring data and performing deviation analysis and retrospective correction on the prediction results, the accuracy and reliability of carbon emission prediction can be further improved.

[0189] Existing carbon emission monitoring and prediction systems typically cannot dynamically analyze changes in generator sets under different operating conditions, thus hindering accurate carbon emission optimization. In one exemplary embodiment, an optional implementation scheme is provided for real-time optimization and adjustment of generator set operating parameters based on carbon emission prediction. Figure 9 As shown, the above method may further include steps 901 to 903. Wherein:

[0190] Step 901: Based on the set of operating parameters and the carbon emission prediction trajectory, construct a joint optimization objective function.

[0191] In this example, the main purpose of the joint optimization objective function is to optimize the generator set's operating strategy by comprehensively considering the predicted carbon emission trajectory and actual operating parameter data. This is achieved by combining the carbon emission prediction results with operating parameters (such as power output, load, and speed) to form a comprehensive optimization model. In this model, the carbon emission level is the key optimization indicator, while other constraints such as the generator set's economy, stability, and safety must also be considered. Through this joint optimization objective function, the optimal combination of operating parameters can be found, minimizing carbon emissions while meeting energy production needs. The specific setting of this joint optimization objective function will depend on the requirements of the industrial control stage and may involve multiple constraints and optimization objectives, such as emission control limits, fuel consumption, and equipment load.

[0192] Step 902: Determine the balance control parameters of the generator set based on the joint optimization objective function.

[0193] For example, by calculating the joint optimization objective function, balanced control parameters containing the optimal values ​​of various control variables during the operation of the generator set can be obtained. Based on these optimal balanced control parameters, the generator set's operation control system can adjust its operation mode in real time to ensure that the generator set maximizes its operating efficiency while meeting the carbon emission prediction target.

[0194] Step 903: Perform operation control of the generator set based on the balance control parameters.

[0195] For example, the computer equipment can send the determined balance control parameters to the generator set to control the generator set to adjust the balance control parameters in real time, and then perform power generation operation based on the adjusted balance control parameters.

[0196] In this embodiment, after generating the carbon emission prediction trajectory, the carbon emission prediction can be further combined with the operation control of the generator set to optimize the performance of the generator set, achieve effective management of carbon emissions, and improve the overall operating performance and efficiency of the generator set.

[0197] In one exemplary embodiment, an optional implementation scheme for providing anomaly warnings during actual operation based on carbon emission prediction trajectories is provided. For example... Figure 10 As shown, the above method may further include steps 1001 to 1003. Wherein:

[0198] Step 1001: Perform data mapping between the operating parameter set and the carbon emission prediction trajectory, and store the correlation between the data-mapped operating parameter set and the carbon emission prediction trajectory in a preset data recording space.

[0199] For example, data mapping is performed on the carbon emission prediction trajectory and the set of operating parameters. This includes mapping the previously generated carbon emission prediction trajectory with the actually collected operating parameters and storing this data in a data recording space. The purpose of data mapping is to ensure that the correlation between the carbon emission prediction trajectory and the actual operating parameters can be clearly recorded and easily accessed and queried in subsequent analyses. The data recording space is a database or storage system used for long-term storage of generator set operating data and carbon emission prediction data. In this way, all historical data will be uniformly organized and saved, facilitating easy retrieval for subsequent analysis, verification, and optimization.

[0200] Step 1002: Based on the correlation stored in the preset data recording space, continuously monitor the working status of the generator set.

[0201] For example, recorded data from the data recording space is invoked to perform continuous operation data anomaly verification for the generator set. The core of this process is to continuously monitor carbon emission data and operating parameters during generator set operation, and promptly detect any potential anomalies in the data. Anomaly verification can be performed by comparing historical and current data. For instance, if the actual carbon emission value of the generator set deviates significantly from the predicted value, or if operating parameters suddenly fluctuate abnormally, it may indicate a system malfunction or operational anomaly. In this way, anomalies in the data can be detected in real time, avoiding inaccurate carbon emission predictions due to data distortion or system failures.

[0202] Step 1003: If an abnormal operating status of the generator set is detected, an abnormal data warning is generated based on the abnormal data.

[0203] For example, based on the results of data anomaly verification, if data anomalies are detected, the computer device can output a data anomaly warning. Warnings can be output in various ways, such as through alarm systems, SMS notifications, or system logs, to promptly notify relevant personnel to take appropriate measures.

[0204] In this embodiment, after generating the carbon emission prediction trajectory, the carbon emissions and operating parameters during the generator set operation can be continuously monitored, and any potential anomalies can be detected and warned in a timely manner. By promptly detecting and handling data anomalies, the generator set can operate in a more efficient and stable mode, while reducing carbon emission errors caused by anomalies, thus supporting more accurate carbon emission management.

[0205] In the above embodiments, by collecting and analyzing the operating parameters of the generator set in real time, and combining the disturbance trajectory map with historical trajectory matching, accurate prediction of carbon emissions is achieved, overcoming the low accuracy and lag problems of traditional methods. In addition, based on the parameter disturbance tensor sequence and disturbance velocity sequence, the operating status of the generator set can be dynamically monitored, providing real-time carbon emission monitoring and early warning. Furthermore, through multi-dimensional index scoring and similarity weighted fusion, accurate carbon emission prediction trajectories are generated, optimizing carbon emission control and power generation efficiency. This not only achieves the technical effect of improving the accuracy and real-time performance of carbon emission accounting through generator set parameter disturbance analysis and multi-dimensional index scoring, but also realizes real-time detection and early warning of abnormal data of the generator set, ensuring the stable operation of the generator equipment and avoiding carbon emission exceeding the standard or efficiency decline.

[0206] In one exemplary embodiment, a complete implementation of a method for calculating carbon dioxide emissions based on generator sets is provided, such as... Figure 11 As shown, it includes the following steps:

[0207] Step 1: Collect the operating parameters of the generator set at fixed time intervals, establish the operating parameter set, perform parameter change rate analysis on the operating parameter set within a predetermined sliding time window, and construct the parameter perturbation tensor sequence.

[0208] Step 2: After calculating the rate of change between adjacent parameter perturbation tensors in the parameter perturbation tensor sequence, establish the perturbation velocity sequence based on the calculation results. The perturbation velocity sequence is used to characterize the dynamic trend of the perturbation trajectory.

[0209] Step 3: After establishing the perturbation trajectory map using the parameter perturbation tensor sequence and perturbation velocity sequence, perform similar trajectory search and matching from the historical trajectory embedding library to establish a search matching set.

[0210] Step 4: Perform multi-dimensional index scoring on the perturbation trajectory map of the search matching set, and establish a similarity score vector and a set of historical perturbation-carbon emission trajectory pairs; the multi-dimensional index scoring includes DTW matching distance index, parameter perturbation direction consistency index, and outlier overlap index.

[0211] Step 5: Based on the historical perturbation-carbon emission trajectory set and similarity score vector, perform weighted fusion to generate carbon emission prediction trajectory.

[0212] This method involves collecting and analyzing generator unit operating parameters at fixed time intervals, calculating parameter change rates, constructing parameter perturbation tensor sequences and perturbation velocity sequences, and then establishing a perturbation trajectory map. It then performs similar trajectory search and matching from a historical trajectory embedding library and multi-dimensional index scoring to establish a set of historical perturbation-carbon emission trajectory pairs. Finally, it performs weighted fusion based on the historical perturbation-carbon emission trajectory pair set and the similarity score vector to generate a carbon emission prediction trajectory. The method proposed in this application solves the technical problem that existing carbon emission accounting methods cannot accurately reflect the dynamic changes in generator unit operating status, leading to insufficient accuracy and real-time performance in carbon emission accounting. It achieves the technical effect of improving the accuracy and real-time performance of carbon emission accounting through generator unit parameter perturbation analysis and multi-dimensional index scoring.

[0213] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0214] Based on the same inventive concept, this application also provides a generator set-based carbon dioxide emission accounting device for implementing the above-described generator set-based carbon dioxide emission accounting method. The solution provided by this device is similar to the solution described in the above-described method. Therefore, the specific limitations of one or more generator set-based carbon dioxide emission accounting device embodiments provided below can be found in the limitations of the generator set-based carbon dioxide emission accounting method described above, and will not be repeated here.

[0215] In one exemplary embodiment, such as Figure 12As shown, a carbon dioxide emission accounting device 1200 based on a generator set is provided, including: a parameter acquisition module 1201, a disturbance construction module 1202, and a carbon emission prediction module 1203, wherein:

[0216] The parameter acquisition module 1201 is used to acquire the set of operating parameters of the generator set in the current time period; the set of operating parameters includes the operating data of multiple operating parameters in the current time period.

[0217] The disturbance construction module 1202 is used to construct a parameter disturbance tensor sequence and a disturbance velocity sequence based on the set of operating parameters. The parameter disturbance tensor sequence is used to characterize the disturbance situation of the generator set at different time points, and the disturbance velocity sequence is used to characterize the dynamic trend of the disturbance trajectory.

[0218] The carbon emission prediction module 1203 is used to predict the carbon emission trajectory of the generator set based on the parameter perturbation tensor sequence, perturbation velocity sequence, historical trajectory embedding library and reference carbon emission trajectory corresponding to historical perturbation trajectory, and generate carbon emission prediction trajectory.

[0219] In one embodiment, the perturbation construction module 1202 includes:

[0220] The perturbation tensor construction unit is used to perform parameter change rate analysis on each running parameter in the running parameter set using a preset sliding time window and a preset moving step size, and generate a parameter perturbation tensor sequence. The parameter perturbation tensor sequence includes parameter perturbation tensors for multiple time windows within the current time period, and the parameter perturbation tensor includes the change rate of multiple running parameters within the corresponding time window.

[0221] The perturbation velocity construction unit is used to perform rate of change analysis on adjacent parameter perturbation tensors in the parameter perturbation tensor sequence to generate a perturbation velocity sequence.

[0222] In one embodiment, the carbon emission prediction module 1203 includes:

[0223] The disturbance trajectory construction unit is used to construct the target disturbance trajectory map of the generator set in the current time period based on the parameter disturbance tensor sequence and the disturbance velocity sequence;

[0224] The trajectory matching unit is used to match the target disturbance trajectory map with the historical trajectory embedding library to determine the search matching set; the search matching set includes at least one historical disturbance trajectory with a similarity greater than or equal to the target disturbance trajectory map.

[0225] The carbon emission prediction unit is used to predict the carbon emission trajectory of the generator set based on the reference carbon emission trajectory corresponding to each historical disturbance trajectory in the search matching set, and generate the carbon emission prediction trajectory.

[0226] In one embodiment, the carbon emission prediction unit is specifically used to determine the fusion weights corresponding to each historical disturbance trajectory in the search matching set; based on the fusion weights corresponding to each historical disturbance trajectory in the search matching set, the reference carbon emission trajectories corresponding to each historical disturbance trajectory in the search matching set are weighted and fused to generate carbon emission prediction trajectories.

[0227] In one embodiment, the carbon emission prediction unit is specifically used to quantitatively evaluate the similarity between each historical disturbance trajectory in the search matching set and the target disturbance trajectory map using multiple evaluation indicators, and generate a similarity score vector; the similarity score vector includes the similarity score value between each historical disturbance trajectory and the target disturbance trajectory map; based on the similarity score vector, the fusion weight corresponding to each historical disturbance trajectory in the search matching set is determined; the fusion weight is positively proportional to the corresponding similarity score value.

[0228] In one embodiment, the carbon emission prediction unit is specifically used to perform multi-scale segmentation on the target perturbation trajectory map to obtain segmentation results at each segmentation scale. The segmentation results include multiple perturbation trajectory sub-maps corresponding to the target perturbation trajectory map at the corresponding segmentation scale. For each historical perturbation trajectory in the search matching set, each evaluation index is used to determine the fusion matching score between the historical perturbation trajectory and the segmentation results at each segmentation scale. Based on the fusion matching score corresponding to each evaluation index, the similarity score value between the historical perturbation trajectory and the target perturbation trajectory map is determined. Based on the similarity score value between each historical perturbation trajectory and the target perturbation trajectory map, a similarity score vector is generated.

[0229] In one embodiment, the carbon emission prediction unit is specifically used to obtain a multi-scale segmentation scheme; based on the segmentation scheme of each scale, the target perturbation trajectory map is segmented to obtain the segmentation result at each segmentation scale.

[0230] In one embodiment, the carbon emission prediction unit is specifically used to establish a joint perturbation gradient sequence based on the parameter perturbation tensor sequence and the perturbation velocity sequence; extract multiple resolution views from the joint perturbation gradient sequence based on a multi-scale smoothing filter to obtain multiple joint perturbation gradient views at different resolutions; segment the joint perturbation gradient sequence based on the local peak values ​​of the perturbation intensity under each resolution joint perturbation gradient view to obtain multiple perturbation trajectory subgraph segmentation structures under the resolution joint perturbation gradient view; and determine the scale segmentation scheme corresponding to the resolution based on the multiple perturbation trajectory subgraph segmentation structures under the resolution joint perturbation gradient view.

[0231] In one embodiment, the multiple evaluation metrics include at least two of the following: DTW matching distance metric, perturbation direction consistency metric, and anomaly point overlap metric.

[0232] The various modules in the aforementioned generator set-based carbon dioxide emission accounting device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the computer device's memory as software, so that the processor can call and execute the corresponding operations of each module.

[0233] Based on the same inventive concept, this application also provides a generator set-based carbon dioxide emission accounting system for implementing the above-mentioned generator set-based carbon dioxide emission accounting method. The solution provided by this system is similar to the implementation scheme described in the above method. Therefore, the specific limitations of one or more generator set-based carbon dioxide emission accounting system embodiments provided below can be found in the limitations of the generator set-based carbon dioxide emission accounting method described above, and will not be repeated here.

[0234] For example, the module structure of the carbon dioxide emission accounting system based on generator sets can refer to the module structure of the carbon dioxide emission accounting device based on generator sets described above, and will not be repeated here.

[0235] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 13 As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs in the non-volatile storage media to run. The database stores generator set operating data and intermediate and final data from the carbon emission prediction process. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When executed by the processor, the computer program implements a method for calculating carbon dioxide emissions based on generator sets.

[0236] Those skilled in the art will understand that Figure 13 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0237] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the generator set-based carbon dioxide emission accounting method in any of the above embodiments.

[0238] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the generator set-based carbon dioxide emission accounting method in any of the above embodiments.

[0239] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps of the generator set-based carbon dioxide emission accounting method in any of the above embodiments.

[0240] It should be noted that the data involved in this application (including but not limited to data used for analysis, data stored, data displayed, etc.) are all information and data that have been fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0241] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0242] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0243] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for calculating carbon dioxide emissions based on generator sets, characterized in that, The method includes: Obtain the set of operating parameters for the generator set during the current time period; the set of operating parameters includes operating data of multiple operating parameters during the current time period. Based on the set of operating parameters, a parameter perturbation tensor sequence and a perturbation velocity sequence are constructed; the parameter perturbation tensor sequence is used to characterize the perturbation of the generator set at different time points, and the perturbation velocity sequence is used to characterize the dynamic trend of the perturbation trajectory. Based on the parameter perturbation tensor sequence and the perturbation velocity sequence, a target perturbation trajectory diagram of the generator set in the current time period is constructed; The target perturbation trajectory map and the historical trajectory embedding library are matched to determine a search matching set; the search matching set includes at least one historical perturbation trajectory with a similarity greater than or equal to a preset similarity with the target perturbation trajectory map; Based on the reference carbon emission trajectories corresponding to each historical disturbance trajectory in the search matching set, the carbon emission trajectory of the generator set is predicted to generate the carbon emission prediction trajectory.

2. The method according to claim 1, characterized in that, The construction of the parameter perturbation tensor sequence and perturbation velocity sequence based on the set of operating parameters includes: Using a preset sliding time window and a preset moving step size, parameter change rate analysis is performed on each of the operating parameters in the set of operating parameters to generate the parameter perturbation tensor sequence; the parameter perturbation tensor sequence includes parameter perturbation tensors for multiple time windows within the current time period, and the parameter perturbation tensor includes the change rate of multiple operating parameters within the corresponding time window; The rate of change of adjacent parameter perturbation tensors in the parameter perturbation tensor sequence is analyzed to generate the perturbation velocity sequence.

3. The method according to claim 1, characterized in that, The step of predicting the carbon emission trajectory of the generator set based on the reference carbon emission trajectories corresponding to each historical disturbance trajectory in the search matching set, and generating the carbon emission prediction trajectory, includes: Determine the fusion weights corresponding to each historical disturbance trajectory in the search matching set; Based on the fusion weights corresponding to each historical disturbance trajectory in the search matching set, the reference carbon emission trajectories corresponding to each historical disturbance trajectory in the search matching set are weighted and fused to generate the carbon emission prediction trajectory.

4. The method according to claim 3, characterized in that, Determining the fusion weights corresponding to each historical disturbance trajectory in the search matching set includes: Multiple evaluation metrics are used to quantitatively evaluate the similarity between each historical perturbation trajectory in the search matching set and the target perturbation trajectory map, generating a similarity score vector; the similarity score vector includes the similarity score value between each historical perturbation trajectory and the target perturbation trajectory map; Based on the similarity score vector, the fusion weight corresponding to each historical disturbance trajectory in the search matching set is determined; the fusion weight is positively proportional to the corresponding similarity score value.

5. The method according to claim 4, characterized in that, The method employs multiple evaluation metrics to quantitatively assess the similarity between each historical perturbation trajectory in the search matching set and the target perturbation trajectory map, generating a similarity score vector, including: The target perturbation trajectory map is segmented at multiple scales to obtain the segmentation result at each segmentation scale; the segmentation result includes multiple perturbation trajectory sub-maps corresponding to the target perturbation trajectory map at the corresponding segmentation scale; For each historical perturbation trajectory in the search matching set, the evaluation indexes are used to determine the fusion matching score between the historical perturbation trajectory and the segmentation results at each segmentation scale. Based on the fusion matching score corresponding to each of the evaluation indicators, the similarity score between the historical disturbance trajectory and the target disturbance trajectory map is determined. The similarity score vector is generated based on the similarity score between each of the historical perturbation trajectories and the target perturbation trajectory map.

6. The method according to claim 5, characterized in that, The multi-scale segmentation of the target disturbance trajectory map, to obtain the segmentation result at each segmentation scale, includes: Obtain multi-scale segmentation schemes; Based on the segmentation schemes described above, the target perturbation trajectory map is segmented to obtain the segmentation result at each segmentation scale.

7. The method according to claim 6, characterized in that, The acquisition of the multi-scale segmentation scheme includes: Based on the parameter perturbation tensor sequence and the perturbation velocity sequence, a joint perturbation gradient sequence is established; Based on a multi-scale smoothing filter, the perturbation joint gradient sequence is extracted using a multi-resolution view to obtain multiple perturbation joint gradient views at different resolutions. Under the joint gradient view of the perturbation at each resolution, the joint gradient sequence of the perturbation is segmented based on the local peak value of the perturbation intensity to obtain a segmented structure of multiple perturbation trajectory subgraphs under the joint gradient view of the perturbation at each resolution; Based on the segmented structure of multiple perturbation trajectory subgraphs under the perturbation joint gradient view at the specified resolution, the scale segmentation scheme corresponding to the specified resolution is determined.

8. The method according to claim 4, characterized in that, The multiple evaluation indicators include at least two of the following: DTW matching distance indicator, disturbance direction consistency indicator, and anomaly point overlap indicator.

9. A carbon dioxide emission accounting device based on a generator set, characterized in that, The device includes: The parameter acquisition module is used to acquire the set of operating parameters of the generator set in the current time period; the set of operating parameters includes the operating data of multiple operating parameters in the current time period. The disturbance construction module is used to construct a parameter disturbance tensor sequence and a disturbance velocity sequence based on the set of operating parameters; the parameter disturbance tensor sequence is used to characterize the disturbance situation of the generator set at different time points, and the disturbance velocity sequence is used to characterize the dynamic trend of the disturbance trajectory; The carbon emission prediction module includes: The disturbance trajectory construction unit is used to construct the target disturbance trajectory map of the generator set in the current time period based on the parameter disturbance tensor sequence and the disturbance velocity sequence; A trajectory matching unit is used to match the target disturbance trajectory map with the historical trajectory embedding library to determine a search matching set; the search matching set includes at least one historical disturbance trajectory with a similarity greater than or equal to a preset similarity to the target disturbance trajectory map; The carbon emission prediction unit is used to predict the carbon emission trajectory of the generator set based on the reference carbon emission trajectory corresponding to each historical disturbance trajectory in the search matching set, and generate the carbon emission prediction trajectory.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 8.

Images