A method, system, medium and device for monitoring carbon emissions in space

Abstract

The present invention relates to a spatial carbon emission monitoring method, system, medium and equipment, which adopts automatic monitoring means such as remote sensing and radar, integrates multiple data sources such as spectral information, radar detectors, and process parameters, and can fully perceive the carbon emission status inside industrial buildings, and can realize real-time collection and dynamic monitoring of spatial carbon emissions of industrial buildings, instead of relying on manual records and data reporting, reducing the problems of time lag and missing edge data. Through deep learning technologies such as kernel PCA, ICA, neural networks, etc., more effective data features can be extracted from multidimensional features, making the predicted carbon emissions more accurate, thereby improving the accuracy of monitoring. By integrating monitoring data into the BIM model, the visual presentation of carbon emission information is realized, which is convenient for managers to grasp and analyze the carbon emission dynamics of industrial buildings in a timely manner.

Description

Technical Field

[0001] The present invention relates to the field of carbon emission monitoring, and in particular to a method, system, medium and equipment for monitoring spatial carbon emission. Background Art

[0002] Carbon emissions, also known as carbon footprint or greenhouse gas emissions, refer to the total amount of carbon dioxide or other greenhouse gas emissions produced by an entity over a certain period of time, usually expressed in carbon dioxide equivalents. Carbon emissions are an important indicator to measure the impact of human activities on climate change.

[0003] At present, factory carbon emissions monitoring mainly relies on statistics and analysis of production processes, energy usage, etc., which makes it difficult to fully and accurately reflect the factory's carbon emissions. In addition, the recording and reporting of carbon emissions data in many factories often have time lags, and edge data collection mostly relies on manual recording or updating, resulting in the existing carbon emissions accounting methods being unable to achieve accurate monitoring. Summary of the invention

[0004] In view of the above problems, the present application provides a spatial carbon emission monitoring method, system, medium and equipment.

[0005] To achieve the above object, in a first aspect, the present invention provides a spatial carbon emission monitoring method, which is applicable to industrial areas, where the industrial areas include industrial buildings, and the method comprises:

[0006] Acquire spectral information within a first preset time period, where the spectral information is configured to be acquired through a remote sensing sensor, and obtain first gas concentration distribution information in the industrial area through an inversion algorithm using the spectral information, where the first gas concentration distribution information includes a first greenhouse gas category, a first greenhouse gas concentration, and a first greenhouse gas distribution feature;

[0007] and obtaining gas reflection information in the industrial building within a first preset time period, the gas reflection information being configured to be obtained by measuring with a radar detector, and obtaining second gas concentration distribution information in the industrial building by using an echo time difference algorithm for the gas reflection information, the second gas concentration distribution information including a second greenhouse gas category, a second greenhouse gas concentration, and a second greenhouse gas distribution characteristic;

[0008] and, obtaining process information in the industrial building within a first preset time period, the process information including process equipment information, process raw material information and personnel configuration information, and generating third gas concentration distribution information in the industrial building according to the process information, the third gas concentration distribution information including third greenhouse gas category, third greenhouse gas concentration and third greenhouse gas distribution characteristics;

[0009] Performing feature fusion on the first gas concentration distribution information, the second gas concentration distribution information, and the third gas concentration distribution information to obtain a first fusion feature, where the first fusion feature is a high-dimensional fusion feature;

[0010] Perform kernel PCA dimensionality reduction on the first fusion feature to obtain a first dimensionality reduction feature, the first dimensionality reduction feature includes a principal component subspace, perform ICA dimensionality reduction on the principal component subspace to obtain a second dimensionality reduction feature, the second dimensionality reduction feature includes multiple independent component features, and the second dimensionality reduction feature is recorded as a second fusion feature;

[0011] Inputting the second fusion feature into a carbon emission prediction model to obtain first carbon emission change information within a second preset time period, wherein the carbon emission prediction model is configured to be constructed based on a neural network model;

[0012] Structural parameter information of the industrial building is obtained, and a BIM model is constructed according to the structural parameter information and the first carbon emission change information, so as to obtain and display carbon emission evaluation information of the industrial building, wherein the carbon emission evaluation information includes a plurality of carbon emission thermal distribution diagrams in the industrial building.

[0013] In some embodiments, the inversion algorithm includes a differential absorption method inversion operation and a radiation transmission optimization operation, and obtaining the first gas concentration distribution information in the industrial area through the inversion algorithm using the spectral information includes:

[0014] The spectral information is preprocessed, and the spectral information preprocessing includes radiation correction and Gaussian filtering to obtain first processed spectral information, which is expressed by formula (1). Formula (1) is as follows:

[0015]

[0016] In formula (1), L′(λ) is the spectral information after radiation correction, L 1 (λ) is the radiation information of the atmospheric scattering path, τ is the atmospheric transmittance, L 2 (λ) is the surface reflected radiation information, L 3 (λ) is the atmospheric downlink radiation information, L″(λ) is the first processed spectrum information, λ 0 is the central wavelength of the spectral information, λ is the wavelength value to be processed in the spectral information, σ is the standard deviation of the Gaussian kernel, and N is the total wavelength value in the spectral information;

[0017] Spectral features are extracted from the first processed spectral information to obtain a plurality of first spectral features, and the plurality of first spectral features are matched according to greenhouse gas categories to obtain a plurality of first greenhouse gas categories, which are expressed by formula (2). Formula (2) is as follows:

[0018] Q i =f(F)=UΣ′V T ;

[0019] In formula (2), Q i is the i-th greenhouse gas category, f(F) is the classification function, F is the spectral feature matrix formed by multiple first spectral features, U is the left singular vector matrix, the column vector of U is the principal component vector of F, Σ′ is the diagonal matrix containing the singular values ​​of the principal component vector of F, V T is the transpose of the right singular vector matrix, V T The row vector of is the projection of the spectral information in the principal component space of F;

[0020] The first processed spectral information is subjected to a differential absorption inversion operation according to the first greenhouse gas category to obtain the first initial gas concentration information, which is expressed by formula (3). Formula (3) is as follows:

[0021]

[0022] In formula (3), is the first initial gas concentration information corresponding to the i-th first greenhouse gas category, dL″(λ) is the differential change of the spectral brightness of the λ wavelength value, α i is the gas absorption coefficient of the ith greenhouse gas category, l is the optical path length;

[0023] The first initial gas concentration information is subjected to a radiation transmission optimization operation to obtain the first final gas concentration information, which is expressed by formula (4). Formula (4) is as follows:

[0024]

[0025] In formula (4), is the first final gas concentration information corresponding to the i-th first greenhouse gas category, I 0 is the initial radiation intensity, I(s,k,j) is the radiation intensity after passing through the atmosphere, k is the extinction coefficient, j is the emissivity, s is the optical path length from the observation point where the spectral information is collected to the upper boundary of the atmosphere, s′ is the micro optical path length from the atmospheric incident point to the integral position, and s′′ is the micro optical path length from the integral position to the observation point where the spectral information is collected;

[0026] The first final gas concentration information is mapped and stored with the first greenhouse gas category to obtain the first gas concentration distribution information.

[0027] In some embodiments, the industrial building is divided into a plurality of preset areas, a radar detector is provided in each preset area, and each preset area corresponds to a gas reflection information;

[0028] The gas reflection information is obtained by using the echo time difference algorithm to obtain the second gas concentration distribution information in the industrial building, including:

[0029] The following steps are performed for the gas reflection information collected in each preset area:

[0030] Acquire multiple radar measurement information collected by the radar detector in the current preset area within the first preset time period, and organize the multiple radar measurement information into gas reflection information corresponding to the current preset area in chronological order, which is expressed by formula (5). Formula (5) is as follows:

[0031] R mp ={(r m1 ,t m1 ),(r m2 ,t m2 ),…,(r mp ,t mp )};

[0032] In formula (5), R mp is the gas reflection information corresponding to m preset areas, (r m1 ,t m1 ) is the radar measurement information of the first time node in the mth preset area, (r m2 ,t m2 ) is the radar measurement information of the second time node in the mth preset area, (r mp ,t mp ) is the radar measurement information of the p-th time node in the m-th preset area;

[0033] The echo time difference algorithm is used to calculate the current gas reflection information to obtain the gas layer distance between two adjacent time nodes, which is expressed by formula (6). Formula (6) is as follows:

[0034]

[0035] In formula (6), Δr mp is the distance between the gas layers at two adjacent time nodes, v 0 is the speed of light constant;

[0036] Multiple gas layer distances are input into the support vector machine model to obtain the regional gas concentration distribution information corresponding to the current preset area, which is expressed by formula (7). Formula (7) is as follows:

[0037]

[0038] In formula (7), ρ m (x, y, z) is the regional gas concentration distribution information of the prediction point with coordinates (x, y, z) in the current preset area, β a is the first parameter of the support vector machine model, b is the second parameter of the support vector machine model, is a polynomial kernel function, γ is the mapping dimension of the polynomial kernel function, r represents the predicted radar measurement information of the prediction point with coordinates (x, y, z), and t represents the prediction of the prediction point with coordinates (x, y, z) at time node t;

[0039] Repeat the above steps until each preset area corresponds to a regional gas concentration distribution information, and organize the regional gas concentration distribution information into second gas concentration distribution information, which is expressed by formula (8). Formula (8) is as follows:

[0040]

[0041] In formula (8), ρ′(x, y, z) is the second gas concentration distribution information, ρ 1 (x, y, z) is the regional gas concentration distribution information of the first preset area, ρ m (x, y, z) is the regional gas concentration distribution information of the mth preset area.

[0042] In some embodiments, generating third gas concentration distribution information in the industrial building according to the process information includes:

[0043] According to the process flow, the process equipment information, process raw material information and personnel configuration information are associated to obtain a process processing data set, which includes a plurality of process unit information arranged in the order of the process flow, and each process unit information corresponds to the process equipment, process raw material and processing personnel required for a process step;

[0044] Construct a linear regression model between process unit information and carbon emissions, and obtain room information of process unit information in industrial buildings;

[0045] A mapping relationship between each process unit information and the industrial building is established based on the linear regression model and the room information, and the diffusion information of the greenhouse gas of the process unit information in the industrial building is calculated based on the concentration field-flow field coupling equation;

[0046] The mapping relationship and the diffusion information are mapped and stored according to the categories of greenhouse gases to obtain the third gas concentration distribution information.

[0047] In some embodiments, the first fusion feature is subjected to kernel PCA dimensionality reduction to obtain a first dimensionality reduction feature, where the first dimensionality reduction feature includes a principal component subspace including:

[0048] Calculate the kernel function matrix of the first fusion feature, and perform eigenvalue decomposition on the kernel function matrix to obtain a plurality of principal component features sorted in a decomposition order, each principal component feature having an eigenvalue and an eigenvector;

[0049] Select several principal component features that are ranked first to form a principal component matrix, and project the first fused feature onto the principal component matrix to generate a principal component subspace;

[0050] Arrange the principal component subspace to obtain the first dimension reduction feature;

[0051] Perform ICA dimensionality reduction on the principal component subspace to obtain the second dimensionality reduction feature, which includes multiple independent component features. The second dimensionality reduction feature is recorded as the second fusion feature, including:

[0052] Perform dimensionality reduction preprocessing on the principal component subspace, which includes centering and whitening, to obtain the first subspace feature;

[0053] Applying the FastICA algorithm to the first subspace feature to extract independent components, and obtaining at least one independent component feature corresponding to the current first subspace feature;

[0054] The first subspace feature is concatenated with the independent component feature to obtain the second subspace feature, which is the second dimensionality reduction feature.

[0055] In some embodiments, the carbon emission prediction model is trained by the following steps:

[0056] Constructing a neural network model, and inputting the first sample features of the first sample preset time period into the input layer of the neural network model;

[0057] A GRU time series modeling layer is added to the neural network model, and the first sample feature is input into the GRU time series modeling layer to extract the feature of the time series, so as to obtain the first sample time series feature;

[0058] Input the sample time series features into the fully connected layer to obtain the second sample time series features;

[0059] Input the second sample time series feature into the output layer to obtain the second sample feature;

[0060] The aforementioned steps are repeated until the prediction accuracy of the second sample feature reaches a preset accuracy threshold, thereby obtaining a carbon emission prediction model.

[0061] In some embodiments, a BIM model is constructed according to the structural parameter information and the first carbon emission change information to obtain carbon emission assessment information of industrial buildings, including:

[0062] Input the structural parameter information into the BIM model to obtain a three-dimensional model of the industrial building, wherein the industrial building has a plurality of preset areas, and each preset area has a regional three-dimensional model;

[0063] Dividing the first carbon emission change information into a plurality of second carbon emission change information according to a preset area;

[0064] Associating the second carbon emission change information with the regional three-dimensional model to obtain regional carbon emission flow data;

[0065] A carbon emission heat map of a preset area is constructed based on the regional carbon emission flow data, and multiple carbon emission heat maps are correlated and sorted according to the process flow to obtain carbon emission assessment information.

[0066] In a second aspect, the present invention provides a spatial carbon emission monitoring system, which is applicable to the spatial carbon emission monitoring method described in the first aspect. The system includes an information processing module, an information fusion module, a carbon emission processing module and a building carbon emission analysis module. The information processing module includes a spectral information processing unit, a radar information processing unit and a process information processing unit. The spectral information processing unit is used to obtain spectral information within a first preset time period. The spectral information is configured to be acquired through a remote sensing sensor. The spectral information is obtained through an inversion algorithm to obtain first gas concentration distribution information in an industrial area. The first gas concentration distribution information includes a first greenhouse gas category, a first greenhouse gas concentration and a first greenhouse gas distribution feature; the radar information processing unit is used to obtain a first Gas reflection information in an industrial building within a preset time period, the gas reflection information is configured to be obtained by measuring with a radar detector, and the gas reflection information is used to obtain second gas concentration distribution information in the industrial building through an echo time difference algorithm, and the second gas concentration distribution information includes a second greenhouse gas category, a second greenhouse gas concentration, and a second greenhouse gas distribution characteristic; a process information processing unit is used to obtain process information in the industrial building within a first preset time period, the process information includes process equipment information, process raw material information, and personnel configuration information, and generate third gas concentration distribution information in the industrial building according to the process information, and the third gas concentration distribution information includes a third greenhouse gas category, a third greenhouse gas concentration, and a third greenhouse gas distribution characteristic;

[0067] The information fusion module is used to fuse the first gas concentration distribution information, the second gas concentration distribution information and the third gas concentration distribution information to obtain a first fusion feature, which is a high-dimensional fusion feature; the first fusion feature is subjected to kernel PCA dimensionality reduction to obtain a first dimensionality reduction feature, which includes a principal component subspace, and the principal component subspace is subjected to ICA dimensionality reduction to obtain a second dimensionality reduction feature, which includes multiple independent component features, and the second dimensionality reduction feature is recorded as a second fusion feature; the carbon emission processing module is used to input the second fusion feature into the carbon emission prediction model to obtain the first carbon emission change information within a second preset time period, and the carbon emission prediction model is configured to be constructed based on a neural network model; the building carbon emission analysis module is used to obtain the structural parameter information of the industrial building, and construct a BIM model based on the structural parameter information and the first carbon emission change information, to obtain and display the carbon emission evaluation information of the industrial building, and the carbon emission evaluation information includes multiple carbon emission thermal distribution diagrams in the industrial building.

[0068] In a third aspect, the present invention further provides a computer-readable storage medium storing computer program instructions, which implement the method described in the first aspect when executed by a processor.

[0069] In a fourth aspect, the present invention further provides an electronic device, comprising a memory and a processor, wherein the memory is used to store one or more computer program instructions, wherein the one or more computer program instructions are executed by the processor to implement the method described in the first aspect.

[0070] Different from the existing technology, the above technical solution adopts automated monitoring methods such as remote sensing and radar, and integrates multiple data sources such as spectral information, radar detectors, and process parameters. It can fully perceive the carbon emission status inside industrial buildings, and can realize real-time collection and dynamic monitoring of spatial carbon emissions of industrial buildings, instead of relying on manual records and data reporting, reducing the problems of time lag and missing edge data. Through deep learning technologies such as kernel PCA, ICA, and neural networks, more effective data features can be extracted from multidimensional features, making the predicted carbon emissions more accurate, thereby improving the accuracy of monitoring. By integrating monitoring data into the BIM model, the visualization of carbon emission information is realized, which is convenient for managers to grasp and analyze the carbon emission dynamics of industrial buildings in a timely manner.

[0071] The above-mentioned records related to the invention content are only an overview of the technical solution of the present application. In order to enable ordinary technicians in the field to more clearly understand the technical solution of the present application, and then implement it according to the text of the specification and the contents recorded in the drawings, and to make the above-mentioned purpose and other purposes, features and advantages of the present application easier to understand, the following is an explanation in combination with the specific implementation mode and drawings of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] The drawings are only used to illustrate the principles, implementation methods, applications, characteristics and effects of the specific embodiments of the present invention and other related contents, and shall not be considered as limiting the present application.

[0073] In the drawings of the specification:

[0074] Figure 1 This is a schematic diagram of steps S101 to S105 of the spatial carbon emission monitoring method described in the specific implementation method;

[0075] Figure 2 This is a schematic diagram of steps S201 to S204 of the spatial carbon emission monitoring method described in the specific implementation method;

[0076] Figure 3 This is a schematic diagram of steps S301 to S306 of the spatial carbon emission monitoring method described in the specific implementation method;

[0077] Figure 4 It is a schematic diagram of the spatial carbon emission monitoring system described in a specific implementation method.

[0078] The reference numerals in the above drawings are described as follows:

[0079] 1. Spatial carbon emission monitoring system;

[0080] 11. Information processing module;

[0081] 12. Information fusion module;

[0082] 13. Carbon emission processing module;

[0083] 14. Building carbon emission analysis module. DETAILED DESCRIPTION

[0084] In order to explain in detail the possible application scenarios, technical principles, specific schemes that can be implemented, and the purposes and effects that can be achieved, the following is a detailed description of the specific embodiments listed in conjunction with the accompanying drawings. The embodiments described herein are only used to more clearly illustrate the technical solutions of the present application, and are therefore only used as examples, and cannot be used to limit the scope of protection of the present application.

[0085] Reference to "embodiment" herein means that the specific features, structures or characteristics described in conjunction with the embodiment may be included in at least one embodiment of the present application. The term "embodiment" appearing in various places in the specification does not necessarily refer to the same embodiment, nor does it particularly limit its independence or association with other embodiments. In principle, in the present application, as long as there is no technical contradiction or conflict, the various technical features mentioned in the embodiments can be combined in any way to form a corresponding implementable technical solution.

[0086] Unless otherwise defined, the technical terms used in this document have the same meanings as those generally understood by those skilled in the art to which this application belongs; the use of relevant terms in this document is only for describing specific embodiments and is not intended to limit this application.

[0087] In the description of this application, the term "and / or" is an expression used to describe the logical relationship between objects, indicating that three relationships may exist, for example, A and / or B, which means: A exists, B exists, and A and B exist at the same time. In addition, the character " / " in this article generally indicates that the objects before and after are in an "or" logical relationship.

[0088] In the present application, terms such as “first” and “second” are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship of quantity, priority or sequence between these entities or operations.

[0089] Without further limitations, in this application, the words "include", "comprises", "has" or other similar open-ended expressions used in sentences are intended to cover non-exclusive inclusion. These expressions do not exclude the presence of additional elements in the process, method or product including the elements, so that the process, method or product including a series of elements may include not only those limited elements, but also other elements not explicitly listed, or also include elements inherent to such process, method or product.

[0090] Similar to the understanding in the Examination Guidelines, in this application, expressions such as "greater than", "less than", "exceed" and the like are understood to exclude the number itself; expressions such as "above", "below", "within" and the like are understood to include the number itself. In addition, in the description of the embodiments of this application, "multiple" means more than two (including two), and similar expressions related to "multiple" are also understood in this way, such as "multiple groups", "multiple times", etc., unless otherwise clearly and specifically limited.

[0091] See also Figure 1In a first aspect, this embodiment provides a spatial carbon emission monitoring method, which is applicable to industrial areas, where the industrial areas include industrial buildings. The method includes:

[0092] S101, obtaining spectral information within a first preset time period, the spectral information is configured to be obtained through remote sensing sensor collection, and the spectral information is used to obtain first gas concentration distribution information in the industrial area through an inversion algorithm, and the first gas concentration distribution information includes a first greenhouse gas category, a first greenhouse gas concentration, and a first greenhouse gas distribution characteristic; and, obtaining gas reflection information in an industrial building within the first preset time period, the gas reflection information is configured to be obtained through radar detector measurement, and the gas reflection information is used to obtain second gas concentration distribution information in the industrial building through an echo time difference algorithm, and the second gas concentration distribution information includes a second greenhouse gas category, a second greenhouse gas concentration, and a second greenhouse gas distribution characteristic; and, obtaining process information in the industrial building within the first preset time period, the process information includes process equipment information, process raw material information, and personnel configuration information, and generating third gas concentration distribution information in the industrial building according to the process information, and the third gas concentration distribution information includes a third greenhouse gas category, a third greenhouse gas concentration, and a third greenhouse gas distribution characteristic;

[0093] S102, performing feature fusion on the first gas concentration distribution information, the second gas concentration distribution information, and the third gas concentration distribution information to obtain a first fusion feature, where the first fusion feature is a high-dimensional fusion feature;

[0094] S103, performing kernel PCA dimensionality reduction on the first fusion feature to obtain a first dimensionality reduction feature, the first dimensionality reduction feature includes a principal component subspace, performing ICA dimensionality reduction on the principal component subspace to obtain a second dimensionality reduction feature, the second dimensionality reduction feature includes a plurality of independent component features, and recording the second dimensionality reduction feature as a second fusion feature;

[0095] S104, inputting the second fusion feature into a carbon emission prediction model to obtain first carbon emission change information within a second preset time period, wherein the carbon emission prediction model is configured to be constructed based on a neural network model;

[0096] S105. Obtain structural parameter information of the industrial building, and build a BIM model according to the structural parameter information and the first carbon emission change information, obtain and display carbon emission evaluation information of the industrial building, and the carbon emission evaluation information includes multiple carbon emission thermal distribution diagrams in the industrial building.

[0097] In this embodiment, the industrial area can be understood as an area containing multiple industrial buildings, which includes landscapes, roads, street lights, transportation tools, etc. Industrial buildings can be understood as factories, plants or buildings used for a certain process. The industrial area can include multiple industrial buildings. In this embodiment, there is no limit on the number of industrial buildings.

[0098] In step S101, the first preset time period can be set according to actual needs, such as the first preset time period can be set to year, month, quarter, day, week, etc. It should be noted that the spectral information in the first preset time period is acquired regularly, which is convenient for the collection of the time series characteristics of the subsequent carbon emission prediction model. It should be noted that in this embodiment, the spectral information is acquired by remote sensing sensors, and the specific remote sensing sensors may include remote sensing satellites with spectral acquisition functions. The spectral image information in the current industrial area can be obtained by using remote sensing sensors. Further, the spectral information is reversely deduced from the first gas concentration distribution information using an inversion algorithm. It can be understood that the current first gas concentration distribution information includes the first greenhouse gas category, the first greenhouse gas concentration and the first greenhouse gas distribution characteristics. The first greenhouse gas category is the greenhouse gas category extracted from the spectral information. The greenhouse gas category is the gas category that causes the greenhouse effect, specifically including fluorine-containing compounds, nitrogen-containing compounds, carbon-containing compounds, etc., such as perfluorocarbons, nitrous oxide, methane, carbon dioxide, sulfur hexafluoride, etc. The first greenhouse gas concentration is the greenhouse gas concentration in the industrial area corresponding to the first greenhouse gas category, and the first greenhouse gas distribution characteristics can be understood as the distribution characteristics of each category of greenhouse gas in the industrial area.

[0099] In step S101, it is necessary to collect gas reflection information in industrial buildings in industrial areas. Similar to spectral information, gas reflection information is also collected regularly within the first preset time period, which facilitates the subsequent extraction of time series features. Specifically, the gas reflection information is obtained by measuring with a radar detector, and the acquisition method is as follows: pre-install the radar detector at multiple locations in the industrial building, control the radar detector to regularly emit electromagnetic waves, and correspondingly collect the reflected waves of the emitted electromagnetic waves, so as to confirm the current content of greenhouse gases in the air. For ease of expression, the timestamp and wavelength content data obtained in this step of the emitted electromagnetic waves and reflected waves are recorded as gas reflection information. After obtaining the gas reflection information, it is subjected to an echo time difference algorithm to obtain the content and distribution characteristics of greenhouse gases in the space detected by each radar detector, and recorded as the second gas concentration distribution information.

[0100] In step S101, process information within a first preset time period is obtained, where the process information is information about the manufacturing process in the industrial building, specifically including process equipment information, process raw material information and personnel configuration information. The process equipment information includes the equipment model required for use in the process, the energy consumption of the equipment and the working time of the equipment. The carbon emissions generated by the process equipment in a complete product production process can be learned from the process equipment information; the process raw material information includes the category of process raw materials required for use in the process, the carbon emissions information released by the process raw materials during the production process and the additional carbon emissions information required for recycling of waste generated by the process raw materials after production. The personnel configuration information includes the personnel invested in each process flow, the personnel working hours, the work content and the carbon emissions information generated by the personnel during the mobilization process. The carbon emission information in the process information is linearly fitted to obtain the third gas concentration distribution information, where the third gas concentration distribution information includes the third greenhouse gas category, the third greenhouse gas concentration and the third greenhouse gas distribution characteristics.

[0101] In step S102, the first gas concentration distribution information, the second gas concentration distribution information and the third gas concentration distribution information are feature fused. It can be understood that the feature fusion can use Gaussian fusion to adapt the association between the first gas concentration distribution information, the second gas concentration distribution information and the third gas concentration distribution information. The feature fusion process can also be implemented in the form of direct splicing to obtain a first fusion feature. The current first fusion feature is a high-dimensional fusion feature.

[0102] In step S103, for the convenience of calculation, the first fused feature is reduced in dimension to further purify the feature and improve the accuracy of the subsequent carbon emission prediction model. This embodiment uses global PCA dimensionality reduction processing and local ICA dimensionality reduction processing to reduce the dimension of the first fused feature. Specifically, the first fused feature is first subjected to kernel PCA dimensionality reduction, and the feature after dimensionality reduction is recorded as the first dimensionality reduction feature. The current first dimensionality reduction feature includes the principal component subspace. On this basis, the principal component subspace is subjected to ICA dimensionality reduction processing. After the ICA dimensionality reduction processing, the principal component subspace will obtain at least one independent component feature, but it is not unique. When the principal component subspace is subjected to the ICA dimensionality reduction processing, multiple independent component features can be obtained. After sorting out the multiple independent component features, the second dimensionality reduction feature is obtained. For the convenience of expression, the second dimensionality reduction feature is recorded as the second fused feature. That is, the second fused feature is the first fused feature after dimensionality reduction. On this basis, the second fused feature is input into the carbon emission prediction model.

[0103] In step S104, the second fusion feature is input into the carbon emission prediction model to obtain the first carbon emission change information within the second preset time period. Preferably, the second preset time period corresponds to the time length of the first preset time period. For example, when the first preset time period is the time length of one month in the past, the second preset time period is the time length of one month in the future. For example, when the first preset time period is the time length of one quarter in the past, the second preset time period is the time length of one quarter in the future. In this embodiment, the first carbon emission change information is the carbon emission prediction information in this industrial building within the second preset time period. The first carbon emission change information can provide managers with a multi-dimensional adjustment reference for personnel management, process management, and equipment management of industrial buildings within the second preset time period. In this embodiment, the carbon emission prediction model is constructed using a neural network model, and the second fusion feature is extracted and predicted in a rich manner using deep learning, so as to obtain more accurate first carbon emission change information.

[0104] In step S105, the structural parameter information of the industrial building is obtained. It should be noted that the structural parameter information includes the floor, height and size of each room in the current industrial building, as well as the building materials, building layout and other information of the current industrial building. The structural parameters are input into the BIM model to construct the current industrial building, and the digital modeling process of the industrial building is completed. On this basis, the first carbon emission change information is used as the carbon emission data in the current industrial building, integrated into the BIM model, and the carbon emission evaluation information of the industrial building is obtained. The carbon emission evaluation information includes the overall carbon emission thermal distribution map, the local carbon emission thermal distribution map and the carbon emission thermal distribution map that needs to be focused on in the industrial building, etc. The carbon emission evaluation information is displayed to realize the visual display of the industrial building. The digital modeling of the industrial building is more helpful for managers to replace process equipment, transfer personnel, adjust process flow and other operations according to actual needs, so as to reduce the carbon emission of the entire industrial building and optimize the carbon emission management of the industrial building.

[0105] This embodiment uses a variety of means such as remote sensing sensors, radar detectors, and process information to comprehensively obtain greenhouse gas concentration distribution information in industrial areas and buildings, and fully grasp the carbon emission status. Through feature fusion and dimensionality reduction technology, the features of multi-source data are effectively integrated, and key features richer in time series information are extracted, laying the foundation for subsequent carbon emission predictions. Further use of a carbon emission prediction model based on a neural network can accurately predict the changes in carbon emissions of industrial buildings within the second preset time period, integrate carbon emission assessment information into the BIM model, and realize the visualization of industrial buildings, which helps managers to more intuitively understand carbon emission hotspots, and can help managers to adjust personnel, processes, equipment, etc. in a targeted manner, thereby reducing the carbon emissions of the entire industrial building and optimizing carbon emission management.

[0106] In some embodiments, the inversion algorithm includes a differential absorption method inversion operation and a radiation transmission optimization operation, and obtaining the first gas concentration distribution information in the industrial area through the inversion algorithm using the spectral information includes:

[0107] The spectral information is preprocessed, and the spectral information preprocessing includes radiation correction and Gaussian filtering to obtain first processed spectral information, which is expressed by formula (1). Formula (1) is as follows:

[0108]

[0109] In formula (1), L′(λ) is the spectral information after radiation correction, L 1 (λ) is the radiation information of the atmospheric scattering path, τ is the atmospheric transmittance, L 2 (λ) is the surface reflected radiation information, L 3 (λ) is the atmospheric downlink radiation information, L″(λ) is the first processed spectrum information, λ 0 is the central wavelength of the spectral information, λ is the wavelength value to be processed in the spectral information, σ is the standard deviation of the Gaussian kernel, and N is the total wavelength value in the spectral information;

[0110] Spectral features are extracted from the first processed spectral information to obtain a plurality of first spectral features, and the plurality of first spectral features are matched according to greenhouse gas categories to obtain a plurality of first greenhouse gas categories, which are expressed by formula (2). Formula (2) is as follows:

[0111] Q i =f(F)=UΣ′V T ;

[0112] In formula (2), Q iis the i-th greenhouse gas category, f(F) is the classification function, F is the spectral feature matrix formed by multiple first spectral features, U is the left singular vector matrix, the column vector of U is the principal component vector of F, Σ′ is the diagonal matrix containing the singular values ​​of the principal component vector of F, V T is the transpose of the right singular vector matrix, V T The row vector of is the projection of the spectral information in the principal component space of F;

[0113] The first processed spectral information is subjected to a differential absorption inversion operation according to the first greenhouse gas category to obtain the first initial gas concentration information, which is expressed by formula (3). Formula (3) is as follows:

[0114]

[0115] In formula (3), is the first initial gas concentration information corresponding to the first greenhouse gas category i, dL"(λ) is the difference change of the spectral brightness of the wavelength value λ, α i is the gas absorption coefficient of the ith greenhouse gas category, l is the optical path length;

[0116] The first initial gas concentration information is subjected to a radiation transmission optimization operation to obtain the first final gas concentration information, which is expressed by formula (4). Formula (4) is as follows:

[0117]

[0118] In formula (4), is the first final gas concentration information corresponding to the i-th first greenhouse gas category, I 0 is the initial radiation intensity, I(s,k,j) is the radiation intensity after passing through the atmosphere, k is the extinction coefficient, j is the emissivity, s is the optical path length from the observation point where the spectral information is collected to the upper boundary of the atmosphere, s′ is the micro optical path length from the atmospheric incident point to the integral position, and s′′ is the micro optical path length from the integral position to the observation point where the spectral information is collected;

[0119] The first final gas concentration information is mapped and stored with the first greenhouse gas category to obtain the first gas concentration distribution information.

[0120] In this embodiment, the preprocessing of spectral information includes radiation correction and Gaussian filtering. The radiation correction can correct the spectral wavelength change caused by atmospheric radiation in the spectral information, and the Gaussian filtering can filter out the noise in the spectral information to make the spectral features more obvious. Specifically, as shown in formula (1), the spectral information after the spectral information preprocessing is recorded as the first processed spectral information, and the spectral features of the first processed spectral information are extracted to obtain multiple first spectral features.

[0121] f(F) used in formula (2) is a classification function based on singular value decomposition, and different greenhouse gas categories are identified by classifying the singular value information of the first processed spectral information. That is, by utilizing the principle that greenhouse gases have different wavelength absorption bands for visible light, singular value decomposition is performed in the first processed spectral information to obtain a plurality of first spectral features, and then the category of greenhouse gases is identified based on the first spectral features, and finally the first greenhouse gas category is obtained.

[0122] In this embodiment, the first processed spectral information also needs to be subjected to an inversion algorithm, and the inversion algorithm includes two steps: differential absorption inversion operation and radiation transfer optimization operation. The first processed spectral information is first subjected to a differential absorption inversion operation to obtain the first initial gas concentration information. It should be noted that in this process, the first initial gas concentration information is synchronously matched with the first greenhouse gas category, so as to obtain the first initial gas concentration information corresponding to the i-th first greenhouse gas category. Further, the first initial gas concentration information is subjected to a radiation transfer optimization operation to optimize and correct the first initial gas concentration information, filter out the influence of atmospheric and surface factors on the first initial gas concentration information, and finally obtain the first final gas concentration information.

[0123] Finally, the plurality of first final gas concentration information are mapped and stored with the first greenhouse gas category, so as to obtain the first gas concentration distribution information.

[0124] This embodiment improves the quality and availability of spectral information through preprocessing methods such as radiation correction and Gaussian filtering. It adopts a classification function based on singular value decomposition, which can effectively utilize spectral feature information and accurately identify different greenhouse gas categories. Through the two steps of differential absorption inversion and radiation transfer optimization, the preprocessed spectral information is converted into initial and final concentration information of greenhouse gases, thereby improving the accuracy of inversion. Finally, the greenhouse gas categories and corresponding concentration information are comprehensively mapped and stored to form comprehensive gas concentration distribution information in the industrial area.

[0125] In some embodiments, the industrial building is divided into a plurality of preset areas, a radar detector is provided in each preset area, and each preset area corresponds to a gas reflection information;

[0126] The gas reflection information is obtained by using the echo time difference algorithm to obtain the second gas concentration distribution information in the industrial building, including:

[0127] The following steps are performed for the gas reflection information collected in each preset area:

[0128] Acquire multiple radar measurement information collected by the radar detector in the current preset area within the first preset time period, and organize the multiple radar measurement information into gas reflection information corresponding to the current preset area in chronological order, which is expressed by formula (5). Formula (5) is as follows:

[0129] R mp ={(r m1 ,t m1 ),(r m2 ,t m2 ),…,(r mp ,t mp )};

[0130] In formula (5), R mp is the gas reflection information corresponding to m preset areas, (r m1 ,t m1 ) is the radar measurement information of the first time node in the mth preset area, (r m2 ,t m2 ) is the radar measurement information of the second time node in the mth preset area, (r mp ,t mp ) is the radar measurement information of the p-th time node in the m-th preset area;

[0131] The echo time difference algorithm is used to calculate the current gas reflection information to obtain the gas layer distance between two adjacent time nodes, which is expressed by formula (6). Formula (6) is as follows:

[0132]

[0133] In formula (6), Δr mp is the distance between the gas layers at two adjacent time nodes, v 0 is the speed of light constant;

[0134] Multiple gas layer distances are input into the support vector machine model to obtain the regional gas concentration distribution information corresponding to the current preset area, which is expressed by formula (7). Formula (7) is as follows:

[0135]

[0136] In formula (7), ρ m (x, y, z) is the regional gas concentration distribution information of the prediction point with coordinates (x, y, z) in the current preset area, β a is the first parameter of the support vector machine model, b is the second parameter of the support vector machine model, is a polynomial kernel function, γ is the mapping dimension of the polynomial kernel function, r represents the predicted radar measurement information of the prediction point with coordinates (x, y, z), and t represents the prediction of the prediction point with coordinates (x, y, z) at time node t;

[0137] Repeat the above steps until each preset area corresponds to a regional gas concentration distribution information, and organize the regional gas concentration distribution information into second gas concentration distribution information, which is expressed by formula (8). Formula (8) is as follows:

[0138]

[0139] In formula (8), ρ′(x, y, z) is the second gas concentration distribution information, ρ 1 (x, y, z) is the regional gas concentration distribution information of the first preset area, ρ m (x, y, z) is the regional gas concentration distribution information of the mth preset area.

[0140] This embodiment shows the specific relationship between the radar detector and the industrial building. For the convenience of description, the industrial building is divided into multiple preset areas. The preset area can be understood as a single independent room in the industrial building. The preset area can also be a relatively independent area roughly divided by partitions, boundary lines, etc. in the industrial building. Preferably, the preset area can be defined as an area that can complete a certain process step independently. In this embodiment, at least one radar detector is set in the preset area, and each preset area corresponds to a gas reflection information. It should be noted that the radar detector can measure the gas concentration in the space, and the determination of the gas category needs to rely on the process parameters. Optionally, a spectral image acquisition device can be arranged near the radar detector to make a more accurate determination of the gas category with the help of spectral images. This embodiment will not be described in detail.

[0141] A radar detector is used for each preset area to collect multiple radar measurement information at regular intervals, and then the radar measurement information is sorted in chronological order to obtain the gas reflection information corresponding to the current preset area. This method can obtain the dynamic change data of greenhouse gases in industrial buildings, which is convenient for the generation of the first carbon emission change information in the second preset time period.

[0142] This embodiment divides industrial buildings into multiple preset areas, and deploys radar detectors in each preset area to monitor the greenhouse gas concentration inside industrial buildings. By regularly collecting radar measurement data, a time series of gas reflection information is formed, which lays the foundation for subsequent echo time difference algorithm processing. The echo time difference algorithm is used to calculate the gas layer distance between adjacent time nodes, and input into the support vector machine model to obtain more accurate regional gas concentration distribution information. The gas concentration distribution information of each preset area is integrated to form a complete second gas concentration distribution information, which provides comprehensive data support for industrial carbon emission analysis.

[0143] See also Figure 2 In some embodiments, generating third gas concentration distribution information in an industrial building according to process information includes:

[0144] S201, according to the process flow, the process equipment information, the process raw material information and the personnel configuration information are associated to obtain a process processing data set, wherein the process processing data set includes a plurality of process unit information arranged in the order of the process flow, and each process unit information corresponds to the process equipment, process raw material and processing personnel required for a process step;

[0145] S202, constructing a linear regression model of process unit information and carbon emissions, and obtaining room information of the process unit information in the industrial building;

[0146] S203, establishing a mapping relationship between each process unit information and the industrial building according to the linear regression model and the room information, and calculating the diffusion information of the greenhouse gas of the process unit information in the industrial building according to the concentration field-flow field coupling equation;

[0147] S204: Map and store the mapping relationship and the diffusion information according to the categories of greenhouse gases to obtain third gas concentration distribution information.

[0148] In step S201, it is necessary to integrate the process equipment information, process raw material information and personnel configuration information according to the process flow. It can be understood that the process flow includes multiple process steps, and each process step corresponds to at least one item of process equipment information, process raw material information and personnel configuration information. For the convenience of expression, this correspondence is recorded as process unit information.

[0149] In step S202, a linear regression model of process unit information and carbon emissions is constructed. Since this process does not involve a complex fitting process of carbon emissions information, the carbon emissions information of each process unit information is fixed, and the use of a linear regression model can meet actual needs. At the same time, the room information of the process unit information in the current industrial building is obtained. The room information can be understood as the building space divided by the industrial building to complete this process step. It can be a separate room or a space divided by partitions, warning lines, etc., corresponding to the aforementioned preset area.

[0150] In step S203, a mapping relationship between each process unit information and the industrial building is established based on the linear regression model and the room information. This mapping relationship combines the diffusion characteristics of greenhouse gases and the architectural characteristics of industrial buildings. On this basis, a concentration field-flow field coupling equation is constructed. Specifically, the concentration field-flow field coupling equation is expressed by formula (9), which is as follows:

[0151]

[0152] In formula (9), is the partial derivative of greenhouse gas concentration P with respect to time t, is the gradient operator, θ is the velocity field, P is the greenhouse gas concentration, Ω is the diffusion coefficient, and Λ represents the source of the greenhouse gas. In this step, the source of the greenhouse gas can be understood as the gas output section corresponding to the process unit information. Indicates the change information of greenhouse gas concentration following the flow field, is the gradient vector of the greenhouse gas concentration field in space, It represents the change information of greenhouse gas concentration following the diffusion process, that is, diffusion information.

[0153] In step S204, the mapping relationship and diffusion information are mapped and stored according to the category of greenhouse gas to obtain the third gas concentration distribution information. It should be noted that the category of this greenhouse gas is derived in the process information according to the process and raw material category and the chemical reaction produced, which will not be described in detail in this embodiment.

[0154] This embodiment makes full use of industrial process information, and by analyzing the equipment, raw materials and personnel configuration involved in the process unit, establishes a mapping relationship between the process unit and greenhouse gas emissions, and forms a more comprehensive greenhouse gas distribution analysis. At the same time, combined with the spatial layout information of industrial buildings, the concentration field-flow field coupling equation is used to simulate the diffusion process of greenhouse gases in the industrial space, realizing a refined simulation of greenhouse gas emissions and diffusion, and improving the accuracy of the monitoring results. The mapping relationship and diffusion information of the process unit are integrated to obtain complete third gas concentration distribution information, which is convenient for improving the accuracy of subsequent data processing.

[0155] See also Figure 3 In some embodiments, the first fusion feature is subjected to kernel PCA dimensionality reduction to obtain a first dimensionality reduction feature, and the first dimensionality reduction feature includes a principal component subspace including:

[0156] S301, calculating the kernel function matrix of the first fusion feature, and performing eigenvalue decomposition on the kernel function matrix to obtain a plurality of principal component features sorted in a decomposition order, each principal component feature having an eigenvalue and an eigenvector;

[0157] S302, selecting a plurality of principal component features that are ranked first to form a principal component matrix, and projecting the first fused feature onto the principal component matrix to generate a principal component subspace;

[0158] S303, sorting the principal component subspace to obtain the first dimension reduction feature;

[0159] Perform ICA dimensionality reduction on the principal component subspace to obtain the second dimensionality reduction feature, which includes multiple independent component features. The second dimensionality reduction feature is recorded as the second fusion feature, including:

[0160] S304, performing dimensionality reduction preprocessing on the principal component subspace, the dimensionality reduction preprocessing including centering and whitening, to obtain a first subspace feature;

[0161] S305, applying the FastICA algorithm to the first subspace feature to extract independent components, and obtaining at least one independent component feature corresponding to the current first subspace feature;

[0162] S306: Concatenate the first subspace feature and the independent component feature to obtain a second subspace feature, which is a second dimensionality reduction feature.

[0163] In step S301, the first fused feature is converted into a kernel function matrix, and on this basis, the kernel function matrix is ​​subjected to eigenvalue decomposition to obtain a plurality of principal component features, wherein the eigenvalue of the principal component feature shows the importance of the principal component feature, and the eigenvector of the principal component feature shows the direction of the principal component feature. It should be noted that the decomposition order shown in this embodiment is consistent with the descending order of the eigenvalue size.

[0164] In step S302 and step S303, several principal component features that are sorted in the front are selected, that is, the main principal component features in the first fusion feature are selected, and these principal component features are sorted to obtain a principal component matrix. On this basis, the original first fusion feature is projected onto the principal component matrix, that is, the first fusion feature is expressed as a linear combination of the principal component matrix, thereby obtaining the projection coefficient of the first fusion feature on the principal component matrix, that is, constituting the spatial coordinate representation of the first fusion feature in this principal component matrix, that is, generating a principal component subspace, and then sorting the principal component subcontrols into the first dimensionality reduction feature.

[0165] In step S304, the principal component subspace is preprocessed for dimensionality reduction, wherein centering can be understood as mean processing of the principal component subspace to satisfy the zero mean distribution, and whitening can be understood as normalization processing to remove the linear correlation in the principal component subcontrols, thereby obtaining the first subspace feature.

[0166] In step S305, an independent component analysis algorithm, namely, the FastICA algorithm, is applied to the first subspace feature to select mutually independent component features, which are recorded as independent component features. It can be understood that the first subspace feature may include multiple independent component features, which are specifically determined according to the actual analysis results.

[0167] In step S306, the first subspace feature is concatenated with the independent component feature to obtain a second subspace feature, thereby forming a richer feature representation, namely, a second dimensionality reduction feature.

[0168] This embodiment can capture the main change pattern in the first fusion feature through eigenvalue decomposition and principal component selection of kernel PCA. Combined with independent component analysis of ICA, it can further extract independent potential features to form a richer feature representation, which can capture the linear and nonlinear correlations in the original features, making the final features more rich and robust in expression; effectively avoiding overfitting and improving the generalization performance of machine learning models on new data.

[0169] In some embodiments, the carbon emission prediction model is trained by the following steps:

[0170] Constructing a neural network model, and inputting the first sample features of the first sample preset time period into the input layer of the neural network model;

[0171] A GRU time series modeling layer is added to the neural network model, and the first sample feature is input into the GRU time series modeling layer to extract the feature of the time series, so as to obtain the first sample time series feature;

[0172] Input the sample time series features into the fully connected layer to obtain the second sample time series features;

[0173] Input the second sample time series feature into the output layer to obtain the second sample feature;

[0174] The aforementioned steps are repeated until the prediction accuracy of the second sample feature reaches a preset accuracy threshold, thereby obtaining a carbon emission prediction model.

[0175] Specifically, this embodiment constructs a neural network model as the basis for carbon emission prediction. The neural network model includes an input layer, a hidden layer, and an output layer. In the input layer of the neural network model, the first sample feature is input. It should be noted that the construction of the neural network model in this embodiment requires the introduction of a sample database, in which the sample database has a first sample feature of a preset time period of the first sample and a second actual sample feature of a preset time period of the second sample. For the convenience of distinction, the prediction result obtained based on the neural network model is recorded as the second sample feature. On this basis, a GRU time series modeling layer is added to the neural network model. The GRU time series modeling layer is a recurrent neural network unit that is good at modeling time series data. The first sample feature is input into the GRU time series modeling layer, and the GRU will extract the features of the sample in the time series. The output of the GRU time series modeling layer is the first sample time series feature, which already contains the potential pattern of the first sample feature in the time dimension. The first sample time series feature is input into the fully connected layer, and the fully connected layer can further extract and combine these time series features to obtain the second sample time series feature. The second sample time series feature is input into the output layer, and the output layer will give the prediction result of the carbon emission of the second sample. Repeat the above steps and use different sample data for training until the prediction accuracy of the neural network model on the validation set reaches the preset threshold, and the final carbon emission prediction model can be obtained.

[0176] This embodiment shows the training process of the carbon emission prediction model. The carbon emission prediction model shown in this embodiment can effectively use time series information to improve prediction performance. By introducing the GRU time series modeling layer in the neural network, the carbon emission prediction model can capture the changing rules of sample features in the time dimension, thereby more accurately describing the dynamic evolution of carbon emissions and greatly improving the accuracy of the prediction. During the training process, the carbon emission prediction model will continuously optimize the parameters until the preset accuracy threshold is reached on the validation set, adapting to various complex carbon emission scenarios and improving the generalization performance on new data.

[0177] In some embodiments, a BIM model is constructed according to the structural parameter information and the first carbon emission change information to obtain carbon emission assessment information of industrial buildings, including:

[0178] Input the structural parameter information into the BIM model to obtain a three-dimensional model of the industrial building, wherein the industrial building has a plurality of preset areas, and each preset area has a regional three-dimensional model;

[0179] Dividing the first carbon emission change information into a plurality of second carbon emission change information according to a preset area;

[0180] Associating the second carbon emission change information with the regional three-dimensional model to obtain regional carbon emission flow data;

[0181] A carbon emission heat map of a preset area is constructed based on the regional carbon emission flow data, and multiple carbon emission heat maps are correlated and sorted according to the process flow to obtain carbon emission assessment information.

[0182] In this embodiment, a complete three-dimensional model of the industrial building is constructed based on the structural parameter information of the industrial building (such as size, material, etc.). At the same time, the entire industrial building is divided into multiple preset areas, and the complete three-dimensional model is synchronously matched according to the preset areas, so that each preset area has a regional three-dimensional model.

[0183] The output result of the carbon emission prediction model, that is, the first carbon emission change information, is used as the input feature of the BIM model. According to the preset area of ​​the industrial building, the first carbon emission change information is divided into multiple second carbon emission change information corresponding to the carbon emission change of each preset area. At the same time, the second carbon emission change information of each area is associated with the regional three-dimensional model. Specifically, the spatial association query algorithm can be used to quickly associate and match the second carbon emission change information with the regional three-dimensional model. According to the process flow, the regional carbon emission flow data corresponding to each preset area can be obtained in the BIM model. Combined with the overall building characteristics of the BIM model, the flow relationship of carbon emissions between multiple preset areas can be clarified.

[0184] On this basis, the regional carbon emission flow data is used to construct a carbon emission heat map for each preset area. The IDW interpolation algorithm can be used to generate a continuous carbon emission heat map. Preferably, the carbon emission heat map can be further combined and rendered with the BIM model by means of a graphics rendering algorithm, so that the carbon emission distribution in the industrial building is more intuitive. Then, multiple carbon emission heat maps are sorted according to the process flow to obtain carbon emission evaluation information.

[0185] This embodiment gives full play to the advantages of BIM technology and can construct a detailed three-dimensional building model, which is conducive to more accurately describing the geometric characteristics of the building. By associating carbon emission data with the regional three-dimensional model, a more detailed regional carbon emission flow situation can be obtained. The carbon emission heat map is used to intuitively display the carbon emission distribution of each region, which can further enhance the visualization effect and improve the intuitiveness and interactivity of the analysis results. The carbon emission heat maps of multiple preset areas are integrated according to the process flow to obtain complete carbon emission evaluation information. It can better reflect the transmission mechanism of carbon emissions inside the building and provide decision support for optimizing production processes and implementing targeted energy conservation and emission reduction.

[0186] See also Figure 4 In the second aspect, the present embodiment provides a spatial carbon emission monitoring system 1, which is applicable to the spatial carbon emission monitoring method described in the first aspect. The system includes an information processing module 11, an information fusion module 12, a carbon emission processing module 13 and a building carbon emission analysis module 14. The information processing module 11 includes a spectral information processing unit, a radar information processing unit and a process information processing unit. The spectral information processing unit is used to obtain spectral information within a first preset time period. The spectral information is configured to be acquired through remote sensing sensors. The spectral information is obtained through an inversion algorithm to obtain first gas concentration distribution information in the industrial area. The first gas concentration distribution information includes a first greenhouse gas category, a first greenhouse gas concentration and a first greenhouse gas distribution feature; the radar information processing unit Used to obtain gas reflection information in an industrial building within a first preset time period, the gas reflection information is configured to be obtained by measuring with a radar detector, and the gas reflection information is used to obtain second gas concentration distribution information in the industrial building through an echo time difference algorithm, and the second gas concentration distribution information includes a second greenhouse gas category, a second greenhouse gas concentration, and a second greenhouse gas distribution characteristic; the process information processing unit is used to obtain process information in the industrial building within the first preset time period, the process information includes process equipment information, process raw material information, and personnel configuration information, and generate third gas concentration distribution information in the industrial building according to the process information, and the third gas concentration distribution information includes a third greenhouse gas category, a third greenhouse gas concentration, and a third greenhouse gas distribution characteristic;

[0187] The information fusion module 12 is used to fuse the first gas concentration distribution information, the second gas concentration distribution information and the third gas concentration distribution information to obtain a first fusion feature, which is a high-dimensional fusion feature; the first fusion feature is subjected to kernel PCA dimensionality reduction to obtain a first dimensionality reduction feature, which includes a principal component subspace, and the principal component subspace is subjected to ICA dimensionality reduction to obtain a second dimensionality reduction feature, which includes multiple independent component features, and the second dimensionality reduction feature is recorded as a second fusion feature; the carbon emission processing module 13 is used to input the second fusion feature into the carbon emission prediction model to obtain the first carbon emission change information within a second preset time period, and the carbon emission prediction model is configured to be constructed based on a neural network model; the building carbon emission analysis module 14 is used to obtain the structural parameter information of the industrial building, and construct a BIM model based on the structural parameter information and the first carbon emission change information, to obtain and display the carbon emission evaluation information of the industrial building, and the carbon emission evaluation information includes multiple carbon emission thermal distribution diagrams in the industrial building.

[0188] The steps shown in this embodiment can be understood by referring to the text described in the first aspect above, and this embodiment will not be elaborated on in detail.

[0189] In a third aspect, this embodiment further provides a computer-readable storage medium on which computer program instructions are stored. When the computer program instructions are executed by a processor, the method described in the first aspect is implemented.

[0190] The computer program involved in this embodiment can be stored in a computer device readable storage medium, which includes but is not limited to a disk, a tape, a magnetic card, a floppy disk, a flash memory, an optical disk, an optical card, a read-only memory (ROM), a random access memory (RAM), an erasable programmable ROM (EPROM) and an electrically erasable programmable ROM (EEPROM), etc., and also includes other biological, physical or chemical structures that can achieve the same or equivalent functions as the storage media listed above, such as DNA, RNA, protein and other units with information storage capabilities. In a specific embodiment, the storage medium involved can be one of the above-mentioned media types, or a combination of the above-mentioned media types. In different embodiments, the computer program involved in the embodiment can be stored in a single medium in a centralized manner, or it can be stored in multiple media in a distributed manner. The memory containing the computer device readable storage medium can be a non-volatile memory or a random access memory. These computer device readable storage media can be built into the device, or they can be connected to the device involved in the embodiment as an external device or a part of an external device. In some embodiments, a memory with a computer device readable storage medium is deployed locally; in other embodiments, a solution of deploying the memory away from the processor may also be adopted, such as a network attached memory accessed via an RF circuit or an external port and a communication network, wherein the communication network may be the Internet, one or more intranets, a local area network (LAN), a wide area network (WLAN), a storage area network (SAN), etc., or a suitable combination thereof, as long as the computer device can access the memory. In addition, the computer program involved in the embodiment may be stored in plaintext / ciphertext form, or may be designed as training data, which may be integrated and reorganized through model training and implicitly stored in the parameter state of a deep neural network or other machine learning model.

[0191] In a fourth aspect, this embodiment further provides an electronic device, comprising a memory and a processor, wherein the memory is used to store one or more computer program instructions, wherein the one or more computer program instructions are executed by the processor to implement the method described in the first aspect.

[0192] The processor described in this embodiment can be implemented by hardware, firmware, software or a combination thereof, and can use a circuit, a single or multiple application-specific integrated circuits (Application Specific Integrated Circuit, ASIC), a digital signal processor (Digital Signal Processor, DSP), a digital signal processing device (Digital Signal Processing Device, DSPD), a programmable logic device (Programmable Logic Device, PLD), a field programmable gate array (Field Programmable Gate Array, FPGA), a central processing unit (Central Processing Unit, CPU), a controller, a microcontroller, at least one of a microprocessor, and also includes other physical, biological or chemical structures that can achieve similar or equivalent functions to the processors listed above, such as biological neurons, quantum computing units, DNA computing units, etc., so that the processor can execute some or all of the steps in the computer program or method involved in the various embodiments of the present application, or any combination of the steps mentioned therein.

[0193] Different from the existing technology, the above technical solution adopts automated monitoring methods such as remote sensing and radar, and integrates multiple data sources such as spectral information, radar detectors, and process parameters. It can fully perceive the carbon emission status inside industrial buildings, and can realize real-time collection and dynamic monitoring of spatial carbon emissions of industrial buildings, instead of relying on manual records and data reporting, reducing the problems of time lag and missing edge data. Through deep learning technologies such as kernel PCA, ICA, and neural networks, more effective data features can be extracted from multidimensional features, making the predicted carbon emissions more accurate, thereby improving the accuracy of monitoring. By integrating monitoring data into the BIM model, the visualization of carbon emission information is realized, which is convenient for managers to grasp and analyze the carbon emission dynamics of industrial buildings in a timely manner.

[0194] Finally, it should be noted that although the above embodiments have been described in the specification and drawings of this application, this does not limit the scope of patent protection of this application. All technical solutions generated by replacing or modifying equivalent structures or equivalent processes based on the essential concept of this application using the contents recorded in the specification and drawings of this application, as well as directly or indirectly implementing the technical solutions of the above embodiments in other related technical fields, are included in the scope of patent protection of this application.

Claims

1. A spatial carbon emission monitoring method, characterized in that: Applicable to industrial areas, including industrial buildings, the method comprises: Acquire spectral information within a first preset time period, the spectral information being configured to be acquired through a remote sensing sensor, and obtain first gas concentration distribution information within the industrial area through an inversion algorithm using the spectral information, the first gas concentration distribution information including a first greenhouse gas category, a first greenhouse gas concentration, and a first greenhouse gas distribution feature; and obtaining gas reflection information in the industrial building within a first preset time period, the gas reflection information being configured to be obtained by measuring with a radar detector, and obtaining second gas concentration distribution information in the industrial building by using an echo time difference algorithm with the gas reflection information, the second gas concentration distribution information including a second greenhouse gas category, a second greenhouse gas concentration, and a second greenhouse gas distribution feature; and, obtaining process information in the industrial building within a first preset time period, the process information including process equipment information, process raw material information and personnel configuration information, and generating third gas concentration distribution information in the industrial building according to the process information, the third gas concentration distribution information including third greenhouse gas category, third greenhouse gas concentration and third greenhouse gas distribution characteristics; Performing feature fusion on the first gas concentration distribution information, the second gas concentration distribution information, and the third gas concentration distribution information to obtain a first fusion feature, where the first fusion feature is a high-dimensional fusion feature; Perform kernel PCA dimensionality reduction on the first fusion feature to obtain a first dimensionality reduction feature, the first dimensionality reduction feature includes a principal component subspace, perform ICA dimensionality reduction on the principal component subspace to obtain a second dimensionality reduction feature, the second dimensionality reduction feature includes a plurality of independent component features, and record the second dimensionality reduction feature as a second fusion feature; Inputting the second fusion feature into a carbon emission prediction model to obtain first carbon emission change information within a second preset time period, wherein the carbon emission prediction model is configured to be constructed based on a neural network model; Structural parameter information of the industrial building is obtained, and a BIM model is constructed according to the structural parameter information and the first carbon emission change information, so as to obtain and display carbon emission evaluation information of the industrial building, wherein the carbon emission evaluation information includes a plurality of carbon emission thermal distribution diagrams in the industrial building.

2. The spatial carbon emission monitoring method according to claim 1, characterized in that: The inversion algorithm includes a differential absorption method inversion operation and a radiation transmission optimization operation. The spectrum information is obtained by the inversion algorithm to obtain the first gas concentration distribution information in the industrial area, including: The spectral information is preprocessed, and the spectral information preprocessing includes radiation correction and Gaussian filtering to obtain first processed spectral information, which is expressed by formula (1). The formula (1) is as follows: In formula (1), L′(λ) is the spectral information after radiation correction, L1(λ) is the atmospheric scattering path radiation information, τ is the atmospheric transmittance, L2(λ) is the surface reflection radiation information, L3(λ) is the atmospheric downlink radiation information, L″(λ) is the first processed spectral information, λ0 is the central wavelength of the spectral information, λ is the wavelength value to be processed in the spectral information, σ is the standard deviation of the Gaussian kernel, and N is the total wavelength value in the spectral information; Spectral features are extracted from the first processed spectral information to obtain a plurality of first spectral features, and the plurality of first spectral features are matched according to greenhouse gas categories to obtain a plurality of first greenhouse gas categories, which are expressed by formula (2). Formula (2) is as follows: Q i =f(F)=UΣ′V T ; In formula (2), Q i is the i-th greenhouse gas category, f(F) is the classification function, F is the spectral feature matrix formed by multiple first spectral features, U is the left singular vector matrix, the column vector of U is the principal component vector of F, Σ′ is the diagonal matrix containing the singular values ​​of the principal component vector of F, V T is the transpose of the right singular vector matrix, V T The row vector of is the projection of the spectral information in the principal component space of F; The first processed spectral information is subjected to a differential absorption inversion operation according to the first greenhouse gas category to obtain first initial gas concentration information, which is expressed by formula (3). Formula (3) is as follows: In formula (3), is the first initial gas concentration information corresponding to the i-th first greenhouse gas category, dL″(λ) is the differential change of the spectral brightness of the λ wavelength value, α i is the gas absorption coefficient of the ith greenhouse gas category, l is the optical path length; The first initial gas concentration information is subjected to a radiation transmission optimization operation to obtain the first final gas concentration information, which is expressed by formula (4). Formula (4) is as follows: In formula (4), is the first final gas concentration information corresponding to the i-th first greenhouse gas category, I0 is the initial radiation intensity, I(s,k,j) is the radiation intensity after passing through the atmosphere, k is the extinction coefficient, j is the emissivity, s is the optical path length from the observation point where the spectral information is collected to the upper boundary of the atmosphere, s′ is the micro optical path length from the atmospheric incident point to the integral position, and s′′ is the micro optical path length from the integral position to the observation point where the spectral information is collected; The first final gas concentration information is mapped and stored with the first greenhouse gas category to obtain the first gas concentration distribution information.

3. The spatial carbon emission monitoring method according to claim 1, characterized in that: Generating the third gas concentration distribution information in the industrial building according to the process information includes: According to the process flow, the process equipment information, the process raw material information and the personnel configuration information are associated to obtain a process processing data set, wherein the process processing data set includes a plurality of process unit information arranged in the order of the process flow, each of the process unit information corresponds to the process equipment, process raw material and processing personnel required for a process step; Constructing a linear regression model between the process unit information and carbon emissions, and obtaining room information of the process unit information in the industrial building; Establishing a mapping relationship between each of the process unit information and the industrial building according to the linear regression model and the room information, and calculating the diffusion information of the greenhouse gas of the process unit information in the industrial building according to the concentration field-flow field coupling equation; The mapping relationship and the diffusion information are mapped and stored according to the categories of greenhouse gases to obtain the third gas concentration distribution information.

4. The spatial carbon emission monitoring method according to claim 1, characterized in that: The first fusion feature is subjected to kernel PCA dimensionality reduction to obtain a first dimensionality reduction feature, where the first dimensionality reduction feature includes a principal component subspace, specifically including: Calculating a kernel function matrix of the first fusion feature, and performing eigenvalue decomposition on the kernel function matrix to obtain a plurality of principal component features sorted in a decomposition order, each of the principal component features having an eigenvalue and an eigenvector; Selecting a plurality of principal component features that are ranked first to form a principal component matrix, and projecting the first fused features onto the principal component matrix to generate a principal component subspace; Arrange the principal component subspace to obtain the first dimensionality reduction feature; Performing ICA dimensionality reduction on the principal component subspace to obtain a second dimensionality reduction feature, wherein the second dimensionality reduction feature includes a plurality of independent component features, and recording the second dimensionality reduction feature as a second fusion feature includes: Performing dimensionality reduction preprocessing on the principal component subspace, wherein the dimensionality reduction preprocessing includes centering and whitening, to obtain a first subspace feature; Applying the FastICA algorithm to the first subspace feature to extract independent components, to obtain at least one independent component feature corresponding to the current first subspace feature; The first subspace feature and the independent component feature are concatenated to obtain a second subspace feature, that is, the second dimensionality reduction feature.

5. The spatial carbon emission monitoring method according to claim 1, characterized in that: The carbon emission prediction model is trained by the following steps: Constructing a neural network model, and inputting first sample features of a first sample preset time period into an input layer of the neural network model; Adding a GRU time series modeling layer to the neural network model, and inputting the first sample feature into the GRU time series modeling layer to extract the feature of the time series, so as to obtain the first sample time series feature; Inputting the sample time series features into a fully connected layer to obtain a second sample time series feature; Inputting the second sample time series feature into the output layer to obtain the second sample feature; The aforementioned steps are repeated until the prediction accuracy of the second sample feature reaches a preset accuracy threshold, thereby obtaining the carbon emission prediction model.

6. The spatial carbon emission monitoring method according to claim 1, characterized in that: The BIM model is constructed according to the structural parameter information and the first carbon emission change information to obtain the carbon emission evaluation information of the industrial building, including: Inputting the structural parameter information into the BIM model to obtain a three-dimensional model of the industrial building, wherein the industrial building has a plurality of preset areas, and each of the preset areas has a regional three-dimensional model; Dividing the first carbon emission change information into a plurality of second carbon emission change information according to preset regions; Associating the second carbon emission change information with the regional three-dimensional model to obtain regional carbon emission flow data; A carbon emission heat map of the preset area is constructed according to the regional carbon emission flow data, and a plurality of the carbon emission heat maps are correlated and sorted according to the process flow to obtain the carbon emission evaluation information.

7. A spatial carbon emission monitoring system, characterized in that: The spatial carbon emission monitoring method according to any one of claims 1 to 6 is applicable, wherein the system comprises: The information processing module includes a spectral information processing unit, a radar information processing unit and a process information processing unit, wherein the spectral information processing unit is used to obtain spectral information within a first preset time period, wherein the spectral information is configured to be acquired through remote sensing sensor collection, and the spectral information is used to obtain first gas concentration distribution information in the industrial area through an inversion algorithm, wherein the first gas concentration distribution information includes a first greenhouse gas category, a first greenhouse gas concentration and a first greenhouse gas distribution characteristic; the radar information processing unit is used to obtain gas reflection information in the industrial building within a first preset time period, wherein the gas reflection information is configured to be acquired through radar detector measurement, and the gas reflection information is used to obtain second gas concentration distribution information in the industrial building through an echo time difference algorithm, wherein the second gas concentration distribution information includes a second greenhouse gas category, a second greenhouse gas concentration and a second greenhouse gas distribution characteristic; the process information processing unit is used to obtain process information in the industrial building within a first preset time period, wherein the process information includes process equipment information, process raw material information and personnel configuration information, and third gas concentration distribution information in the industrial building is generated according to the process information, wherein the third gas concentration distribution information includes a third greenhouse gas category, a third greenhouse gas concentration and a third greenhouse gas distribution characteristic; An information fusion module is used to perform feature fusion on the first gas concentration distribution information, the second gas concentration distribution information and the third gas concentration distribution information to obtain a first fusion feature, where the first fusion feature is a high-dimensional fusion feature; perform kernel PCA dimensionality reduction on the first fusion feature to obtain a first dimensionality reduction feature, where the first dimensionality reduction feature includes a principal component subspace, perform ICA dimensionality reduction on the principal component subspace to obtain a second dimensionality reduction feature, where the second dimensionality reduction feature includes multiple independent component features, and record the second dimensionality reduction feature as a second fusion feature; a carbon emission processing module, configured to input the second fusion feature into a carbon emission prediction model to obtain first carbon emission change information within a second preset time period, wherein the carbon emission prediction model is configured to be constructed based on a neural network model; The building carbon emission analysis module is used to obtain the structural parameter information of the industrial building, and build a BIM model based on the structural parameter information and the first carbon emission change information, to obtain and display the carbon emission evaluation information of the industrial building, wherein the carbon emission evaluation information includes multiple carbon emission thermal distribution diagrams in the industrial building.

8. A computer-readable storage medium storing computer program instructions, characterized in that: The computer program instructions implement the method according to any one of claims 1 to 6 when executed by a processor.

9. An electronic device, comprising a memory and a processor, characterized in that: The memory is used to store one or more computer program instructions, wherein the one or more computer program instructions are executed by the processor to implement the method according to any one of claims 1 to 6.

Images