Method and apparatus for monitoring greenhouse gas emissions of an electronic device manufacturing enterprise

Abstract

The application relates to a greenhouse gas emission monitoring method and device for an electronic equipment manufacturing enterprise. The method comprises the following steps: obtaining equipment-level power parameters, fossil fuel combustion emission and purchased electric heat net emission of the electronic equipment manufacturing enterprise, and performing data preprocessing on the equipment-level power parameters to obtain a time-series power data set; determining a current typical production state mode of the electronic equipment manufacturing enterprise based on the time-series power data set, and extracting running features of key equipment in the electronic equipment manufacturing enterprise in the typical production state mode; inputting the running features into a pre-constructed knowledge graph to obtain a mapping result; and determining greenhouse gas emissions of the electronic equipment manufacturing enterprise according to the mapping result, the running features, the fossil fuel combustion emission and the purchased electric heat net emission. The method can improve the real-time performance of carbon emission monitoring in the electronic equipment manufacturing enterprise.

Description

Technical Field

[0001] This application relates to the field of power big data technology, and in particular to a method and apparatus for monitoring greenhouse gas emissions from an electronic equipment manufacturing enterprise. Background Technology

[0002] With the deepening of the green and low-carbon development concept in the industrial sector, greenhouse gas emissions from electronic equipment manufacturing companies, especially in high-end manufacturing industries such as semiconductors and chips, are receiving increasing attention. In addition to emissions from the purchase of electricity, heat, and the combustion of fossil fuels, electronic equipment manufacturing companies also generate non-carbon dioxide greenhouse gases such as perfluorinated carbon, nitrogen trifluoride, and sulfur hexafluoride during processes such as etching, chemical vapor deposition, and chamber cleaning. These gases have extremely high global warming potential, and their emissions are strongly correlated with production processes and are difficult to monitor, making them a key aspect of corporate carbon emission management.

[0003] Currently, electronic equipment manufacturers generally use periodic statistical accounting methods based on emission factors to monitor greenhouse gas emissions. This method uses an annual or quarterly cycle to make ex-post estimates of the company's overall greenhouse gas emissions based on parameters such as default emission factors and feedstock gas utilization rates.

[0004] However, this method is a post-hoc accounting approach, which cannot reflect changes in greenhouse gas emissions during the production process of electronic equipment manufacturers in a timely manner, nor can it accurately pinpoint specific production lines, equipment, or processes. Attempts to improve real-time performance and accuracy by adding dedicated greenhouse gas monitoring equipment face the problem of high deployment costs. Therefore, how to monitor greenhouse gas emissions from electronic equipment manufacturers without significantly increasing hardware costs has become a pressing technical problem in this field. Summary of the Invention

[0005] Therefore, it is necessary to provide a method and apparatus for monitoring greenhouse gas emissions from electronic equipment manufacturing enterprises, addressing the aforementioned technical problems.

[0006] In a first aspect, this application provides a method for monitoring greenhouse gas emissions from an electronic equipment manufacturing enterprise, comprising:

[0007] The equipment-level power parameters, fossil fuel combustion emissions, and net emissions from purchased electricity and heat of electronic equipment manufacturing enterprises are obtained, and the equipment-level power parameters are preprocessed to obtain a time-series power dataset.

[0008] Based on time-series power datasets, we determine the typical production status patterns of electronic equipment manufacturing enterprises and extract the operational characteristics of key equipment in these enterprises under these typical production status patterns.

[0009] The operational features are input into a pre-built knowledge graph to obtain mapping results, which include the production process of key equipment and the greenhouse gas emission factors corresponding to the production process.

[0010] Based on the mapping results, operating characteristics, fossil fuel combustion emissions, and net emissions from purchased electricity and heat, the greenhouse gas emissions of electronic equipment manufacturing enterprises are determined.

[0011] In one embodiment, the running features are input into a pre-built knowledge graph to obtain a mapping result, including:

[0012] Construct a knowledge graph based on prior knowledge in the field of electronic equipment manufacturing;

[0013] The operational features are input into the knowledge graph to determine the matching relationship between the operational features and entities in the knowledge graph;

[0014] The mapping result is determined based on rule-based reasoning and matching relationships.

[0015] In one embodiment, the greenhouse gas emission monitoring method of the aforementioned electronic equipment manufacturing enterprise further includes:

[0016] The similarity between the operational features and the standard process patterns in the knowledge graph is calculated in multiple dimensions to obtain the similarity values ​​for each dimension.

[0017] Weighted fusion of the various similarity values ​​is performed to determine the global similarity.

[0018] Based on the global similarity, the confidence level of the mapping result is obtained. When the confidence level meets the preset conditions, the mapping result is accepted.

[0019] In one embodiment, the greenhouse gas emissions of an electronic equipment manufacturing company are determined based on mapping results, operational characteristics, fossil fuel combustion emissions, and net emissions from purchased electricity and heat, including:

[0020] Based on the mapping results, the production process of key equipment is determined, including continuous operation processes and batch processes;

[0021] Determine process emissions based on production processes, operational characteristics, and greenhouse gas emission factors;

[0022] Greenhouse gas emissions are determined based on process emissions, fossil fuel combustion emissions, and net emissions from purchased electricity and heat.

[0023] In one embodiment, process emissions are determined based on production processes, operational characteristics, and greenhouse gas emission factors, including:

[0024] When the production process is a continuous operation process, the process emissions are obtained based on the process equivalent power and operating time in the operating characteristics and the greenhouse gas emission factor;

[0025] When the production process is a batch process, the process emissions are obtained based on the number of events and the emissions per event in the operating characteristics.

[0026] In one embodiment, based on a time-series power dataset, the typical current production status pattern of an electronic equipment manufacturing enterprise is determined, including:

[0027] Based on the time-series power dataset, the active power sequence in the time-series power dataset is segmented to obtain multiple load curve segments;

[0028] Based on each load curve segment, the clustering characteristics of each load curve segment are determined;

[0029] Clustering of each load curve segment is performed based on the dynamic time warp distance, and the optimal number of clusters is determined by the silhouette coefficient method to obtain the clustering results;

[0030] Based on the clustering results and clustering characteristics, the typical production status pattern of electronic equipment manufacturing enterprises is obtained.

[0031] In one embodiment, the operational characteristics of key equipment in an electronic equipment manufacturing enterprise are extracted, including:

[0032] Based on the time-series power dataset, determine the total active power sequence of key equipment;

[0033] Wavelet packet decomposition is used to decompose the total active power sequence, and the decomposed sub-band signals are filtered and reconstructed to obtain the reconstructed signal;

[0034] Event detection is performed on the reconstructed signals to determine the operating characteristics of key equipment.

[0035] In one embodiment, device-level power parameters are preprocessed to obtain a time-series power dataset, including:

[0036] Based on the density threshold elimination method, abnormal data in the equipment-level power parameters are eliminated to obtain the power parameters after eliminating abnormal data.

[0037] The power parameters after removing outlier data are standardized to obtain a time-series power dataset.

[0038] In one embodiment, the greenhouse gas emission monitoring method of the aforementioned electronic equipment manufacturing enterprise further includes:

[0039] The time-series power dataset is sorted according to the time series to determine the time-series data matrix;

[0040] The comprehensive power load sequence is determined based on each time point in the time series data matrix;

[0041] Based on the first-order autocorrelation coefficient in the comprehensive power load series, the time-series correlation degree is determined, which characterizes the smoothness of the comprehensive power load series.

[0042] The power parameters of each data point in the time-series power dataset are combined with the time-series correlation to obtain an enhanced feature vector;

[0043] Based on the density clustering algorithm, outlier data in the enhanced feature vectors are removed.

[0044] Secondly, this application also provides a greenhouse gas emission monitoring device for electronic equipment manufacturing enterprises, comprising:

[0045] The acquisition module is used to acquire the equipment-level power parameters, fossil fuel combustion emissions, and net emissions from purchased electricity and heat from electronic equipment manufacturing enterprises, and to preprocess the equipment-level power parameters to obtain a time-series power dataset.

[0046] The extraction module is used to determine the current typical production status mode of electronic equipment manufacturing enterprises based on time-series power datasets, and extract the operating characteristics of key equipment in electronic equipment manufacturing enterprises under the typical production status mode.

[0047] The input module is used to input the operational features into a pre-built knowledge graph to obtain the mapping results, which include the production process of key equipment and the greenhouse gas emission factors corresponding to the production process.

[0048] The determination module is used to determine the greenhouse gas emissions of electronic equipment manufacturing enterprises based on mapping results, operating characteristics, fossil fuel combustion emissions, and net emissions from purchased electricity and heat.

[0049] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0050] The equipment-level power parameters, fossil fuel combustion emissions, and net emissions from purchased electricity and heat of electronic equipment manufacturing enterprises are obtained, and the equipment-level power parameters are preprocessed to obtain a time-series power dataset.

[0051] Based on time-series power datasets, we determine the typical production status patterns of electronic equipment manufacturing enterprises and extract the operational characteristics of key equipment in these enterprises under these typical production status patterns.

[0052] The operational features are input into a pre-built knowledge graph to obtain mapping results, which include the production process of key equipment and the greenhouse gas emission factors corresponding to the production process.

[0053] Based on the mapping results, operating characteristics, fossil fuel combustion emissions, and net emissions from purchased electricity and heat, the greenhouse gas emissions of electronic equipment manufacturing enterprises are determined.

[0054] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0055] The equipment-level power parameters, fossil fuel combustion emissions, and net emissions from purchased electricity and heat of electronic equipment manufacturing enterprises are obtained, and the equipment-level power parameters are preprocessed to obtain a time-series power dataset.

[0056] Based on time-series power datasets, we determine the typical production status patterns of electronic equipment manufacturing enterprises and extract the operational characteristics of key equipment in these enterprises under these typical production status patterns.

[0057] The operational features are input into a pre-built knowledge graph to obtain mapping results, which include the production process of key equipment and the greenhouse gas emission factors corresponding to the production process.

[0058] Based on the mapping results, operating characteristics, fossil fuel combustion emissions, and net emissions from purchased electricity and heat, the greenhouse gas emissions of electronic equipment manufacturing enterprises are determined.

[0059] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0060] The equipment-level power parameters, fossil fuel combustion emissions, and net emissions from purchased electricity and heat of electronic equipment manufacturing enterprises are obtained, and the equipment-level power parameters are preprocessed to obtain a time-series power dataset.

[0061] Based on time-series power datasets, we determine the typical production status patterns of electronic equipment manufacturing enterprises and extract the operational characteristics of key equipment in these enterprises under these typical production status patterns.

[0062] The operational features are input into a pre-built knowledge graph to obtain mapping results, which include the production process of key equipment and the greenhouse gas emission factors corresponding to the production process.

[0063] Based on the mapping results, operating characteristics, fossil fuel combustion emissions, and net emissions from purchased electricity and heat, the greenhouse gas emissions of electronic equipment manufacturing enterprises are determined.

[0064] The aforementioned method and apparatus for monitoring greenhouse gas emissions from electronic equipment manufacturing enterprises acquire equipment-level power parameters, fossil fuel combustion emissions, and net emissions from purchased electricity and heat from these enterprises. The equipment-level power parameters are preprocessed to obtain a time-series power dataset. Based on this dataset, the typical production state mode of the electronic equipment manufacturing enterprise is determined, and the operational characteristics of key equipment within this mode are extracted. These operational characteristics are then input into a pre-constructed knowledge graph to obtain mapping results. These results include the production processes of the key equipment and the corresponding greenhouse gas emission factors. Based on the mapping results, operational characteristics, fossil fuel combustion emissions, and net emissions from purchased electricity and heat, the greenhouse gas emissions of the electronic equipment manufacturing enterprise are determined. This method eliminates the need for dedicated gas monitoring hardware, achieving accurate mapping and real-time calculation from equipment-level electricity consumption to process-level emissions. It effectively overcomes the technical problems of poor real-time performance and ambiguous positioning in traditional periodic statistical methods, significantly improving the real-time performance, traceability, and economic efficiency of carbon emission monitoring in electronic equipment manufacturing enterprises. Attached Figure Description

[0065] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0066] Figure 1 This is a diagram illustrating the application environment of a greenhouse gas emission monitoring method for an electronic equipment manufacturing company in one embodiment.

[0067] Figure 2 This is one of the flowcharts illustrating a method for monitoring greenhouse gas emissions from an electronic equipment manufacturing company in one embodiment;

[0068] Figure 3 This is a second schematic diagram of a greenhouse gas emission monitoring method for electronic equipment manufacturing companies in one embodiment;

[0069] Figure 4 This is the third flowchart illustrating a method for monitoring greenhouse gas emissions from an electronic equipment manufacturing company in one embodiment;

[0070] Figure 5 This is the fourth flowchart illustrating a method for monitoring greenhouse gas emissions from an electronic equipment manufacturing company in one embodiment;

[0071] Figure 6 This is the fifth flowchart illustrating a method for monitoring greenhouse gas emissions from an electronic equipment manufacturing company in one embodiment;

[0072] Figure 7 This is a schematic diagram of the greenhouse gas emission monitoring method for an electronic equipment manufacturing company in one embodiment;

[0073] Figure 8 This is the seventh flowchart illustrating a method for monitoring greenhouse gas emissions from an electronic equipment manufacturing company in one embodiment;

[0074] Figure 9 This is the eighth flowchart illustrating a method for monitoring greenhouse gas emissions from an electronic equipment manufacturing company in one embodiment;

[0075] Figure 10 This is the ninth flowchart illustrating a method for monitoring greenhouse gas emissions from an electronic equipment manufacturing company, as described in one embodiment.

[0076] Figure 11 This is a structural block diagram of a greenhouse gas emission monitoring device for an electronic equipment manufacturing company, as shown in one embodiment. Detailed Implementation

[0077] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0078] It should be noted that the terms "first," "second," etc., used in this application can be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from the second element. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more. The term "and / or" used in this application refers to one of the embodiments, or any combination of multiple embodiments.

[0079] With the deepening of the green and low-carbon development concept in the industrial sector, greenhouse gas emissions from electronic equipment manufacturing companies, especially in high-end manufacturing industries such as semiconductors and chips, are receiving increasing attention. In addition to emissions from the purchase of electricity, heat, and the combustion of fossil fuels, electronic equipment manufacturing companies also generate non-carbon dioxide greenhouse gases such as perfluorinated carbon, nitrogen trifluoride, and sulfur hexafluoride during processes such as etching, chemical vapor deposition, and chamber cleaning. These gases have extremely high global warming potential, and their emissions are strongly correlated with production processes and are difficult to monitor, making them a key aspect of corporate carbon emission management.

[0080] Currently, electronic equipment manufacturers generally use periodic statistical accounting methods based on emission factors to monitor greenhouse gas emissions. This method uses an annual or quarterly cycle to make ex-post estimates of the company's overall greenhouse gas emissions based on parameters such as default emission factors and feedstock gas utilization rates.

[0081] However, this method is a post-hoc accounting approach, which cannot reflect changes in greenhouse gas emissions during the production process of electronic equipment manufacturers in a timely manner, nor can it accurately pinpoint specific production lines, equipment, or processes. Attempts to improve real-time performance and accuracy by adding dedicated greenhouse gas monitoring equipment face the problem of high deployment costs. Therefore, how to monitor greenhouse gas emissions from electronic equipment manufacturers without significantly increasing hardware costs has become a pressing technical problem in this field.

[0082] In view of the above-mentioned technical problems, this application provides a method for monitoring greenhouse gas emissions of electronic equipment manufacturing enterprises. The following embodiments will specifically illustrate the method for monitoring greenhouse gas emissions of electronic equipment manufacturing enterprises.

[0083] The greenhouse gas emission monitoring method for electronic equipment manufacturing enterprises provided in this application embodiment can be applied to, for example... Figure 1 The computer device shown can be a terminal, and its internal structure diagram can be as follows. Figure 1 As shown, the computer device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When the computer program is executed by the processor, it implements a method for monitoring greenhouse gas emissions in an electronic equipment manufacturing enterprise. The display unit is used to form a visually visible image and can be a display screen, projection device, or virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0084] Those skilled in the art will understand that Figure 1 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0085] In one exemplary embodiment, such as Figure 2 As shown, a method for monitoring greenhouse gas emissions from an electronic equipment manufacturing enterprise is provided, which can be applied to... Figure 1 Taking computer devices as an example, the explanation includes:

[0086] S201: Obtain the equipment-level power parameters, fossil fuel combustion emissions, and net emissions from purchased electricity and heat from electronic equipment manufacturing enterprises, and perform data preprocessing on the equipment-level power parameters to obtain a time-series power dataset.

[0087] Equipment-level electrical parameters may include, but are not limited to: voltage, current, active power, reactive power, power factor, frequency, total harmonic distortion, timestamp, equipment identifier, and process step code; fossil fuel combustion emissions include natural gas consumption. Diesel consumption Other fuel consumption Carbon dioxide emissions from fuel consumption; net emissions from purchased electricity and heat include indirect emissions from purchased electricity and heat, minus emissions from electricity and heat exported, specifically including active power purchased from the external grid. Active power output to the external power grid Mass flow rate of purchased steam and the mass flow rate of output steam .

[0088] In this embodiment, a computer device uniformly collects multi-source heterogeneous power data (i.e., device-level power parameters) from smart meters, equipment sensors, production planning systems, etc., to construct a power data space. The multi-source heterogeneous power data is collected every 15 minutes, and each data point represents the power parameters of a single device at a specific time point, thereby forming a dataset. , where each data point It is a multi-dimensional feature vector containing the following features: voltage (V), current (A), active power (kW), reactive power (kVar), power factor (dimensionless), frequency (Hz), total harmonic distortion (%), timestamp (ISO 8601 format), equipment identifier (e.g., equipment ID), and process step code (e.g., etching, CVD). In addition, data points... By associating device identifiers with specific production equipment of electronic equipment manufacturers, the system ensures that the collected multi-source heterogeneous power data can be traced back to a specific stage in the semiconductor manufacturing process. Subsequently, the computer equipment also obtains real-time emissions from fossil fuel combustion and net emissions from purchased electricity and heat by connecting to the enterprise's energy management system. Fossil fuel combustion emissions include carbon dioxide emissions from the consumption of fuels such as natural gas and diesel; net emissions from purchased electricity and heat include indirect emissions from purchased electricity and heat, minus the emissions corresponding to electricity and heat output. Finally, the computer equipment performs data preprocessing on the device-level power parameters to remove outliers from the device-level power parameters (i.e., multi-source heterogeneous power data), resulting in a time-series power dataset.

[0089] S202, based on time-series power datasets, determines the current typical production state mode of electronic equipment manufacturing enterprises, and extracts the operating characteristics of key equipment in electronic equipment manufacturing enterprises under the typical production state mode.

[0090] In this embodiment, the computer device first segments the active power sequence in the time-series power dataset to identify the typical production state mode currently in which the electronic equipment manufacturing enterprise is located. Typical production state modes include, but are not limited to, continuous high-production mode, intermittent production mode, and maintenance low-production mode. Under different production state modes, the start-stop frequency, load stability, and energy consumption distribution of key equipment also differ significantly, and their contributions to greenhouse gas emissions also vary. Based on this, the computer device uses a time-frequency decomposition method to analyze the total load signal for the identified typical production state modes, separating the operating components corresponding to key equipment. Key equipment includes major process equipment such as etching machines and chemical vapor deposition equipment. Subsequently, the computer device further identifies the start-up and stop times of key equipment through an event detection algorithm, and extracts operating characteristics such as runtime, process equivalent power, and start-stop frequency. Among these, process equivalent power reflects the typical energy consumption level of the equipment when performing a specific process, while the start-stop frequency indirectly reflects the intensity of activities such as process switching and chamber cleaning.

[0091] S203, input the running features into the pre-built knowledge graph to obtain the mapping results.

[0092] The mapping results include the production process of key equipment and the corresponding greenhouse gas emission factors.

[0093] In this embodiment, the computer device first constructs a knowledge graph containing equipment classes, process classes, emission source classes, and their relationships based on prior knowledge in the field of electronic equipment manufacturing. This knowledge graph pre-stores the correspondence between different types of production equipment and production processes, as well as the mapping relationship between each production process and greenhouse gas emission factors. During actual operation, the computer device inputs the operational characteristics of key equipment in the electronic equipment manufacturing enterprise into the pre-constructed knowledge graph. Through entity matching and rule reasoning, it determines the type of production process currently being executed by the key equipment and finds the corresponding greenhouse gas emission factor, ultimately obtaining the mapping result. Production processes include continuous operation processes (such as etching and chemical vapor deposition) and batch processes (such as chamber cleaning), and the corresponding greenhouse gases include non-carbon dioxide greenhouse gases such as perfluorocarbons, nitrogen trifluoride, and sulfur hexafluoride.

[0094] S204. Based on the mapping results, operating characteristics, fossil fuel combustion emissions, and net emissions from purchased electricity and heat, determine the greenhouse gas emissions of electronic equipment manufacturing enterprises.

[0095] In this embodiment, the computer equipment first determines the type of production process currently being executed by the key equipment based on the mapping results. Different types of processes employ different emission calculation methods. For example, for continuous operation processes, the computer equipment can calculate the process emissions based on the process equivalent power, runtime, and greenhouse gas emission factor in the operating characteristics; for batch processes, the process emissions can be calculated based on the number of events and emissions per event in the operating characteristics. Subsequently, the computer equipment adds the process emissions, fossil fuel combustion emissions, and net emissions from purchased electricity and heat, and the total greenhouse gas emissions of the electronic equipment manufacturing enterprise are obtained, thereby achieving real-time monitoring and dynamic accounting of the enterprise's carbon emissions.

[0096] The aforementioned method for monitoring greenhouse gas emissions from electronic equipment manufacturing enterprises involves acquiring equipment-level power parameters, fossil fuel combustion emissions, and net emissions from purchased electricity and heat. The equipment-level power parameters are preprocessed to obtain a time-series power dataset. Based on this dataset, the typical production status of the electronic equipment manufacturing enterprise is determined, and the operational characteristics of key equipment within this model are extracted. These operational characteristics are then input into a pre-constructed knowledge graph to obtain mapping results. These results include the production processes of the key equipment and the corresponding greenhouse gas emission factors. Based on the mapping results, operational characteristics, fossil fuel combustion emissions, and net emissions from purchased electricity and heat, the greenhouse gas emissions of the electronic equipment manufacturing enterprise are determined. This method eliminates the need for dedicated gas monitoring hardware, enabling precise mapping and real-time calculation from equipment-level electricity consumption to process-level emissions. It effectively overcomes the technical problems of poor real-time performance and ambiguous positioning in traditional periodic statistical methods, significantly improving the real-time performance, traceability, and economic efficiency of carbon emission monitoring in electronic equipment manufacturing enterprises.

[0097] In an exemplary embodiment, the phrase "inputting the running features into a pre-built knowledge graph to obtain mapping results" in S203 above, such as... Figure 3 As shown, it includes:

[0098] S301 constructs a knowledge graph based on prior knowledge in the field of electronic equipment manufacturing.

[0099] In this embodiment, the computer device first defines the core class system of the entire manufacturing process based on prior knowledge in the field of electronic device manufacturing. This system includes a five-dimensional entity structure: Device, Process, Emission, Behavior, and Parameter. The Device class encompasses major process equipment such as Etcher, CVD Equipment, Ion Implanter, and Vacuum Pump. The Process class includes typical manufacturing processes such as Dry Etching, Chemical Vapor Deposition (CVD), Ion Implantation, and Chamber Cleaning. The Emission class includes perfluorocarbons (PFCs), hydrofluorocarbons (HFCs), and nitrogen trifluoride (NFCs). Sulfur hexafluoride () Non-carbon dioxide greenhouse gases, such as , are included; electrical behavior includes start-stop events, operation duration, and power consumption; and characteristic parameters include process-related parameters such as temperature, pressure, gas flow, and cycle time.

[0100] Building upon this foundation, the computer equipment further defines the semantic relationships between the aforementioned five-dimensional entity structures to construct the relational framework of the knowledge graph. Specifically, this includes: the relationship between equipment and process (has_process), such as the relationship between an etching machine and dry etching (etching machine has_process dry etching); and the relationship between process and emission source (has_emission), such as the relationship between dry etching and emission sources. (dry etching has_emission) The relationships between equipment and electrical behavior (has_behavior), such as etching machine associated with start-up and shutdown events (etching machine has_behavior start-up and shutdown events); the relationships between processes and characteristic parameters (has_parameter), such as chemical vapor deposition associated with chamber temperature (chemical vapor deposition has_parameter chamber temperature); and the relationships between processes (next_process), such as dry etching associated with chemical vapor deposition as a downstream process (dry etching next_process chemical vapor deposition). Meanwhile, corresponding data attributes are defined for various entities. Specifically, the attributes for the Device class include rated power, model, and installation date; the attributes for the Process class include process duration, temperature range, and gas type; the attributes for the Behavior class include daily operation hours, start-stop frequency, and average power; and the attributes for the Emission class include emission factor, GWP value, and monitoring method.

[0101] Computer devices can generate triple instances of a knowledge graph by filling in specific equipment, processes, emission sources, and attribute values ​​into the above ontology structure. For example: (Etcher_001, has_process, Dry_Etching) indicates that a certain etching machine has a dry etching process; (Dry_Etching, has_emission, This indicates that the dry etching process emits emissions. (Etcher_001, rated_power, 150kW) indicates that the rated power of the etching machine is 150kW; (gwp_value, 6630) indicates The global warming potential is 6630. At this point, the computer equipment has completed the offline construction of the knowledge graph, providing foundational data support for subsequent real-time semantic mapping.

[0102] S302, input the running features into the knowledge graph and determine the matching relationship between the running features and entities in the knowledge graph.

[0103] In this embodiment, the computer device first converts the operational features into standardized data vectors that can be processed by the knowledge graph in real time. This standardized data vector Fields include device type, operation duration, start-stop frequency, average power, timestamp, and load pattern. For example, the device type is an etching machine, the operation duration is 75 minutes, and the start-stop frequency is 2 times / day.

[0104] Subsequently, the computer equipment uses this standardized data vector as query conditions to perform entity matching within the pre-constructed knowledge graph. Specifically, based on the device type field (device_type) in the operational characteristics, the computer equipment locates the corresponding specific device instance in the device class of the knowledge graph; based on power consumption behavior characteristics such as runtime and start / stop frequency, it matches the corresponding behavior pattern in the power consumption behavior class; and based on timestamps and load patterns, it helps determine the current production stage and process context.

[0105] S303, based on rule-based reasoning and matching relationships, determines the mapping result.

[0106] In this embodiment, after receiving the matching relationship, the inference engine of the computer device can activate a predefined semantic rule base to perform automatic logical reasoning. This semantic rule base defines the mapping logic from electricity consumption behavior to processes and emissions in the form of logical expressions.

[0107] For example, the etching process identification rule is when the equipment type is an etching machine and the single run time exceeds 60 minutes (Device(type=Etcher)). OperationDuration If 60), then the equipment is currently performing a dry etching process (hasProcess(Dry_Etching)); the association rule for CVD chamber cleaning and venting is that when the equipment type is CVD equipment and the daily start-up and shutdown frequency exceeds 3 times (indicating frequent chamber preparation) (Device(type=CVD_Equipment)). StartStopFrequency 3) Then it is determined that the equipment has performed a chamber cleaning process, and associated with... Gas emissions (hasProcess(Chamber_Cleaning)) hasEmission The production process route inference rule is that if a chemical vapor deposition (CVD) process is started on the same production line within 30 minutes after the dry etching process is completed (hasProcess( Dry_Etching ) hasProcess( CVD ) ( - If <3, then the downstream process of dry etching is determined to be chemical vapor deposition (CVD), thus constructing process route knowledge.

[0108] By combining rule-based reasoning with similarity matching, the computer device automatically maps isolated electricity consumption behavior tags to specific production processes and corresponding greenhouse gas emission sources, ultimately outputting the mapping results. The specific details of these mapping results are shown in Table 1.

[0109] Table 1

[0110]

[0111] In an exemplary embodiment, the greenhouse gas emission monitoring method of the above-mentioned electronic equipment manufacturing enterprise, such as Figure 4 As shown, it also includes:

[0112] S401 performs multi-dimensional similarity calculations between the operational features and the standard process patterns in the knowledge graph to obtain similarity values ​​for each dimension.

[0113] In this embodiment, to ensure the accuracy of the mapping, the computer device introduces multi-dimensional similarity matching as an auxiliary decision-making mechanism based on rule-based reasoning. Specifically, the computer device will use real-time data vectors... With the standard process mode in the knowledge base Multi-dimensional comparisons are performed to calculate the similarity values ​​for each dimension. The multi-dimensional similarity includes equipment type similarity, runtime similarity, and start / stop frequency similarity. Specifically, equipment type similarity is a symbolic attribute; the similarity is 1 when the equipment type in the real-time data is the same as the equipment type in the standard process mode, and 0 otherwise. Runtime similarity is a numerical attribute, and runtime similarity can be represented by equation (1), as shown below:

[0114] (1);

[0115] in, Indicates the real-time runtime. The formula (1) represents the typical runtime in the standard process mode. The similarity is calculated based on the relative difference between the real-time runtime and the standard runtime. The smaller the difference, the closer the similarity is to 1; the larger the difference, the closer the similarity is to 0.

[0116] For the similarity of start and stop frequencies, it is a numerical attribute. The similarity of start and stop frequencies can be represented by the relation (2), which is shown below:

[0117] (2);

[0118] in, This indicates the start / stop frequency in real-time data; This indicates the typical start-stop frequency in a standard process mode. It represents the standard deviation of the start-stop frequency distribution; the greater the deviation between the real-time frequency and the standard frequency, the faster the similarity decays.

[0119] S402, weighted fusion of each similarity value to determine the global similarity.

[0120] In this embodiment, the similarity values ​​of the computer devices across different dimensions are weighted and fused. Specifically, the computer devices are assigned corresponding weight coefficients for the similarity of each dimension. Each weight can be set according to the physical importance of that dimension (for example, the weight of power features can be greater than that of voltage features), or it can be determined through training with historical data. Meanwhile, the global similarity can be represented by relation (3), which is shown below:

[0121] (3);

[0122] in, Represents a real-time data vector; This represents the standard process patterns in the knowledge base. Through weighted summation, computer devices can fuse multi-dimensional comparison results into a comprehensive global similarity score for subsequent confidence assessment.

[0123] S403: Based on the global similarity, obtain the confidence level of the mapping result. When the confidence level meets the preset conditions, accept the mapping result.

[0124] In this embodiment, the computer device further calculates the confidence level of the mapping result based on the global similarity. The confidence level is used to measure the reliability of the current matching result relative to other candidate process modes, and it can be represented by relation (4), which is shown below:

[0125] (4);

[0126] in, Indicates the current matching process mode; This indicates other candidate process modes in the knowledge base; This represents the total number of candidate process modes.

[0127] The computer device compares the calculated confidence level with a preset threshold (which can be adjusted according to the actual data distribution). For example, if the preset threshold is 0.71, when the confidence level is greater than or equal to the preset threshold, it indicates that the current mapping result has high reliability, and the computer device accepts the mapping result; when the confidence level is lower than the preset threshold, it indicates that the reliability of the mapping result is insufficient, and the computer device triggers manual review or marks the result as pending confirmation.

[0128] In an exemplary embodiment, the phrase "determine the greenhouse gas emissions of the electronic equipment manufacturing enterprise based on the mapping results, operating characteristics, fossil fuel combustion emissions, and net emissions from purchased electricity and heat" in S204 above, such as... Figure 5 As shown, it includes:

[0129] S501, Based on the mapping results, determine the production process of key equipment.

[0130] The production process includes continuous operation processes and batch processes.

[0131] In this embodiment, the computer device first extracts key information from the mapping results. Based on the recommended process (Dry_Etching) field in the mapping results, it determines the type of production process currently being executed by the key equipment (including two main categories: continuous operation processes and batch processes). Continuous operation processes include, but are not limited to, dry etching and chemical vapor deposition; these processes, once started, run continuously for a relatively long time, exhibiting stable and continuous energy consumption and emissions. Batch processes include, but are not limited to, chamber cleaning; these processes occur as discrete events, each with a short duration but a high frequency, resulting in intermittent fluctuations in emissions.

[0132] It is worth noting that computer equipment can also extract process activity data directly related to emissions in real time based on the mapping results of equipment-level electricity consumption behavior tags and knowledge graphs. Specifically, this includes equipment runtime. Number of key process events and process equivalent power .

[0133] S502 determines process emissions based on production processes, operational characteristics, and greenhouse gas emission factors.

[0134] In this embodiment, the computer device first pre-calculates the process-emission source correlation stored in the knowledge graph and the default parameters of the national standard for each process. Calculate its process-level carbon emission factor The carbon emission factor of this process-grade electricity It can be represented by relation (5), which is shown below:

[0135] (5);

[0136] in, The initial value is calculated based on the standard process recipe. When the computer equipment monitors the actual power of the equipment... Continuously deviating from its process equivalent power At that time, the emission factor can be dynamically fine-tuned; The global warming potential of the main emissions gases from this process is represented by the default values ​​in Table C.3 of Appendix C of GB / T 32151.24-2024. This indicates the utilization rate of the raw material gas corresponding to the process; and Indicates the collection and removal efficiency of the process waste gas treatment device; The standard gas flow rate (kg / h) is provided by the knowledge graph from the characteristic parameter class, representing the gas consumption rate of the process under the standard formulation. Indicates the process equivalent power This parameter distributes the emissions from gas consumption to each kilowatt-hour of electricity, establishing a direct coupling relationship between electricity and carbon emissions.

[0137] Subsequently, the computer equipment selects the corresponding calculation model based on the type of production process to determine the process emissions.

[0138] For continuous operation processes (such as dry etching and chemical vapor deposition), the computer equipment employs an emission calculation model based on runtime. Specifically, the computer equipment acquires the process equivalent power based on operational characteristics. (Unit: kilowatts) and operating time (Unit: hour), based on the process-level electricity carbon emission factor in the mapping results. (Unit is) ), then the equipment In process Below CO2 equivalent emissions generated at any time It can be represented by relation (6), which is shown below:

[0139] (6);

[0140] For batch-type processes (such as chamber cleaning), the computer equipment uses an event-based emission calculation model. First, based on the standard gas flow rate... Standard event duration Raw material gas utilization rate Waste gas collection efficiency Waste gas removal efficiency And the global warming potential of the greenhouse gases corresponding to this technology. Determine the emissions of a single process event. The emissions from this single process event It can be represented by relation (7), which is shown below:

[0141] = × ×(1- )×(1- × )× (7);

[0142] Subsequently, the computer device acquires the number of events in the operational characteristics. (Unit: times), then the process exist Total emissions generated during the period It can be represented by relation (8), which is shown below:

[0143] = × (8);

[0144] Finally, the computer equipment calculates the process emissions, which can be represented by equation (9), as shown below:

[0145] (9);

[0146] in, Indicates process emissions; This indicates the time interval between data acquisition and computation.

[0147] S503 determines greenhouse gas emissions based on process emissions, fossil fuel combustion emissions, and net emissions from purchased electricity and heat.

[0148] In this embodiment of the application, the computer device first calculates the emissions from the combustion of fossil fuels. For the first... This type of fuel, based on its real-time consumption. Average lower heating value Carbon content per unit of calorific value Carbon oxidation rate and the conversion coefficient from carbon to carbon dioxide Calculate the fuel in Mass of carbon dioxide emissions from constant combustion (The unit is kilogram), which can be represented by relation (10), as shown below:

[0149] (10);

[0150] Subsequently, the computer equipment sums up the emissions of all fuels to obtain the total emissions from fuel combustion, and then calculates the instantaneous emission rate of fuel combustion. (Unit is) ), this instantaneous emission rate It can be represented by relation (11), which is shown below:

[0151] (11);

[0152] Secondly, the computer equipment calculates the net indirect emissions from purchased and exported electricity and heat. The instantaneous rate of indirect emissions from purchased electricity is also calculated. For purchased power Multiply by grid emission factor Multiply by the time interval Instantaneous rate of indirect emissions from purchased heat To purchase steam mass flow rate Multiply by the vapor enthalpy difference Multiply by the steam emission factor Meanwhile, the deduction amounts for electricity and heat exported are calculated separately. and = + × × The net indirect emission instantaneous rate is: = + - - .

[0153] Based on this, the computer equipment summarizes the three types of emissions to obtain the total emissions for the entire plant. Comprehensive real-time emission intensity at any given moment (Unit is) ),Right now = + + .

[0154] Finally, computer equipment for comprehensive real-time emission intensity Integrating the results, we obtain the initial report data. up to the current moment greenhouse gas emissions The greenhouse gas emissions It can be represented by relation (12), which is shown below:

[0155] (12);

[0156] At this point, the computer equipment has completed the real-time monitoring and dynamic calculation of greenhouse gas emissions from electronic equipment manufacturing enterprises, realizing a complete data closed loop from equipment-level power parameters to the total emissions of the entire plant.

[0157] In an exemplary embodiment, the "determining the current typical production state mode of electronic equipment manufacturing enterprises based on time-series power datasets" in S202 above, such as... Figure 6 As shown, it includes:

[0158] S601, based on a time-series power dataset, segments the active power sequence in the time-series power dataset to obtain multiple load curve segments.

[0159] In this embodiment, the computer device first extracts active power data from the time-series power dataset and arranges it in chronological order to form a total load time-series data sequence. ,in Indicates at a point in time The total active power of the entire plant, in kilowatts. This represents the total number of time points.

[0160] To capture the production status characteristics at different times, the computer equipment uses a sliding window to segment the above total load time-series data sequence. The window length is set to... (For example, 24 hours corresponds to 96 data collection points at 15-minute intervals), the sliding step size is... (For example, 1 hour corresponds to 4 points). The computer equipment divides the total load time-series data sequence into segments using a sliding window. bar sequence Each subsequence represents a segment of the load curve within a specific time period. These segments retain the dynamic characteristics of the time series while independently characterizing the production load patterns within their corresponding time periods.

[0161] S602, based on each load curve segment, determine the clustering characteristics of each load curve segment.

[0162] In this embodiment, the computer device extracts two core production status features from each load curve segment: the average daily load rate. and load fluctuation intensity Among them, the average daily load rate The ratio of the average load of the subsequence to the maximum load of the subsequence; load fluctuation intensity. This is the ratio of the standard deviation of the subsequence to the mean of the subsequence. Average daily load rate. This reflects the equipment's load level during that period. The closer to 1, the more fully the equipment is utilized; load fluctuation intensity It reflects the stability of the load. The smaller the value, the more stable the load. The larger the value, the more frequent the starts and stops, and the more drastic the fluctuations. Through the above calculations, the computer equipment represents each load curve segment as a two-dimensional clustering feature vector composed of the average daily load rate and the intensity of load fluctuations, and determines the clustering characteristics of each load curve segment.

[0163] S603 clusters each load curve segment based on the dynamic time warp distance and uses the silhouette coefficient method to determine the optimal number of clusters to obtain the clustering results.

[0164] In this embodiment, the computer device uses the K-Means algorithm based on dynamic time warping distance to cluster each load curve segment. For any two subsequences and ,in, and Let X and Y represent the subsequences X and Y respectively at the 1st, 2nd, and 3rd. and the Active power values ​​at individual time points (unit: kW); and The lengths of subsequences X and Y are respectively (in normal cases) (i.e., window length). The calculation steps for the dynamic time-warped distance are as follows: First, the computer device constructs a distance matrix. ,in , Secondly, calculate the cumulative matrix C, whose recurrence relation is as follows: The initial conditions are: Finally, the dynamic time warping distance is The smaller this value, the more similar the shapes of the two load curves are. By using the dynamic time warping distance, computer equipment can effectively identify load curve segments that are similar in shape but misaligned on the time axis.

[0165] During the clustering process, the computer equipment automatically determines the optimal number of clusters using the silhouette coefficient method. For each candidate Values ​​are used to calculate the global profile coefficient. The global contour coefficient It can be represented by relation (13), which is shown below:

[0166] (13);

[0167] in, Indicates sample The average distance to other samples in the same cluster; Indicates sample The minimum average distance to other cluster samples.

[0168] Therefore, the computer equipment selection makes the global contour coefficient The largest The value is used as the optimal clustering number to ensure the rationality and stability of the clustering results.

[0169] S604, based on the clustering results and clustering characteristics, obtains the current typical production status pattern of electronic equipment manufacturing enterprises.

[0170] In this embodiment of the application, the computer device can obtain the clustering results and clustering characteristics. There are three clusters and their corresponding cluster centers, each representing a typical production state pattern. Simultaneously, based on the clustering characteristics of each cluster (average daily load rate)... and load fluctuation intensity Each cluster of computer equipment is assigned a business label with a clear carbon emission semantic. Specifically, for clusters with high load rates and small fluctuations, the label indicates that the equipment is running at full load continuously, corresponding to a continuous production state with high carbon emission intensity; for clusters with loads exhibiting regular periodic fluctuations, the label indicates that the equipment starts and stops at fixed rhythms, corresponding to an intermittent production state with periodic emissions; and for clusters with significantly low average loads, the label indicates that the equipment is in a shutdown or low-load operation state, corresponding to a maintenance or low-production state with low carbon emissions.

[0171] In an exemplary embodiment, the "extracting operational characteristics of key equipment in an electronic equipment manufacturing enterprise" in S202 above, such as Figure 7 As shown, it includes:

[0172] S701, based on the time-series power dataset, determine the total active power sequence of key equipment.

[0173] In this embodiment, the computer device first selects a set of all daily load curves under a typical production state mode (such as a continuous high-production mode), and aligns them to the same starting time point to form a standardized total load signal dataset for in-depth analysis. By locking onto specific production state patterns, computer equipment can effectively eliminate interference from different production states on signal characteristics, making subsequent analysis more targeted and accurate.

[0174] Based on this, the computer equipment extracts the active power data corresponding to key equipment from the time-series power dataset and arranges it in chronological order to form the total active power sequence of the equipment. This sequence reflects the pattern of power consumption of key equipment under specific production conditions over time.

[0175] S702 uses wavelet packet decomposition to decompose the total active power sequence, and then filters and reconstructs the sub-band signals after decomposition to obtain the reconstructed signal.

[0176] In this embodiment, the computer equipment employs wavelet packet decomposition to perform frequency band separation and feature reconstruction on the total active power sequence, in order to separate the operating components of key equipment from the complex power signal superimposed by multiple devices. Specifically, the computer equipment selects the db4 wavelet basis to analyze the total load signal representing a specific production mode. Perform J-level wavelet packet decomposition to obtain Individual with signal Among them, the first The frequency band range of each sub-band is , The sampling frequency is used. After decomposition, the computer equipment measures the energy of each sub-band signal. (in, Indicates the first The number of belts in the first Wavelet packet decomposition coefficients at each time point), selecting those with significant energy (e.g., wavelet packet decomposition coefficients at each time point). ,in, This represents the low-frequency sub-bands that indicate the threshold coefficients for energy screening. These low-frequency sub-bands primarily contain the operational characteristics of key equipment, while high-frequency noise is effectively filtered out. Computer equipment reconstructs the selected sub-bands to obtain the characteristic signal with high-frequency noise removed. That is, the reconstructed signal.

[0177] To further verify the quality of the characteristic signals and identify the operating patterns of the equipment, the computer equipment... Perform a Fast Fourier Transform to obtain its spectrum. By identifying characteristic frequency peaks in the spectrum. Calculate the operating cycle of the equipment =1 / If there are obvious peaks in the spectrum, it indicates that the characteristic signal has good periodicity and the reconstruction effect meets the requirements; if the peaks are not obvious, the computer equipment adjusts the wavelet packet decomposition parameters and performs decomposition and reconstruction again.

[0178] S703 performs event detection on the reconstructed signal to determine the operating characteristics of key equipment.

[0179] In this embodiment of the application, the computer device is based on the reconstructed signal. An event detection algorithm is used to identify the start-up and shutdown events of key equipment, and the operating characteristics of the key equipment are extracted accordingly.

[0180] Specifically, the computer device uses local statistics and slope threshold criteria based on a sliding window for event detection. When a power rising edge is detected to exceed a preset slope threshold, it is determined to be the device startup moment. When the detected power drop-off edge exceeds the preset slope threshold, it is determined as the moment the device stops. Based on the identified start and stop events, the computer equipment calculates the runtime, average power of events, and start and stop frequency in its operating characteristics.

[0181] In terms of runtime, it represents the duration of a single run, i.e. = - For the event-average power, which characterizes the average power of the device during operation, it can be represented by the relationship (14), as shown below:

[0182] (14);

[0183] in, This represents the average power baseline over a period of time (e.g., 30 minutes) prior to the event.

[0184] The start-stop frequency is represented by the number of times the device is started within a preset time period (such as one day), and the unit is times / day.

[0185] Through the above event detection and feature extraction, the computer equipment finally outputs structured operational feature data, such as: the average daily operating time (hours) of the etching machine, the start-up and shutdown frequency of the etching machine (times / day), the average daily operating time (hours) of the CVD equipment, and the start-up and shutdown frequency of the CVD equipment (times / day), etc.

[0186] In an exemplary embodiment, the "preprocessing of device-level power parameters to obtain a time-series power dataset" in S201 above, such as... Figure 8 As shown, it includes:

[0187] S801 uses a density threshold elimination method to remove abnormal data from equipment-level power parameters, resulting in power parameters after removing abnormal data.

[0188] In this embodiment, a density threshold method is used for data cleaning to eliminate invalid data introduced by equipment failures and transmission errors. For each data point... (in, ), the density of data feature points of computer equipment The density of feature points in this data is calculated. It can be represented by relation (15), which is shown below:

[0189] (15);

[0190] in, Representing data points Feature point density; Representing data points The 3D characteristic values ​​(such as voltage, current, etc.); This indicates the number of dimensions of the data features.

[0191] Subsequently, the computer equipment was set with a noise reduction threshold. And calculate the set of density values ​​for all data points. mean and standard deviation .set up - (The constant 2 can be adjusted according to the actual data distribution).

[0192] Based on the noise reduction threshold, the computer device divides the data space into high-density areas and low-density areas; that is, when... ≥ When a data point is located in a high-density area, it indicates that a large number of normal data points are clustered around it, and it is judged as normal data and retained; when < If a data point is located in a low-density area, it indicates that the point deviates from the normal data distribution and is judged as abnormal data, and is therefore removed.

[0193] S802 standardizes the power parameters after removing outlier data to obtain a time-series power dataset.

[0194] In this embodiment of the application, the computer device obtains the power parameters after removing abnormal data. Then, regarding this power parameter After standardization, it can be represented by relation (16), which is shown below:

[0195] (16);

[0196] in, Indicates the first The data point at the th th The normalized values ​​on each feature dimension; Indicates the first The data point at the th th The original feature values ​​after denoising each feature dimension; express The Middle The mean of all values ​​in each feature dimension; express The Middle The standard deviation of all values ​​in each feature dimension.

[0197] After this processing, all power parameters were standardized to the same dimension, with a mean of 0 and a standard deviation of 1, providing a standardized data foundation for subsequent time series analysis.

[0198] In an exemplary embodiment, the greenhouse gas emission monitoring method of the above-mentioned electronic equipment manufacturing enterprise, such as Figure 9 As shown, it also includes:

[0199] S901, sort the time-series power dataset according to the time series to determine the time-series data matrix.

[0200] In this embodiment of the application, the computer device arranges the standardized time-series power dataset according to the time sequence to construct a time-series data matrix. Let there be a total of There are 1 time points, and each time point has 1... Each electrical characteristic (such as voltage, current, active power, etc.) then It is The matrix can be represented by relation (17), which is shown below:

[0201] (17);

[0202] in, Indicates the first The first time point One electrical characteristic value.

[0203] S902, determine the comprehensive power load sequence based on each time point in the time series data matrix.

[0204] In this embodiment of the application, the computer device processes the time-series data matrix. Each row (i.e., each time point) in the data is subjected to feature aggregation, compressing the multidimensional power features into a single comprehensive power load value, thereby obtaining a comprehensive time series sequence for subsequent analysis. .

[0205] For the At each point in time, the computer equipment calculates the weighted Euclidean norm of all its characteristics to obtain the comprehensive power load value. Its calculation formula can be expressed by relation (18), which is shown below:

[0206] (18)

[0207] in, Indicates the first At the first time point The values ​​of each power characteristic dimension.

[0208] Through the above feature aggregation, the computer equipment reduces the high-dimensional power data to a one-dimensional comprehensive power load sequence. This sequence not only preserves the main information of the original data, but also facilitates subsequent time series analysis and feature extraction.

[0209] S903 determines the time-series correlation degree based on the first-order autocorrelation coefficient in the integrated power load sequence.

[0210] Among them, the temporal correlation degree characterizes the smoothness of the comprehensive power load sequence.

[0211] In this embodiment of the application, the computer equipment processes the integrated power load sequence. Time-series correlation analysis is performed, and the smoothness and inertia of the load series are quantified by calculating the first-order autocorrelation coefficient, thus obtaining the time-series correlation degree. The time series correlation It can be represented by relation (19), which is shown below:

[0212] (19);

[0213] in, Represents the comprehensive power load sequence The arithmetic mean of the time-series correlation. When the value approaches 1, it indicates a strong positive correlation between load values ​​at adjacent time points, a smooth load curve with high inertia, and a stable corresponding process state; when the time sequence correlation... When the load value approaches 0, it indicates that the load values ​​at adjacent time points are weakly correlated, the load curve fluctuates violently and is highly random, and the corresponding process state may be abnormal or frequently switching.

[0214] S904 combines the power parameters of each data point in the time-series power dataset with the time-series correlation degree to obtain an enhanced feature vector.

[0215] In this embodiment, the computer device combines the standardized power parameters of each data point in the time-series power dataset with the time-series correlation degree of its respective equipment to construct an enhanced feature vector, which serves as the input to the density clustering algorithm. For each data point... Its enhanced feature vector The expression is: ,in, For the first data points One original power characteristic. Among them, device-level time-series correlation. The calculation method is shown in relation (19).

[0216] S905, based on the density clustering algorithm, removes outlier data from the enhanced feature vector.

[0217] In this embodiment, the computer device uses the DBSCAN density clustering algorithm to perform cluster analysis on the enhanced feature vectors, identifying and removing outlier data points. First, the computer device adaptively determines the optimal parameters of the DBSCAN algorithm based on the elbow rule. Specifically, it first calculates the clustering of all data points in the enhanced feature space with their corresponding elbows. The distance between the nearest neighbors is called the distance between the nearest neighbors. - Distance; then, all points - Distances are sorted in descending order and plotted as a curve. This curve typically exhibits a distinct inflection point (i.e., the elbow), before which the curve drops sharply (corresponding to dense areas of normal data), and afterwards the change becomes gradual (corresponding to outlier or sparse data). The computer then assigns this inflection point to... - The distance value is selected as the optimal neighborhood radius parameter. Minimum number of points parameter Set directly The value (usually the number of data feature dimensions or a small empirical value).

[0218] Then, the computer device uses the optimal parameters determined above ( , ) Perform DBSCAN clustering to identify and remove all data points that cannot be assigned to any cluster (i.e., noise points) as outliers.

[0219] For outliers that have been removed, the computer equipment uses a differentiated strategy to impute the data based on their continuity in the time series, in order to ensure data quality and temporal consistency:

[0220] For locally missing out points (i.e., the number of consecutive points < threshold) Data imputation is performed based on relation (20), which is shown below:

[0221] (20);

[0222] in, This represents the value after data filling. This indicates the width of the selected normal data window before and after.

[0223] For consecutive missing out points (i.e., the number of consecutive points ≥ the threshold) Data imputation is performed based on relation (21), which is shown below:

[0224] (twenty one);

[0225] in, Indicates the starting time point of a continuous abnormal segment. Indicates the end time point of a continuous abnormal segment. This indicates the last normal value before the abnormal segment; This indicates the first normal value after the abnormal segment.

[0226] In summary, based on all the above embodiments, a method for monitoring greenhouse gas emissions from electronic equipment manufacturing enterprises is also provided, such as... Figure 10 As shown, the method includes:

[0227] S1001, Obtain equipment-level power parameters, fossil fuel combustion emissions, and net emissions from purchased electric heat from electronic equipment manufacturers;

[0228] S1002, execute S1003-S1004 to perform data preprocessing on device-level power parameters and obtain time-series power dataset;

[0229] S1003, based on the density threshold elimination method, abnormal data in the equipment-level power parameters are eliminated to obtain the power parameters after eliminating abnormal data;

[0230] S1004, standardize the power parameters after removing outlier data to obtain a time-series power dataset, and perform S1005-S1009 on the time-series power dataset to remove outliers.

[0231] S1005, Sort the time-series power dataset according to the time series to determine the time-series data matrix;

[0232] S1006, Determine the comprehensive power load sequence based on each time point in the time series data matrix;

[0233] S1007, Determine the time series correlation degree based on the first-order autocorrelation coefficient in the comprehensive power load sequence;

[0234] S1008 combines the power parameters of each data point in the time-series power dataset with the time-series correlation degree to obtain an enhanced feature vector;

[0235] S1009, based on the density clustering algorithm, outlier data in the enhanced feature vector is removed;

[0236] S1010, based on a time-series power dataset, segments the active power sequence in the time-series power dataset to obtain multiple load curve segments;

[0237] S1011, Based on each load curve segment, determine the clustering characteristics of each load curve segment;

[0238] S1012, clustering of each load curve segment based on dynamic time warping distance, and using the silhouette coefficient method to determine the optimal number of clusters to obtain the clustering results;

[0239] S1013. Based on the clustering results and clustering characteristics, the typical production status pattern of electronic equipment manufacturing enterprises is obtained.

[0240] S1014, Under typical production conditions, determine the total active power sequence of key equipment based on time-series power datasets;

[0241] S1015, wavelet packet decomposition is used to decompose the total active power sequence, and the decomposed sub-band signals are filtered and reconstructed to obtain the reconstructed signal;

[0242] S1016, perform event detection on the reconstructed signal to determine the operating characteristics of key equipment;

[0243] S1017, Construct a knowledge graph based on prior knowledge in the field of electronic equipment manufacturing;

[0244] S1018, Input the running features into the knowledge graph and determine the matching relationship between the running features and the entities in the knowledge graph;

[0245] S1019, Determine the mapping result based on rule-based reasoning and matching relationships;

[0246] S1020 calculates the similarity between the operational features and the standard process mode in the knowledge graph in multiple dimensions to obtain the similarity value of each dimension.

[0247] S1021, Weighted fusion of each similarity value is performed to determine the global similarity;

[0248] S1022, Based on the global similarity, obtain the confidence level of the mapping result. When the confidence level meets the preset conditions, accept the mapping result.

[0249] S1023, Based on the mapping results, determine the production process of key equipment;

[0250] S1024, When the production process is a continuous operation process, the process emissions are obtained based on the process equivalent power and operating time in the operating characteristics and the greenhouse gas emission factor; when the production process is a batch process, the process emissions are obtained based on the number of events and the emission amount per event in the operating characteristics.

[0251] S1025 determines greenhouse gas emissions based on process emissions, fossil fuel combustion emissions, and net emissions from purchased electricity and heat.

[0252] The methods described in each of the above steps have been described in the foregoing embodiments. For details, please refer to the foregoing descriptions. They will not be repeated here.

[0253] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.

[0254] Based on the same inventive concept, this application also provides a greenhouse gas emission monitoring device for an electronic equipment manufacturing enterprise to implement the greenhouse gas emission monitoring method for electronic equipment manufacturing enterprises as described above. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more embodiments of the greenhouse gas emission monitoring device for electronic equipment manufacturing enterprises provided below can be found in the limitations of the greenhouse gas emission monitoring method for electronic equipment manufacturing enterprises described above, and will not be repeated here.

[0255] In one exemplary embodiment, such as Figure 11 As shown, a greenhouse gas emission monitoring device for an electronic equipment manufacturing enterprise is provided, comprising:

[0256] The acquisition module 11 is used to acquire the equipment-level power parameters, fossil fuel combustion emissions and net emissions of purchased electricity and heat from electronic equipment manufacturing enterprises, and to preprocess the equipment-level power parameters to obtain a time-series power dataset.

[0257] Extraction module 12 is used to determine the current typical production state mode of electronic equipment manufacturing enterprises based on time-series power datasets, and extract the operating characteristics of key equipment in electronic equipment manufacturing enterprises under the typical production state mode.

[0258] Input module 13 is used to input the operational features into a pre-built knowledge graph to obtain mapping results, which include the production process of key equipment and the greenhouse gas emission factors corresponding to the production process.

[0259] Module 14 is used to determine the greenhouse gas emissions of electronic equipment manufacturing enterprises based on the mapping results, operating characteristics, fossil fuel combustion emissions, and net emissions from purchased electricity and heat.

[0260] The various modules in the greenhouse gas emission monitoring devices of the aforementioned electronic equipment manufacturers can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0261] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0262] The equipment-level power parameters, fossil fuel combustion emissions, and net emissions from purchased electricity and heat of electronic equipment manufacturing enterprises are obtained, and the equipment-level power parameters are preprocessed to obtain a time-series power dataset.

[0263] Based on time-series power datasets, we determine the typical production status patterns of electronic equipment manufacturing enterprises and extract the operational characteristics of key equipment in these enterprises under these typical production status patterns.

[0264] The operational features are input into a pre-built knowledge graph to obtain mapping results, which include the production process of key equipment and the greenhouse gas emission factors corresponding to the production process.

[0265] Based on the mapping results, operating characteristics, fossil fuel combustion emissions, and net emissions from purchased electricity and heat, the greenhouse gas emissions of electronic equipment manufacturing enterprises are determined.

[0266] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0267] The equipment-level power parameters, fossil fuel combustion emissions, and net emissions from purchased electricity and heat of electronic equipment manufacturing enterprises are obtained, and the equipment-level power parameters are preprocessed to obtain a time-series power dataset.

[0268] Based on time-series power datasets, we determine the typical production status patterns of electronic equipment manufacturing enterprises and extract the operational characteristics of key equipment in these enterprises under these typical production status patterns.

[0269] The operational features are input into a pre-built knowledge graph to obtain mapping results, which include the production process of key equipment and the greenhouse gas emission factors corresponding to the production process.

[0270] Based on the mapping results, operating characteristics, fossil fuel combustion emissions, and net emissions from purchased electricity and heat, the greenhouse gas emissions of electronic equipment manufacturing enterprises are determined.

[0271] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, performs the following steps:

[0272] The equipment-level power parameters, fossil fuel combustion emissions, and net emissions from purchased electricity and heat of electronic equipment manufacturing enterprises are obtained, and the equipment-level power parameters are preprocessed to obtain a time-series power dataset.

[0273] Based on time-series power datasets, we determine the typical production status patterns of electronic equipment manufacturing enterprises and extract the operational characteristics of key equipment in these enterprises under these typical production status patterns.

[0274] The operational features are input into a pre-built knowledge graph to obtain mapping results, which include the production process of key equipment and the greenhouse gas emission factors corresponding to the production process.

[0275] Based on the mapping results, operating characteristics, fossil fuel combustion emissions, and net emissions from purchased electricity and heat, the greenhouse gas emissions of electronic equipment manufacturing enterprises are determined.

[0276] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0277] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0278] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for monitoring greenhouse gas emissions from an electronic equipment manufacturing enterprise, characterized in that, The method includes: The equipment-level power parameters, fossil fuel combustion emissions, and net emissions from purchased electricity and heat from electronic equipment manufacturing enterprises are obtained, and the equipment-level power parameters are preprocessed to obtain a time-series power dataset. Based on the time-series power dataset, the typical production status mode of the electronic equipment manufacturing enterprise is determined, and the operating characteristics of key equipment in the electronic equipment manufacturing enterprise are extracted under the typical production status mode. The operational features are input into a pre-constructed knowledge graph to obtain a mapping result, which includes the production process of the key equipment and the greenhouse gas emission factor corresponding to the production process. The greenhouse gas emissions of the electronic equipment manufacturing enterprise are determined based on the mapping results, the operating characteristics, the emissions from fossil fuel combustion, and the net emissions from purchased electricity and heat.

2. The method according to claim 1, characterized in that, The step of inputting the operational features into a pre-constructed knowledge graph to obtain the mapping result includes: The knowledge graph is constructed based on prior knowledge in the field of electronic device manufacturing; The operational features are input into the knowledge graph to determine the matching relationship between the operational features and entities in the knowledge graph; The mapping result is determined based on rule-based reasoning and the matching relationship.

3. The method according to claim 2, characterized in that, The method further includes: The operational features are compared with the standard process patterns in the knowledge graph using multi-dimensional similarity calculations to obtain similarity values ​​for each dimension. The various similarity values ​​are weighted and fused to determine the global similarity. Based on the global similarity, the confidence level of the mapping result is obtained. When the confidence level meets a preset condition, the mapping result is accepted.

4. The method according to claim 1, characterized in that, The determination of the greenhouse gas emissions of the electronic equipment manufacturing enterprise based on the mapping results, the operating characteristics, the emissions from fossil fuel combustion, and the net emissions from purchased electricity and heat includes: Based on the mapping results, the production process of the key equipment is determined, and the production process includes continuous operation process and batch process; Based on the production process, the operating characteristics, and the greenhouse gas emission factor, the process emissions are determined. The greenhouse gas emissions are determined based on the process emissions, the fossil fuel combustion emissions, and the net emissions from the purchased electricity.

5. The method according to claim 4, characterized in that, Determining process emissions based on the production process, the operating characteristics, and the greenhouse gas emission factor includes: When the production process is the continuous operation process, the process emissions are obtained based on the process equivalent power and operating time in the operating characteristics and the greenhouse gas emission factor; When the production process is a batch-type process, the process emission is obtained based on the number of events and the emission amount per event in the operating characteristics.

6. The method according to claim 1, characterized in that, The process of determining the current typical production status mode of the electronic equipment manufacturing enterprise based on the time-series power dataset includes: Based on the time-series power dataset, the active power sequence in the time-series power dataset is segmented to obtain multiple load curve segments; Based on each of the load curve segments, the clustering characteristics of each of the load curve segments are determined; Clustering of each load curve segment is performed based on the dynamic time warp distance, and the optimal number of clusters is determined by the silhouette coefficient method to obtain the clustering results; Based on the clustering results and clustering characteristics, the typical production status pattern of the current electronic equipment manufacturing enterprise is obtained.

7. The method according to claim 1, characterized in that, The extraction of operational characteristics of key equipment in the electronic equipment manufacturing enterprise includes: Based on the time-series power dataset, determine the total active power sequence of the key equipment; The total active power sequence is decomposed using wavelet packet decomposition, and the decomposed sub-band signals are filtered and reconstructed to obtain the reconstructed signal. Event detection is performed on the reconstructed signal to determine the operating characteristics of the key equipment.

8. The method according to claim 1, characterized in that, The process of preprocessing the device-level power parameters to obtain a time-series power dataset includes: Based on the density threshold elimination method, abnormal data in the equipment-level power parameters are eliminated to obtain the power parameters after eliminating abnormal data. The power parameters after removing outlier data are standardized to obtain the time-series power dataset.

9. The method according to claim 8, characterized in that, The method further includes: The time-series power dataset is sorted according to the time series to determine the time-series data matrix; Based on each time point in the time series data matrix, determine the comprehensive power load sequence; Based on the first-order autocorrelation coefficient in the comprehensive power load sequence, the time-series correlation degree is determined, which characterizes the smoothness of the comprehensive power load sequence. The power parameters of each data point in the time-series power dataset are combined with the time-series correlation degree to obtain an enhanced feature vector; Abnormal data in the enhanced feature vector are removed using a density clustering algorithm.

10. A greenhouse gas emission monitoring device for an electronic equipment manufacturing enterprise, characterized in that, The device includes: The acquisition module is used to acquire the equipment-level power parameters, fossil fuel combustion emissions, and net emissions from purchased electricity and heat from electronic equipment manufacturing enterprises, and to perform data preprocessing on the equipment-level power parameters to obtain a time-series power dataset. The extraction module is used to determine the current typical production state mode of the electronic equipment manufacturing enterprise based on the time-series power dataset, and extract the operating characteristics of key equipment in the electronic equipment manufacturing enterprise under the typical production state mode; An input module is used to input the operational features into a pre-constructed knowledge graph to obtain a mapping result, the mapping result including the production process of the key equipment and the greenhouse gas emission factor corresponding to the production process; The determination module is used to determine the greenhouse gas emissions of the electronic equipment manufacturing enterprise based on the mapping results, the operating characteristics, the emissions from fossil fuel combustion, and the net emissions from purchased electricity and heat.

Images