Thermal power generation carbon emission simulation modeling method and system based on multi-source big data

Abstract

The invention discloses a thermal power generation carbon emission simulation modeling method and system based on multi-source big data, and relates to the technical field of electric carbon emission calculation, and the method comprises the steps: obtaining thermal power generation data, carrying out the preprocessing of the data, building an electric carbon coupling refined model based on the preprocessed data, and carrying out the simulation modeling of the thermal power generation carbon emission. And generating a dynamic coupling relationship between equipment through time sequence feature mapping and node information aggregation, performing real-time electricity-energy-carbon modeling on coal-electricity and gas-electricity type thermal power generation enterprises under all working conditions, distinguishing data relationships under different operation working conditions, and constructing a hierarchical correlation model. According to the method, a gating mechanism of a graph neural network is introduced into electric carbon coupling modeling, the problem of information screening in dynamic interaction of equipment is solved, high-precision prediction covered by all working conditions is realized through working condition feature extraction and residual iteration, and an integrated carbon management scheme from prediction to regulation is provided for thermal power generation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electric carbon emission calculation, and particularly relates to a thermal power carbon emission simulation modeling method and system based on multi-source big data. BACKGROUND

[0002] With the increasing severity of global climate change, reducing carbon emissions and achieving green and low-carbon development have become a consensus. Thermal power enterprises, as an important source of energy consumption and carbon emissions, have a high coupling relationship with multi-dimensional data such as power consumption, energy consumption, product output, and production process, providing a new way for energy and carbon emission monitoring and accounting.

[0003] However, the existing carbon emission accounting system mainly relies on the IPCC inventory method and the national and provincial greenhouse gas inventory compilation guidelines, and there are problems such as insufficient timely data collection, lagging statistical accounting information, and inconsistent accounting boundaries and methods.

[0004] In addition, the electric-energy-carbon numerical relationship of thermal power enterprises under different operating conditions such as power start-stop, climbing / slide, rated / non-rated operating points is inconsistent, which is prone to overfitting, and the accuracy of predicting carbon emissions is poor. Therefore, it is necessary to design a thermal power enterprise carbon emission simulation model construction method based on multi-source big data. SUMMARY

[0005] In view of the above existing problems, the present application provides a thermal power carbon emission simulation modeling method and system based on multi-source big data, to solve the problems of insufficient timely data collection, lagging statistical accounting information, and inconsistent accounting boundaries and methods in the prior art.

[0006] To solve the above technical problems, a thermal power carbon emission simulation modeling method based on multi-source big data is proposed, which includes,

[0007] Thermal power data is obtained and preprocessed. Based on the preprocessed data, a first model is constructed, and a dynamic coupling relationship between devices is generated through time sequence feature mapping and node information aggregation. Real-time electric-energy-carbon modeling of coal-fired and gas-fired thermal power enterprises under all operating conditions is performed, and the data relationship of different operating conditions is distinguished and a second model is constructed.

[0008] As a preferred scheme of the thermal power carbon emission simulation modeling method based on multi-source big data, the data preprocessing includes screening abnormal points in the data based on multi-dimensional distance calculation, determining abnormal data according to the local density distribution characteristics of the abnormal points, and cleaning and correcting the abnormal data. The corrected data is divided into a training set and a test set.

[0009] As a preferred scheme of the thermal power carbon emission simulation modeling method based on multi-source big data, the first model is constructed by numbering the equipment nodes and constructing a time sequence relationship matrix according to the production process of the thermal power enterprise, generating an initial embedding representation for each equipment node, and introducing a gating mechanism to dynamically control the information transmission ratio between nodes by iteratively aggregating the time sequence feature information of adjacent nodes, and outputting the aggregated embedding representation of the equipment.

[0010] As a preferred scheme of the thermal power carbon emission simulation modeling method based on multi-source big data, the aggregation of the time sequence feature information of adjacent nodes includes defining the connection weight between nodes based on the time sequence relationship matrix, updating the embedding representation of the node by iteration, combining the current input feature and historical state information in each iteration, and using the gating mechanism to filter and fuse the key information at different iteration stages to generate dynamically updated node features.

[0011] As a preferred scheme of the thermal power carbon emission simulation modeling method based on multi-source big data, the real-time electric-energy-carbon modeling includes clustering data by operating condition type based on the dynamic coupling relationship output by the first model, extracting representative data nodes in each operating condition category, and establishing a mapping relationship between operating condition features and carbon emissions, and training and optimizing the mapping relationship by gradient boosting algorithm to generate an electric-energy-carbon correlation model under all operating conditions.

[0012] As a preferred scheme of the thermal power carbon emission simulation modeling method based on multi-source big data, the training and optimization of the mapping relationship by the gradient boosting algorithm includes initializing a base model and calculating the residual between the predicted value and the actual value, constructing a decision tree model based on the residual, minimizing the objective function error by splitting nodes, integrating the decision tree model into the base model by weighting, and optimizing the model precision by multiple iterations.

[0013] The splitting rule of the decision tree model includes traversing all features and splitting points, calculating the change of mean square error before and after splitting, selecting the feature and splitting point that reduces the error the most for node division, and determining the node prediction value according to the mean value of the sample in the leaf node.

[0014] As a preferred scheme of the thermal power carbon emission simulation modeling method based on multi-source big data, the optimization of the model precision includes setting a learning rate to control the weight of the new model in each iteration, terminating the iteration when the model prediction accuracy reaches a preset threshold, outputting a second model, dynamically binding the second model with real-time monitoring data, updating the parameters and mapping relationship of the model, and outputting the real-time electric-energy-carbon dynamic monitoring result of the thermal power enterprise according to the updated model.

[0015] The output dynamic monitoring result comprises collecting real-time data under the current operation condition and inputting the second model, calculating the real-time carbon emission through the model, comparing and analyzing with the historical data, adjusting the operation parameters of the thermal power generation enterprise based on the analysis result, and optimizing the carbon emission efficiency.

[0016] As a preferred scheme of the thermal power carbon emission simulation modeling system based on multi-source big data, characterized in that, comprising a data integration and preprocessing module, an electricity-carbon coupling modeling module, an operating condition analysis and optimization module, and a real-time monitoring and dynamic adjustment module.

[0017] The data integration and preprocessing module is used to collect multi-source data of the thermal power generation enterprise, filter abnormal points through multi-dimensional distance calculation, clean and correct abnormal data based on local density distribution, and time sequence number the equipment nodes according to the production process and construct a relation matrix to divide the training set and the test set.

[0018] The electricity-carbon coupling modeling module is used to generate an initial embedding representation for the equipment nodes based on the preprocessed data, iteratively aggregate the time sequence feature information of adjacent nodes, dynamically control the information transmission ratio between nodes by combining the gating mechanism, generate dynamically updated node features, and construct a refined model of the electricity-carbon coupling relationship to represent the dynamic correlation between energy conversion and carbon emission of the equipment.

[0019] The operating condition analysis and optimization module is used to cluster the data based on the electricity-carbon coupling model, extract representative node data of the operating condition category, and construct an electricity-energy-carbon hierarchical correlation model through the gradient boosting algorithm, and iteratively optimize the model parameters using residual calculation and decision tree splitting rules.

[0020] The real-time monitoring and dynamic adjustment module is used to dynamically bind the optimized model and real-time monitoring data, update the model parameters and mapping relationship, input the current operating condition data in real time, calculate the carbon emission and compare it with the historical data, generate a dynamic monitoring result, adjust the operation parameters of the thermal power generation enterprise according to the analysis feedback, optimize the carbon emission efficiency, and form a closed loop control.

[0021] A computer device comprising a memory and a processor, the memory storing a computer program, and the processor implementing the steps of the method for thermal power carbon emission simulation modeling based on multi-source big data when executing the computer program.

[0022] A computer readable storage medium having a computer program stored thereon, the computer program being executed by a processor to implement the steps of the method for thermal power carbon emission simulation modeling based on multi-source big data.

[0023] The beneficial effects of the present application: the present application screens abnormal points by multi-dimensional distance calculation, cleans and corrects data by using local density distribution characteristics, and divides training set and test set, improves data reliability and enhances model generalization ability; and constructs a time sequence relationship matrix based on the production process of the equipment node number, generates an initial embedding representation for the equipment, iteratively aggregates the time sequence characteristics of adjacent nodes, introduces a gating mechanism to control the information transmission ratio, realizes fine representation ability and optimizes noise resistance and convergence speed; based on the coupling relationship output by the fine-grained model of electric-carbon coupling, the representative node data is extracted according to the working condition, the mapping relationship is trained by gradient boosting algorithm, the electric-energy-carbon correlation model is generated, and the real-time data optimization model is dynamically bound, the monitoring result is output and the operation parameter is adjusted, realizing the adaptability of all working conditions and real-time closed-loop optimization. BRIEF DESCRIPTION OF DRAWINGS

[0024] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0025] Figure 1 The overall flowchart of the method for simulating and modeling carbon emission of thermal power generation based on multi-source big data provided by an embodiment of the present application.

[0026] Figure 2 The system scheme flowchart of the simulation and modeling system for carbon emission of thermal power generation based on multi-source big data provided by an embodiment of the present application. DETAILED DESCRIPTION

[0027] In order to make the above-mentioned purposes, features and advantages of the present application more apparent and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the drawings in the specification. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should be within the scope of protection of the present application.

[0028] In the following description, many specific details are set forth in order to provide a thorough understanding of the present application, but the present application can also be implemented in other ways different from the description, and those skilled in the art can make similar generalizations without departing from the connotation of the present application, therefore the present application is not limited to the specific embodiments disclosed below.

[0029] Secondly, the "one embodiment" or "an embodiment" referred to herein means a specific feature, structure, characteristic, or combination of features and / or characteristics described herein that can be included in at least one implementation of the present application. The various appearances of "in one embodiment" or "an embodiment" in the specification do not all refer to the same embodiment, although they can.

[0030] The application is described in detail in conjunction with the schematic diagram. In the detailed description of the embodiments of the application, the sectional view of the device structure is partially enlarged without the general proportion for the convenience of description, and the schematic diagram is only an example, which should not limit the scope of protection of the application herein. In addition, the three-dimensional spatial dimensions of length, width and depth should be included in actual production.

[0031] Meanwhile, in the description of the present application, it should be noted that the terms "upper, lower, inner and outer" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the indicated device or element must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present application. In addition, the terms "first, second or third" are only for the purpose of description, and cannot be understood as indicating or implying relative importance.

[0032] Unless otherwise expressly specified and limited, the terms "mounting, connecting, connecting" in the present application should be understood broadly, for example: it can be fixed connection, detachable connection or integral connection; it can also be mechanical connection, electrical connection or direct connection, it can also be indirectly connected through intermediate medium, or it can be the communication inside two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0033] Embodiment 1, refer to Figure 1 As the first embodiment of the present application, the embodiment provides a thermal power carbon emission simulation modeling method based on multi-source big data, which comprises:

[0034] S1: Obtain thermal power data and pre-process the data.

[0035] Furthermore, the thermal power generation data obtained include data of power generation enterprises during operation, data of different equipment in thermal power generation enterprises and emission factor calculation data. The data of power generation enterprises during operation include energy usage, power consumption characteristics, loss characteristics, non-power consumption characteristics and proportions, real-time hourly monitoring carbon emission data and real-time power generation data throughout the year; the data of different equipment in thermal power generation enterprises include historical electricity consumption, historical energy consumption, output data and carbon emission data; the emission factor calculation data include as-received basis moisture, air-dry basis moisture, as-received basis elemental carbon content, air-dry basis elemental carbon content, dry basis elemental carbon content, as-received basis lower calorific value, air-dry basis lower calorific value, dry basis lower calorific value, carbon content per unit calorific value, carbon oxidation rate, as-received basis hydrogen content, air-dry basis hydrogen content, dry basis hydrogen content, as-received basis total sulfur, air-dry basis total sulfur, dry basis total sulfur.

[0036] In the embodiment of the present application, the data preprocessing includes processing the data of thermal power generation enterprises and equipment based on the local outlier factor method (LOF), including screening abnormal data, cleaning abnormal data and correcting data, and calculating the distance between sample points. The Euclidean distance formula is used to calculate the distance between sample points. The specific formula is expressed as follows:

[0037]

[0038] Among them, A and B are two n-dimensional vectors, n is the dimension of the vector, that is, the number of elements in the vector, p is a parameter, p of the Euclidean distance is 2, i is the vector dimension index, a i is the value of vector A in the i-th dimension, b i is the value of vector B in the i-th dimension, and dist(A,B) is the distance between sample points.

[0039] In an optional embodiment, the data preprocessing includes randomly extracting subsamples from the thermal power generation data, constructing multiple isolation trees, recursively splitting the data for each isolation tree, randomly selecting features and splitting points until all samples reach the tree depth limit, calculating the anomaly score based on the path length of the sample in the tree (the number of edges from the root node to the isolated node), setting a threshold to filter out abnormal points, and correcting the abnormal values ​​in combination with equipment operating parameters.

[0040] In another optional embodiment, the data preprocessing includes setting a radius and a minimum number of samples, calculating the number of samples in the neighborhood of each data point, and if the number of samples in the neighborhood of the point is greater than or equal to the minimum number of samples, it is marked as a core point, otherwise it is a noise point. Starting from the core point, the points with reachable density are recursively merged to form a cluster, and the unclustered points are regarded as abnormal. In combination with the operating status of the equipment, the noise points are corrected and eliminated.

[0041] The reachable distance of the local sample point is calculated based on the distance between sample points, the reachable distance of the kth cluster sample point of point O is the maximum value of the distance between the sample point and O and the distance between the kth point P and O, and the reachable distance formula of the kth cluster point of point O and O is:

[0042] dist k (O,P)=MAX{dist k (O),dist(O,P)}

[0043] Wherein, dist k (O,P) is the reachable distance of the kth cluster sample point of point O, dist k (O) is the distance between the sample point and O, and dist(O,P) is the distance between the kth point P and O.

[0044] The local reachable density is calculated based on the reachable distance of the local sample point, the local reachable density refers to the average reachable distance of all points P in the neighborhood of point O to point O, the smaller the sum of the reachable distances between point O and the neighborhood points in the same cluster, the greater the local reachable density, and the formula is:

[0045]

[0046] Wherein, lrd k (O) is the local reachable density of point O, N k (O) is the set of all points in the data set whose distance to point P is less than or equal to dist k (O,P).

[0047] The local outlier factor is calculated based on the local reachable density, the abnormal data points are judged based on the local outlier factor, and cleaning or correction is performed, the local outlier factor LOF is the basis for judging whether point O is an abnormal point, the calculation method is to use the average value of the ratio of the local reachable density of the remaining points in the neighborhood of point O to the local reachable density of point O itself, when LOF is less than 2, it means that the local reachable density of point O is similar to the local reachable density of the neighborhood points, and point O and the neighborhood points belong to a cluster, when LOF is less than 1, it means that the reachable density of point O is greater than the density of the neighborhood points, and O is a dense point, when LOF is greater than 1, it means that the reachable density of point O is less than the density of the neighborhood points, and point O is an abnormal point, and the formula for calculating the local outlier factor is:

[0048]

[0049] Wherein, LOF k (O) is the local outlier factor of point O, lrd k (O) is the local reachable density of point O, and lrd k (P) is the local reachable density of point P.

[0050] S2: based on the pre-processed data, a first model is constructed, and a dynamic coupling relationship between devices is generated through time sequence feature mapping and node information aggregation.

[0051] Further, the first model is an electricity-carbon coupling refinement model, which is based on the processed data and takes each device as a node to construct the electricity-carbon coupling refinement model. Specifically, the pre-processed data is divided to obtain a training set and a test set, and each device in the device sequence of the thermal power generation enterprise is initialized according to a 9:1 ratio to obtain an initial item embedding representation.

[0052] The production process of the thermal power generation enterprise is obtained, the sequences of different devices are determined, and the sequences are numbered by natural numbers and assigned to the corresponding edges to represent the current time sequence feature of the session. The relationship matrix about the current graph is obtained. To reduce the calculation complexity of dimension influence, the tanh function can be used to adjust the x-axis to obtain the tanh'(x) function, which reasonably maps the matrix elements between 0 and 1, and is expressed as:

[0053]

[0054] where x is the input data, and tanh'(x) is the hyperbolic tangent function.

[0055] The gated recurrent unit (GRU) is introduced into the graph neural network, and the number of graph neural network learning is set to a fixed number of steps. The information filtering is controlled through the forget gate and the update gate. Based on the training set, the graph neural network is trained to complete the propagation of each item information in the sequence and the aggregation of neighbor information, and the aggregated embedding representation of each item is obtained, i.e. the electricity-carbon coupling refinement model, which is expressed as:

[0056]

[0057] where j represents the jth item node, m represents the number of item nodes in the graph, is the target embedding representation of item j obtained by aggregating the neighbor information of t layers, is the initial d-dimensional embedding representation of item node 1 to node m, A s,j: is the standardized relationship matrix, H is the identity matrix, b is the bias matrix, and T is the transpose operation.

[0058] The calculation formula of the item embedding representation based on the gated recurrent unit is expressed as:

[0059]

[0060] where t represents the time step, represents the current time, and v represents the item number. is the update gate state value of item v at time t, the update gate is used to control the fusion proportion of the previous time information and the current input information, is the reset gate state value of item v at time t, the reset gate is used to control the influence degree of the previous time information on the current calculation, σ is an activation function, W z is the weight matrix multiplied by the current input av(t) in the update gate calculation, used to learn the feature representation of the current input, U z is the weight matrix multiplied by the node feature representation of the previous time in the update gate calculation, used to learn the feature representation of the previous time information, W r is the weight matrix multiplied by the current input in the reset gate calculation, used to learn the feature representation of the current input, U r is the weight matrix multiplied by the node feature representation of the previous time in the reset gate calculation, used to learn the feature representation of the previous time information, is the input feature vector of item v at time t, representing the information input into the model at the current time, is the node feature representation of item v at time t-1, which is the feature vector output by the model at the previous time.

[0061] S3: Real-time electricity-energy-carbon modeling of coal-fired and gas-fired thermal power enterprises under all working conditions, distinguishing the data relationship of different operating conditions and building a second model.

[0062] Further, the second model is a hierarchical correlation model, based on the electricity-carbon coupling refinement model, the real-time electricity-energy-carbon modeling of coal-fired and gas-fired thermal power enterprises under all operating conditions, the clustering processing of the electricity-carbon coupling relationship obtained by the electricity-carbon coupling refinement model of a single thermal power enterprise, the distinction of the electricity-energy-carbon numerical relationship under different operating conditions of power start-stop, climbing / slide, rated / non-rated working points, and the selection of representative node data.

[0063] In the embodiments of the present application, building a thermal power enterprise hierarchical correlation model includes machine learning modeling based on data, building a thermal power enterprise hierarchical correlation model through a CatBoost algorithm, and the formula is represented as:

[0064]

[0065] wherein y g is the actual value of the gth sample, x g is the input feature vector of sample g, F(x g ) is the predicted value of the gth sample by the model, L is a loss function, and G is the number of samples.

[0066] In an alternative embodiment, the construction of the hierarchical correlation model of the thermal power enterprise comprises defining a base model and calculating an initial prediction value, calculating a residual of a current prediction value and an actual carbon emission, constructing a decision tree based on the residual, selecting an optimal feature and a split point through a greedy algorithm, introducing a regularization term to control the complexity of the model and prevent overfitting, adding the newly generated tree to the base model with a weight controlled by a learning rate, stopping iteration when the accuracy of the model on the test set reaches a preset threshold (RMSE < 5%), and outputting the final model.

[0067] In another alternative embodiment, the construction of the hierarchical correlation model of the thermal power enterprise comprises discretizing continuous features into histograms to reduce the amount of calculation, in each iteration, preferentially retaining samples with the largest gradient to improve training efficiency, bundling sparse and mutually exclusive features into a single feature to reduce dimensionality, and using a leaf splitting strategy to directly optimize the loss function and accelerate model convergence.

[0068] The calculated residual (regression task) is used as a new target value, and a new decision tree is constructed using a "symmetric tree structure". The formula for calculating the mean square error is:

[0069]

[0070] Where R is the sample set of the current node, D R is the number of samples in node R, R1 and R2 are the left child node and the right child node formed after splitting, and are the mean values of the sample target values in node R, y u is the sample target value, MSE R is the mean square error of node R before splitting, ΔMSE is the reduction of the mean square error, is the mean square error of the left child node after splitting, is the mean square error of the right child node, is the number of samples in node R1, is the number of samples in node R2, and u is the variable index.

[0071] The new weak learner is weighted into the model to obtain a new model, and the weight is updated by iteration to reduce the residual until a set threshold is reached, i.e., the model accuracy is greater than 95%, and the hierarchical correlation model of the thermal power enterprise is obtained.

[0072] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and are not limiting. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced equivalently without departing from the spirit and scope of the present application, and they should be covered in the scope of the claims of the present application.

[0073] Embodiment 2, reference Figure 2 As a second embodiment of the present application, the embodiment provides a thermal power carbon emission simulation modeling system based on multi-source big data, which comprises a data integration and preprocessing module, an electricity-carbon coupling modeling module, a working condition analysis and optimization module, and a real-time monitoring and dynamic adjustment module.

[0074] The data integration and preprocessing module is used to collect multi-source data of thermal power enterprises, filter abnormal points through multi-dimensional distance calculation, clean and correct abnormal data based on local density distribution, and number equipment nodes in time sequence and construct a relation matrix according to the production process, and divide training set and test set.

[0075] The electricity-carbon coupling modeling module is used to generate an initial embedding representation for the equipment node based on the preprocessed data, iteratively aggregate the time sequence feature information of adjacent nodes, dynamically control the information transmission ratio between nodes by combining the gating mechanism, generate dynamically updated node features, and construct a refined model of electricity-carbon coupling relationship, representing the dynamic correlation between energy conversion and carbon emission among devices.

[0076] The working condition analysis and optimization module is used to cluster data based on the electricity-carbon coupling model, extract representative node data of working condition categories, and construct an electricity-energy-carbon hierarchical correlation model by gradient boosting algorithm, and iteratively optimize model parameters using residual calculation and decision tree splitting rules.

[0077] The real-time monitoring and dynamic adjustment module is used to dynamically bind the optimized model with real-time monitoring data, update model parameters and mapping relationship, input current running condition data in real time, calculate carbon emission and compare with historical data, generate dynamic monitoring results, and adjust the running parameters of thermal power enterprises according to the analysis feedback, optimize carbon emission efficiency, and form a closed loop control.

[0078] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and are not limited. Although the present application has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solutions of the present application can be modified or replaced equivalently without departing from the spirit and scope of the present application, which should be covered in the scope of the claims of the present application.

[0079] Embodiment 3, the third embodiment of the present application, which is different from the first two embodiments:

[0080] If the functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application or the part of the technical solutions that essentially contribute to the prior art or the part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0081] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered a list of executable instructions for implementing logic functions, and can be specifically embodied in any computer-readable medium for use by an instruction execution system, apparatus, or device, such as a computer-based system, a system including a processor, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions, or in conjunction with these instructions execution systems, apparatuses, or devices. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport programs for use by an instruction execution system, apparatus, or device, or in conjunction with these instruction execution systems, apparatuses, or devices.

[0082] More specific examples (a non-exhaustive list) of the computer-readable medium include the following: an electrical connection having one or more wires (electrical devices), a portable computer diskette (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium can even be paper or other suitable medium on which the program can be printed, as the program can be electronically obtained, for example, by optical scanning of the paper or other medium, followed by editing, interpreting, or otherwise processing, if necessary, in other suitable ways to be electronically obtained, and then stored in the computer memory.

[0083] It should be understood that aspects of the application can be implemented in hardware, software, firmware or combinations thereof. In the embodiments described above, various steps or methods can be implemented, for example, by software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, and in another embodiment, any of the following technology, known in the art, or combinations thereof, can be used: discrete logic circuitry having logic gates for implementing logic functions upon an application of data signals, application specific integrated circuits having appropriate combinational logic gates, programmable gate arrays (PGA), field programmable gate arrays (FPGA), and the like.

Claims

1. A simulation modeling method for carbon emissions from thermal power generation based on multi-source big data, characterized by: include, Obtain thermal power generation data and preprocess the data; Based on the preprocessed data, a first model is constructed, and the dynamic coupling relationship between devices is generated through time series feature mapping and node information aggregation; Conduct real-time electricity-energy-carbon modeling under all operating conditions for coal-fired and gas-fired power generation enterprises, distinguish the data relationships under different operating conditions and build a second model.

2. The method for simulating and modeling carbon emissions from thermal power generation based on multi-source big data according to claim 1, characterized in that: The data preprocessing includes screening outliers in the data based on multidimensional distance calculation, determining the outlier data according to the local density distribution characteristics of the outliers, cleaning and correcting the outliers, and dividing the corrected data into a training set and a test set.

3. The method for simulating and modeling carbon emissions from thermal power generation based on multi-source big data according to claim 2, characterized in that: The construction of the first model includes numbering device nodes and constructing a timing relationship matrix based on the production process of the thermal power generation enterprise, generating an initial embedding representation for each device node, iteratively aggregating the timing feature information of adjacent nodes, introducing a gating mechanism to dynamically control the information transmission ratio between nodes, and outputting an aggregated embedding representation of the device.

4. The method for simulating and modeling carbon emissions from thermal power generation based on multi-source big data according to claim 3, characterized in that: The aggregating temporal feature information of adjacent nodes includes defining the connection weights between nodes based on a temporal relationship matrix, iteratively updating the embedded representation of the nodes, combining the current input features with historical state information in each iteration, and using a gating mechanism to filter and fuse key information from different iteration stages to generate dynamically updated node features.

5. The method for simulating and modeling carbon emissions from thermal power generation based on multi-source big data according to claim 4, characterized in that: The real-time electricity-energy-carbon modeling includes clustering data according to operating condition type based on the dynamic coupling relationship output by the first model, extracting representative data nodes in each operating condition category, and establishing a mapping relationship between operating condition characteristics and carbon emissions. The mapping relationship is trained and optimized through a gradient boosting algorithm to generate an electricity-energy-carbon correlation model under all operating conditions.

6. The method for simulating and modeling carbon emissions from thermal power generation based on multi-source big data according to claim 5, characterized in that: The training and optimization of the mapping relationship by the gradient boosting algorithm includes initializing the basic model and calculating the residual between the predicted value and the actual value, constructing a decision tree model based on the residual, minimizing the objective function error by splitting the nodes, integrating the decision tree model into the basic model with weights, and optimizing the model accuracy through multiple rounds of iterations; The splitting rules of the decision tree model include traversing all features and splitting points, calculating the change in mean square error before and after the split, selecting the features and splitting points that maximize the error reduction for node division, and determining the node prediction value based on the mean value of the samples in the leaf node.

7. The method for simulating and modeling carbon emissions from thermal power generation based on multi-source big data according to claim 6, characterized in that: The optimization model accuracy includes setting a learning rate to control the weight of the newly added model in each iteration, terminating the iteration when the model prediction accuracy reaches a preset threshold, outputting a second model, dynamically binding the second model to the real-time monitoring data, updating the model parameters and mapping relationships, and outputting the power-energy-carbon dynamic monitoring results of the thermal power generation enterprise in real time based on the updated model; Outputting dynamic monitoring results includes collecting real-time data under current operating conditions and inputting it into a second model, obtaining real-time carbon emissions through model calculations, and comparing and analyzing them with historical data. Based on the analysis results, the operating parameters of the thermal power generation enterprise are adjusted and carbon emission efficiency is optimized.

8. A system using the thermal power generation carbon emission simulation modeling method based on multi-source big data as described in any one of claims 1 to 7, characterized in that: It includes data integration and preprocessing module, electric-carbon coupling modeling module, operating condition analysis and optimization module, and real-time monitoring and dynamic adjustment module; The data integration and preprocessing module is used to collect multi-source data from thermal power generation enterprises, screen outliers through multi-dimensional distance calculation, clean and correct outlier data based on local density distribution, and sequentially number equipment nodes according to the production process and construct a relationship matrix to divide the data into training and test sets. The electric-carbon coupling modeling module is used to generate an initial embedded representation for the device node based on the preprocessed data. By iteratively aggregating the temporal feature information of adjacent nodes and combining a gating mechanism to dynamically control the information transmission ratio between nodes, it generates dynamically updated node features and constructs a refined model of the electric-carbon coupling relationship to characterize the dynamic relationship between energy conversion and carbon emissions between devices. The operating condition analysis and optimization module is used to cluster data based on the electric-carbon coupling model, extract representative node data of the operating condition category, and construct an electric-energy-carbon hierarchical correlation model through the gradient boosting algorithm, and iteratively optimize the model parameters using residual calculation and decision tree splitting rules; The real-time monitoring and dynamic adjustment module is used to dynamically bind the optimized model with the real-time monitoring data, update the model parameters and mapping relationships, input the current operating condition data in real time, calculate the carbon emissions and compare and analyze them with historical data, generate dynamic monitoring results, and adjust the operating parameters of the thermal power generation enterprise based on the analysis feedback, optimize the carbon emission efficiency, and form a closed-loop control.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the thermal power generation carbon emission simulation modeling method based on multi-source big data described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the thermal power generation carbon emission simulation modeling method based on multi-source big data described in any one of claims 1 to 7 are implemented.

Images