CN-121070980-B - Log report reiteration method and system based on carbon emission monitoring

CN121070980BCN 121070980 BCN121070980 BCN 121070980BCN-121070980-B

Abstract

The invention provides a journal report recordation method and a journal report recordation system based on carbon emission monitoring, which are characterized in that original carbon emission monitoring journal data comprising a time stamp sequence and consisting of a plurality of monitoring items are firstly obtained, each item comprises fields such as emission source identification, and the like, then feature extraction is carried out on the original carbon emission monitoring journal data to obtain a structured journal feature group comprising core emission parameter features and cross-item time sequence association features, then a pre-built model is called to carry out semantic reconstruction on the feature group, a journal recordation logic framework comprising emission event causal relation chains and key information distribution patterns is generated, an initial recordation journal report is generated based on the journal recordation logic framework, and finally semantic consistency check and structural integrity check are carried out on the initial recordation journal report to generate a final recordation journal report, so that the carbon emission monitoring journal data can be processed efficiently, and an accurate, clear and strong-logic report is generated.

Inventors

XUE SHIWEI
SUN JIAN
ZHANG ZHIJIE
WANG HAIBIN
ZUO TINGTING
Miao Zhenxuan
WANG JIANZHEN
ZHAO GUOZHENG
HU XIAOYAN

Assignees

中建碳科技有限公司
中建生态环境集团有限公司

Dates

Publication Date: 20260505
Application Date: 20250805

Claims (7)

1. A method for carbon emission monitoring-based log reporting replication, the method comprising: Acquiring original carbon emission monitoring log data comprising a time stamp sequence, wherein the original carbon emission monitoring log data consists of a plurality of continuously acquired monitoring items, and each monitoring item comprises an emission source identification field, a monitoring parameter description field and a state change record field; Extracting features of the original carbon emission monitoring log data to obtain a structured log feature set, wherein the structured log feature set comprises core emission parameter features and cross-item time sequence association features of each monitoring item; invoking a pre-constructed log report re-description logic model to carry out semantic reconstruction processing on the structured log feature group, and generating a log re-description logic framework, wherein the log re-description logic framework comprises a causal relationship chain and an important information distribution map of an emission event; Executing text generation operation based on the journal re-description logic framework, and generating an initial re-description journal report, wherein the initial re-description journal report comprises emission event description contents organized according to causal relationship chains; performing two-dimensional verification processing on the initial review log report to generate a final review log report passing verification, wherein the two-dimensional verification processing comprises semantic consistency verification and structural integrity verification; the invoking the pre-constructed log report restating logic model performs semantic reconstruction processing on the structured log feature group to generate a log restating logic framework, which comprises the following steps: Inputting the structured log feature group into a feature input layer of the log report replication logic model, and performing feature dimension adjustment processing to obtain an adjusted feature group matched with the input dimension of the log report replication logic model; Carrying out relation reasoning processing on the adjusted feature groups through a causal relation extraction module of the log report and review logic model, and identifying causal relation types among monitoring items, wherein the causal relation types comprise state change relations caused by parameter changes and trend continuation relations caused by time continuity; constructing a causal relationship chain according to the causal relationship type, wherein the causal relationship chain consists of a node and an edge which represent causal relationship, the node represents a monitoring item, and the edge represents the causal relationship type; Performing information density calculation processing on the adjusted feature set through a key information detection module of the log report re-description logic model, and extracting monitoring items with information density exceeding a preset threshold as key information nodes; performing weight accumulation processing on the core emission parameter characteristics of the key information nodes to obtain comprehensive weight values of the key information nodes; Performing weight labeling processing on the nodes in the causal relationship chain based on the comprehensive weight values to generate a key information distribution map containing node weight information; combining the causal relation chain and the key information distribution map to generate a log replication logic framework; The causal relation extraction module for re-describing the logic model through the log report carries out relation reasoning processing on the adjusted feature group, and identifies causal relation types among monitoring items, and the causal relation extraction module comprises the following steps: Extracting core emission parameter characteristics of adjacent monitoring items in the adjusted characteristic group to obtain precursor item parameter characteristics and subsequent item parameter characteristics; performing differential analysis processing on the precursor item parameter characteristics and the subsequent item parameter characteristics to obtain parameter change direction and change amplitude information; Extracting cross-item time sequence association features of adjacent monitoring items in the adjusted feature group to obtain a time sequence association strength value; The method comprises the steps of establishing a causal relation judging rule, wherein the causal relation judging rule is used for judging a state change relation caused by parameter change when a parameter change direction is consistent with a state change direction and a time sequence association strength value exceeds a preset threshold value, and judging a trend continuation relation caused by time continuity when the parameter change direction is inconsistent with the state change direction and the time sequence association strength value exceeds the preset threshold value; performing relationship judgment processing on adjacent monitoring items through the causality judgment rule to obtain a direct causality type; Performing skip relation reasoning processing on the non-adjacent monitoring items, and deducing an indirect causal relation type of the non-adjacent monitoring items based on the causal relation type of the intermediate monitoring items; taking the direct causal relationship type and the indirect causal relationship type as causal relationship types between monitoring items; The text generation operation is executed based on the journal replication logic framework to generate an initial replication journal report, which comprises the following steps: Extracting a causal relation chain in the log review logic framework, and determining the description sequence of the emission event, wherein the description sequence is arranged according to the node sequence of the causal relation chain; Extracting key information distribution patterns in the log review logic framework, and determining the description detail degree of each node, wherein the description detail degree and the comprehensive weight value of the node are in positive correlation; performing text template matching processing on each node in the causal relationship chain, wherein the text template comprises a sentence structure corresponding to the causal relationship type; performing content filling processing on the parameter placeholders in the text template according to the description detail degree, wherein the content filling processing comprises filling the specific description contents of the core emission parameter characteristics; performing sentence consistency adjustment processing on the filled text template, and enhancing the logic linkage between sentences by adding connection adverbs and time-like words; And splicing the adjusted text templates according to the description sequence to generate an initial reiteration log report containing the emission event description contents organized according to the causal relationship chain.
2. The carbon emission monitoring-based log reporting method of claim 1, wherein the feature extraction of the raw carbon emission monitoring log data to obtain a structured log feature set comprises: Performing field analysis processing on the original carbon emission monitoring log data, and extracting an emission source identification field value, a monitoring parameter description field value and a state change record field value of each monitoring item; Carrying out keyword filtering processing on the monitoring parameter description field value, screening out parameter words directly related to carbon emission as core parameter candidate words, wherein the core parameter candidate words comprise emission type limiting words, measurement unit words and threshold description words; Carrying out semantic intensity analysis processing on the state change record field value, calculating a degree adverb weight value of state change description, and carrying out importance ranking processing on core parameter candidate words based on the degree adverb weight value to generate core emission parameter characteristics containing weight information; Performing interval calculation processing on the timestamp sequences of the adjacent monitoring items to obtain time interval values of the adjacent monitoring items; performing similarity calculation processing on the core emission parameter characteristics of the adjacent monitoring items to obtain parameter variation similarity values of the adjacent monitoring items; Constructing a cross-item time sequence association model based on the time interval value and the parameter variation similarity value, calculating time sequence association strength values of adjacent monitoring items through the cross-item time sequence association model, and generating cross-item time sequence association characteristics containing association strength information; And carrying out dimension standardization operation processing on the core emission parameter characteristics and the cross-entry time sequence association characteristics to generate a structured log characteristic group.
3. The carbon emission monitoring-based log report replication method of claim 2, wherein the performing semantic intensity analysis processing on the state change record field value, calculating a degree adverb weight value of a state change description, performing importance ranking processing on core parameter candidate words based on the degree adverb weight value, and generating core emission parameter features including weight information includes: establishing a carbon emission state description degree sub word library, wherein the degree sub word library comprises a vocabulary set for representing the variation amplitude; Performing adverb recognition processing on the state change record field value, and extracting an adverb vocabulary matched with the degree adverb library as a state degree adverb; Distributing a preset weight coefficient value for each state degree adverb, wherein the weight coefficient value and the remarkable degree of state change are in positive correlation; counting the co-occurrence times of each core parameter candidate word and the state degree adverbs, and calculating the co-occurrence frequency value of each core parameter candidate word; Multiplying the co-occurrence frequency value by a weight coefficient value of the corresponding state degree adverb to obtain an importance grading value of each core parameter candidate word; Performing descending order arrangement processing on the core parameter candidate words according to the importance score value to generate a core parameter sequence ordered according to the importance; And allocating a corresponding importance scoring value for each parameter in the core parameter sequence, and generating a core emission parameter characteristic containing weight information.
4. The carbon emission monitoring-based log reporting method of claim 2, wherein the constructing a cross-entry timing correlation model based on the time interval value and the parameter variation similarity value, calculating a timing correlation strength value of an adjacent monitoring entry by the cross-entry timing correlation model, generating a cross-entry timing correlation feature including correlation strength information, comprises: normalizing the time interval value to obtain a time interval normalized value, and normalizing the parameter variation similarity value to obtain a parameter similarity normalized value; calculating a time sequence correlation intensity value of the adjacent monitoring item through a time sequence correlation intensity calculation formula, wherein the time sequence correlation intensity calculation formula is a time sequence correlation intensity value = inverse of a time interval normalization value multiplied by a parameter similarity normalization value; carrying out sliding window average processing on the time sequence correlation intensity values of the continuous K monitoring items to obtain a long-range time sequence correlation intensity value; Taking the time sequence association strength value and the long-range time sequence association strength value as components of a cross-item time sequence association characteristic; And allocating different feature dimensions for the time sequence association strength value and the long-range time sequence association strength value, and generating a cross-item time sequence association feature containing association strength information.
5. The carbon emission monitoring-based log report replication method of claim 1, wherein the text template matching process is performed on each node in the causal relationship chain, the text template includes a sentence structure corresponding to a causal relationship type, and the method comprises: Establishing a mapping relation table of causal relation types and text templates, wherein a state change relation caused by parameter change in the mapping relation table corresponds to a first type of text templates, and a trend continuation relation caused by time continuity corresponds to a second type of text templates; identifying a causal relationship type corresponding to each node in the causal relationship chain; matching corresponding text templates from the mapping relation table according to the causal relation types; Carrying out parameter placeholder marking processing on the text template, wherein the parameter placeholders comprise a precursor parameter placeholder, a follow-up state placeholder, a time interval placeholder and a parameter trend placeholder; extracting core emission parameter characteristics of a precursor monitoring item and a subsequent monitoring item corresponding to the node to obtain contents needing to be filled into a precursor parameter placeholder and a subsequent state placeholder; extracting the time stamp information corresponding to the node, and calculating a time interval value to obtain the content needing to be filled into the time interval placeholder; Extracting parameter change trend information corresponding to the nodes to obtain target contents which need to be filled into parameter trend placeholders; and carrying out one-to-one correspondence processing on the target content and the parameter placeholders to complete text template matching processing.
6. The carbon emission monitoring-based log report replication method of claim 1, wherein performing a two-dimensional verification process on the initial replication log report to generate a verified final replication log report comprises: Inputting the initial review log report and the original carbon emission monitoring log data into a pre-trained semantic similarity model, and calculating semantic similarity scoring values of the initial review log report and the original carbon emission monitoring log data; extracting the core emission parameter description content in the initial review log report, and performing matching degree check processing on the core emission parameter description content and the core emission parameter characteristics in the original carbon emission monitoring log data to obtain a parameter matching degree scoring value; Extracting causal relation chain description content in the initial re-description log report, carrying out node quantity checking processing with causal relation chains in the log re-description logical framework to obtain node integrity grading values, extracting key information description content in the initial re-description log report, carrying out weight matching checking processing with a key information distribution map in the log re-description logical framework to obtain key integrity grading values, and carrying out weighted average processing on the node integrity grading values and the key integrity grading values to obtain structural integrity checking scores; and performing verification judgment processing on the initial repeated log report according to a verification passing rule, outputting the initial repeated log report as a final repeated log report if the verification passing rule is judged to pass, and returning to a correction step of the log repeated logic framework to execute text generation operation if the verification passing rule is judged not to pass, wherein the verification passing rule is that the semantic consistency verification score exceeds a first preset threshold value and the structural integrity verification score exceeds a second preset threshold value.
7. A carbon emission monitoring based log report copying system comprising a processor and a memory, the memory being coupled to the processor, the memory being for storing programs, instructions or code, the processor being for executing the programs, instructions or code in the memory to implement the carbon emission monitoring based log report copying method of any of the preceding claims 1-6.

Description

Log report reiteration method and system based on carbon emission monitoring Technical Field The invention relates to the technical field of artificial intelligence, in particular to a log report reiteration method and system based on carbon emission monitoring. Background In the big background of actively coping with climate change nowadays, carbon emission monitoring and management become key links. Various carbon emission bodies such as enterprises, institutions and the like need to accurately and clearly record and report carbon emission conditions so as to meet supervision requirements, optimize self carbon emission management and show environmental responsibility to the public. Currently, carbon emissions monitoring typically produces a large amount of raw log data in the form of a plurality of continuously collected monitoring entries, each entry containing information such as emissions source identification, monitoring parameter descriptions, and status change records. However, the existing log processing method is often simply to store and basically query the raw carbon emission monitoring log data, and lacks an effective integrating and analyzing means. Because the original log data is large in volume and complex in format, the manual reading and analysis of the data is not only low in efficiency, but also easy to make mistakes, and the overall situation, the causality and the key information of the carbon emission event are difficult to quickly and accurately grasp. In addition, when generating the log report, a scientific and reasonable method is also lacked, so that the report content may have the problems of incomplete information, unclear logic and the like. Disclosure of Invention In view of the above-mentioned problems, with reference to the first aspect of the present invention, an embodiment of the present invention provides a method for reporting and repeating a log based on carbon emission monitoring, the method including: Acquiring original carbon emission monitoring log data comprising a time stamp sequence, wherein the original carbon emission monitoring log data consists of a plurality of continuously acquired monitoring items, and each monitoring item comprises an emission source identification field, a monitoring parameter description field and a state change record field; Extracting features of the original carbon emission monitoring log data to obtain a structured log feature set, wherein the structured log feature set comprises core emission parameter features and cross-item time sequence association features of each monitoring item; invoking a pre-constructed log report re-description logic model to carry out semantic reconstruction processing on the structured log feature group, and generating a log re-description logic framework, wherein the log re-description logic framework comprises a causal relationship chain and an important information distribution map of an emission event; Executing text generation operation based on the journal re-description logic framework, and generating an initial re-description journal report, wherein the initial re-description journal report comprises emission event description contents organized according to causal relationship chains; and carrying out two-dimensional verification processing on the initial repeated log report to generate a final repeated log report passing verification, wherein the two-dimensional verification processing comprises semantic consistency verification and structural integrity verification. In yet another aspect, an embodiment of the present invention further provides a log report replication system based on carbon emission monitoring, including a processor, and a machine-readable storage medium, where the machine-readable storage medium is connected to the processor, and the machine-readable storage medium is used to store a program, an instruction, or a code, and the processor is used to execute the program, the instruction, or the code in the machine-readable storage medium, so as to implement the method described above. Based on the above aspects, the embodiment of the invention obtains the original carbon emission monitoring log data comprising the time stamp sequence, and performs feature extraction on the original carbon emission monitoring log data to obtain the structured log feature group comprising the core emission parameter feature and the cross-item time sequence association feature, so that the effective integration and deep mining of the original log data are realized, the pre-constructed log report reiteration logic model is called to perform semantic reconstruction processing on the structured log feature group, and a log reiteration logic framework comprising a causal relationship chain of emission events and an important information distribution map is generated, so that the logic relationship and key information between the carbon emission events can be clearly presented. The initial repeated l