A product type and process-based carbon accounting method and device and related medium

Abstract

The application discloses a kind of based on product type and process carbon accounting method, device and related medium, the method includes: obtaining industrial product data, and the product type of industrial product data is marked;Classify prediction training is carried out to industrial product data by logistic regression algorithm, to build product classification prediction model;Based on product classification, the preset product carbon accounting model is classified into discrete product carbon accounting model and process type product carbon accounting model;Classify prediction is carried out to the target industrial product to be carried out carbon accounting using product classification prediction model, and the classification result of target industrial product is obtained;According to classification result, discrete product carbon accounting model or process type product carbon accounting model is selected as target product carbon accounting model, and carbon footprint accounting is carried out.The application can reduce the difficulty of carbon footprint accounting by predicting the type of industrial product, to select corresponding carbon accounting model to carry out carbon footprint accounting, can improve carbon accounting applicability and accuracy.

Description

Technical Field

[0001] This invention relates to the field of carbon emission accounting technology, and in particular to a carbon accounting method, apparatus and related media based on product type and process. Background Technology

[0002] To raise public awareness of low-carbon practices and encourage businesses to reduce carbon emissions, low-carbon products have emerged. At the macro level, promoting the consumption of low-carbon products can effectively control greenhouse gas emissions during production. At the micro level, it can improve the energy efficiency of energy-consuming products and other products, enhance material utilization, and enable businesses to produce products that are both energy-efficient and low-carbon.

[0003] The most crucial step in creating low-carbon products is calculating the carbon emissions generated throughout the product's entire lifecycle. Since product carbon accounting requires building a carbon footprint model, which presents a significant barrier for companies, currently, companies primarily invite carbon accounting implementation providers to conduct calculations according to international carbon footprint standards. However, because international carbon footprint accounting standards are only a basic framework and guidelines, and companies' production processes vary, carbon accounting implementation providers need to communicate with companies multiple times and even conduct on-site investigations to understand the actual production processes and bills of materials before finally completing the carbon accounting model. This is time-consuming and labor-intensive for both parties. Furthermore, current mainstream product carbon accounting systems are built based on process-oriented products, which are completely unsuitable for discrete products. For process-oriented products, current carbon accounting models are mainly built according to product categories. However, even within the same company producing the same product category, manufacturing processes can differ, meaning that carbon accounting models categorized solely by product category are only applicable to a specific process product, resulting in limited applicability. Summary of the Invention

[0004] This invention provides a carbon accounting method, apparatus, computer equipment, and storage medium based on product type and process, aiming to reduce the difficulty of carbon footprint accounting and improve the applicability of carbon accounting models.

[0005] In a first aspect, embodiments of the present invention provide a carbon accounting method based on product type and process, including:

[0006] Acquire industrial product data, and label and preprocess the product types of the industrial product data; wherein, product types include discrete products and process products;

[0007] The industrial product data is trained using a logistic regression algorithm for classification and prediction to construct a product classification prediction model.

[0008] Based on product classification, the pre-set product carbon accounting models are divided into discrete product carbon accounting models and process-type product carbon accounting models.

[0009] The product classification prediction model is used to classify and predict the target industrial products to be carbon accounted for, and the classification results of the target industrial products are obtained.

[0010] Based on the classification results, either a discrete product carbon accounting model or a process product carbon accounting model is selected as the target product carbon accounting model, and the carbon footprint of the target industrial product is calculated using the target product carbon accounting model.

[0011] Secondly, embodiments of the present invention provide a carbon accounting device based on product type and process, comprising:

[0012] The data preprocessing unit is used to acquire industrial product data and to label and preprocess the product types of the industrial product data; wherein, the product types include discrete products and process products;

[0013] A model unit is constructed to train the industrial product data for classification and prediction using a logistic regression algorithm, so as to construct a product classification prediction model.

[0014] The model classification unit is used to classify the pre-set product carbon accounting models into discrete product carbon accounting models and process-type product carbon accounting models based on product classification.

[0015] The classification prediction unit is used to use the product classification prediction model to classify and predict the target industrial products to be carbon accounted for, and to obtain the classification results of the target industrial products.

[0016] The carbon footprint accounting unit is used to select either a discrete product carbon accounting model or a process product carbon accounting model as the target product carbon accounting model based on the classification results, and to perform carbon footprint accounting on the target industrial product using the target product carbon accounting model.

[0017] Thirdly, embodiments of the present invention provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement a carbon accounting method based on product type and process as described in the first aspect.

[0018] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements a carbon accounting method based on product type and process as described in the first aspect.

[0019] This invention provides a carbon accounting method, apparatus, computer equipment, and storage medium based on product type and process. The method includes: acquiring industrial product data and labeling and preprocessing the product types of the industrial product data; training the industrial product data for classification prediction using a logistic regression algorithm to construct a product classification prediction model; classifying the pre-set product carbon accounting model into a discrete product carbon accounting model and a process-oriented product carbon accounting model based on product classification; using the product classification prediction model to classify and predict the target industrial product to be carbon accounted for, obtaining the classification result of the target industrial product; selecting either the discrete product carbon accounting model or the process-oriented product carbon accounting model as the target product carbon accounting model according to the classification result, and performing carbon footprint accounting on the target industrial product using the target product carbon accounting model. Compared with existing technologies, this invention constructs a product classification prediction model to classify and predict the target industrial product to be carbon accounted for, and then selects the corresponding target product carbon accounting model for carbon footprint accounting based on the classification result. This approach has a wider applicable carbon accounting range, reduces the difficulty and cost of carbon footprint accounting for products, and improves the accuracy of carbon footprint accounting. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 A flowchart illustrating a carbon accounting method based on product type and process, provided for an embodiment of the present invention;

[0022] Figure 2 A schematic diagram of a sub-process of a carbon accounting method based on product type and process provided in an embodiment of the present invention;

[0023] Figure 3 A schematic block diagram of a carbon accounting device based on product type and process provided for an embodiment of the present invention;

[0024] Figure 4 This is a schematic block diagram of a carbon accounting device based on product type and process, provided for an embodiment of the present invention. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0027] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0028] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0029] Please see below. Figure 1 , Figure 1 The flowchart of a carbon accounting method based on product type and process provided in this embodiment of the invention specifically includes steps S101 to S105.

[0030] S101. Acquire industrial product data, and label and preprocess the product types of the industrial product data; wherein, the product types include discrete products and process products;

[0031] S102. The industrial product data is trained for classification and prediction using a logistic regression algorithm to construct a product classification prediction model.

[0032] S103. Based on product classification, the pre-set product carbon accounting models are classified into discrete product carbon accounting models and process-type product carbon accounting models.

[0033] S104. Use the product classification prediction model to classify and predict the target industrial products to be carbon accounted for, and obtain the classification results of the target industrial products.

[0034] S105. Based on the classification results, select either a discrete product carbon accounting model or a process product carbon accounting model as the target product carbon accounting model, and perform carbon footprint accounting on the target industrial product using the target product carbon accounting model.

[0035] In this embodiment, industrial product data is first acquired and labeled as discrete or process-oriented products. Then, the labeled industrial product data is input into a logistic regression algorithm to construct a product classification prediction model for product classification. Simultaneously, based on the type of industrial product, a pre-defined product carbon accounting model is divided into discrete and process-oriented carbon accounting models. Subsequently, this product classification prediction model can be used to classify and predict the carbon footprint of a specified target industrial product, determining whether it is a discrete or process-oriented product. The corresponding discrete or process-oriented carbon accounting model is then selected for carbon footprint accounting.

[0036] This embodiment constructs a product classification prediction model to classify and predict the target industrial products to be carbon accounted for. Then, based on the classification results, it selects the corresponding target product carbon accounting model to perform carbon footprint accounting. This can broaden the scope of carbon accounting, reduce the difficulty and cost of carbon footprint accounting for industrial products, and improve the accuracy of carbon footprint accounting.

[0037] In specific application scenarios, when performing carbon accounting for discrete products, the embodiments of the present invention can directly match product components according to the enterprise's product BOM (Bill of Material), which can reduce the manpower and time costs of the carbon accounting implementer; when performing carbon accounting for process products, carbon accounting can be performed according to the corresponding process steps, making the modeling of process carbon accounting models simpler.

[0038] In one embodiment, step S101 includes:

[0039] The industrial product data is obtained through web crawling technology, and the industrial product data is then cleaned.

[0040] The cleaned industrial product data is then subjected to word segmentation, encoding, and normalization processes in sequence.

[0041] Record the product attributes corresponding to the industrial product data, including product name, product type, product specifications, product unit, product material, and product manufacturing materials;

[0042] The product attributes are subjected to feature engineering to transform each product attribute into a product feature, wherein the product feature includes numerical features, category features and / or word vector features.

[0043] In this embodiment, a massive amount of industrial product data (in the hundreds of thousands) of various types is first collected from the Internet using web crawling technology. The industrial product data is then labeled, and then the industrial product data is sequentially cleaned, segmented, encoded, and normalized for feature engineering to obtain product features. With the support of massive data, the accuracy of industrial product data classification is improved.

[0044] In one specific embodiment, the product type in the product attributes is converted into the category feature in the product characteristics. For example, the "electronic product" type in the product type is converted into [1,0,0,0,0,0,0,0] through feature engineering, and the "daily necessities" type in the product type is converted into [0,1,0,0,0,0,0,0] through feature engineering. [1,0,0,0,0,0,0,0] and [0,1,0,0,0,0,0,0] are the category features, which can be denoted as p1.

[0045] The product specifications in the product attributes are converted into numerical features between 0 and 1 using a normalization method. For example, if the product specification is 1kg, it is converted to 1. 1 is the numerical feature and can be denoted as p2.

[0046] The product name in the product attributes is converted into word vector features using word vector technology. For example, for the product name "plastic pencil box", the keywords "plastic" and "pencil box" are extracted using an open-source word segmentation and extraction program. Then, a third-party open word vector database (such as Tencent Word Vector) is referenced to obtain the 200-dimensional length of the two keywords, which are denoted as v1 and v2 respectively. v1 and v2 are summed and averaged to finally obtain the word vector features [-1.2, 0.9, 0.1, ..., x199, x200], which can be denoted as p3.

[0047] In one embodiment, step S102 includes:

[0048] The input feature vector of the product classification prediction model is constructed according to the following formula:

[0049] P = [p1, p2, ..., p] n ] T

[0050] In the formula, P is the input feature vector, p1, p2, ..., p n All of these are the product features described above;

[0051] The weight vector of the product classification prediction model is constructed according to the following formula:

[0052] W = [w 1， w2, ..., w n ] T

[0053] In the formula, W is the weight vector, w1, w2, ..., w n All of these are weight vectors corresponding to the product features;

[0054] The input feature vector and the weight vector are input into the logistic regression algorithm according to the following formula to construct the product classification prediction model:

[0055] y=σ(W T p+b)

[0056] in:

[0057]

[0058] In the formula, b is the deviation vector, z is the variable, e is the constant, and y is the output of the product classification prediction model.

[0059] In this embodiment, the labeled industrial product data is trained using a logistic regression algorithm to construct the input feature vector and weight vector of the product classification prediction model, ultimately obtaining the product classification prediction model. This model is then used to classify the target industrial products. The product classification prediction model outputs either y = 1 or y = -1. Here, y = 1 represents a discrete product, and y = -1 represents a process product. Of course, in other application scenarios, y = -1 can also represent a discrete product, and y = 1 can represent a process product, or other numerical values ​​can be used to represent different meanings. It is not limited to y = 1 representing a discrete product and y = -1 representing a process product.

[0060] Logistic regression is a classification method that includes a training process that derives variables from real results (i.e., the process of building a product classification prediction model by training classification prediction in this embodiment) and a testing process that guesses the results by predicting variables (i.e., the process of performing classification prediction using the product classification prediction model in this embodiment).

[0061] Furthermore, the parameters of the product classification prediction model can be combined using the following formula:

[0062] z = W T P+b

[0063] Where W represents the weight vector corresponding to each input feature vector P, and in specific applications, w1, w2, ..., w n The initial values ​​can all be 0, or all be 0.001 or other values.

[0064] In one specific embodiment, the carbon accounting method based on product type and process further includes:

[0065] A model validation set is generated using the industrial product data, and the product classification prediction model is validated and evaluated using the model validation set.

[0066] The prediction results of the product classification prediction model on the model validation set are statistically analyzed, and it is determined whether the accuracy of the prediction results is higher than a preset threshold.

[0067] If the accuracy of the prediction result is higher than a preset threshold, the product classification prediction model is deemed qualified.

[0068] If the accuracy of the prediction result is not higher than a preset threshold, the product classification prediction model is deemed unqualified, and the product classification prediction model is optimized and updated until the accuracy of the prediction result is higher than the preset threshold.

[0069] In this embodiment, the product classification prediction model is judged to be qualified by statistically analyzing the prediction results of the product classification prediction model on the model validation set and determining whether the accuracy of the prediction results is higher than a preset threshold. For example, if the preset threshold is set to 60%, and the accuracy of the prediction results is higher than 60%, the product classification prediction model is deemed qualified; if the accuracy of the prediction results is lower than 60%, the product classification prediction model is deemed unqualified, and it is optimized and updated. Industrial product data is then re-input into the product classification prediction model for classification prediction training. The product classification prediction model is repeatedly trained until it is qualified, that is, the accuracy of the prediction results is higher than 60%.

[0070] After the product classification prediction model is evaluated and deemed qualified, it is applied to predict product classification. Based on the product classification, the pre-set product carbon accounting model is classified into discrete product carbon accounting model and process product carbon accounting model. Then, the product classification prediction model is used to classify and predict the target industrial product to be carbon accounted for. The process includes: first, selecting or inputting the product and product attribute data to be carbon accounted for; second, performing data preprocessing and feature engineering on the input data (using the same method as the product classification prediction model training process); then loading the product classification prediction model to classify and predict the feature data, obtaining the classification result y of the target industrial product; and judging whether the target industrial product to be carbon accounted for is a discrete product or a process product based on the value of y.

[0071] Combination Figure 2 As shown, in one embodiment, step S105 includes steps S201 to S204.

[0072] S201. When the classification result is a discrete product, the discrete product carbon accounting model is selected as the target product carbon accounting model.

[0073] S202. Based on the discrete product carbon accounting model, the discrete product is divided into several components, the carbon emissions of each component are obtained, and the component carbon emissions of the discrete product are obtained by combining them.

[0074] S203. Obtain the carbon emissions from the assembly of the discrete product;

[0075] S204. The carbon emissions of the components and the carbon emissions of the assembly are summed into the carbon emissions of the discrete product, and the carbon emissions of the discrete product are output as the carbon footprint accounting result of the target industrial product.

[0076] In this embodiment, the carbon emissions of the discrete product include component carbon emissions and assembly carbon emissions. The discrete product is divided into several components, and the carbon emissions of each component are obtained and combined to obtain the component carbon emissions. The component carbon emissions and the assembly carbon emissions are then summed to obtain the carbon emissions of the discrete product.

[0077] For example, when it is necessary to calculate the carbon footprint of an air conditioner, the air conditioner is first broken down into components such as compressor, condenser, evaporator, four-way valve, and capillary tube. Then, the corresponding components are determined by matching specifications and / or parameters (for example, compressors are divided into positive displacement and dynamic types, of which positive displacement is further divided into piston, spiral, scroll, and rolling rotor types, and dynamic types are divided into centrifugal and axial types). Next, the carbon emissions of each component are obtained. Then, the carbon emissions of all components are combined to obtain the component carbon emissions of the air conditioner. Finally, the carbon emissions of the assembly process are combined to obtain the carbon emissions of the air conditioner, which is the carbon footprint calculation result of the air conditioner.

[0078] The carbon emissions of a component are mainly related to its quality or quantity, material, and manufacturing process. For example, to calculate the carbon emissions of the handle of a cup, you need to select the material (e.g., plastic), weight (e.g., 50 grams), and manufacturing process (e.g., blow molding). By calculating the carbon emissions generated by producing 50 grams of plastic using the blow molding process, you can obtain the carbon emissions of the handle.

[0079] Carbon emissions generated during the assembly process include input carbon emissions and output carbon emissions. Input carbon emissions include carbon emissions from energy, natural resources, and the materials themselves, as well as transportation carbon emissions. The carbon emissions from energy, natural resources, and the materials themselves are obtained by multiplying their corresponding quantities by a carbon footprint factor. Transportation carbon emissions mainly consist of emissions generated during the transportation of product components, and are obtained by multiplying the weight of the transported material by the distance and then by the transportation carbon emission factor. Furthermore, energy carbon emissions are calculated as the quantity (or mass) of energy consumed multiplied by the energy carbon footprint factor, where the energy carbon footprint factor represents the carbon emissions generated per unit mass of energy produced. The carbon emissions from natural resources and materials, in addition to the quantity consumed multiplied by the corresponding carbon footprint factor, also need to include the carbon emissions generated during transportation, which can be obtained by multiplying the mass of the transported resources or materials by the transportation distance and then by the transportation carbon emission factor.

[0080] Carbon emissions mainly include by-product carbon emissions, waste carbon emissions, and greenhouse gas carbon emissions. By-products themselves do not generate carbon emissions. By-product carbon emissions are obtained by calculating the conversion factor of the raw materials consumed in production. For example, if the input is n grams of main material and the output is x grams of by-product, that is, the actual assembly process consumes nx grams. The carbon footprint factor of the energy quantity (or mass) of the by-product is calculated to obtain the by-product carbon emissions. Waste carbon emissions are the carbon emissions generated by waste disposal plus the carbon emissions generated by waste transportation. This can be obtained by multiplying the disposal amount by the carbon emission factor and then adding the transportation weight by the distance by the transportation carbon emission factor. Greenhouse gas carbon emissions are mainly greenhouse gases directly emitted during the production process. For example, cement production involves the decomposition of calcium carbonate. We will multiply the amount of calcium carbonate by the corresponding emission factor to obtain the final amount of greenhouse gases.

[0081] The difference between carbon footprint factor and carbon emission factor is as follows: taking acetylene as an example, the carbon footprint factor of gasoline is the carbon emissions generated in the entire process of producing a unit mass of gasoline, while the carbon emission factor is the carbon emissions produced by burning a unit mass of gasoline.

[0082] In one embodiment, step S105 further includes:

[0083] When the classification result is a process-type product, the carbon accounting model for process-type products is selected as the carbon accounting model for the target product.

[0084] Obtain the manufacturing process of the process-type product, and perform adaptive process editing on the manufacturing process to obtain the process-type process of the process-type product;

[0085] Each process step in the process flow and the corresponding input / output list of each process step are combined to form a process unit.

[0086] The carbon emissions of each process unit are obtained, and the carbon emissions of all process units are aggregated into the carbon emissions of the process product. The carbon emissions of the process product are then output as the carbon footprint accounting result of the target industrial product.

[0087] In this embodiment, the manufacturing process of the process-type product is obtained and adaptive process editing is performed to obtain the process-type process. Then, the process steps of the process-type process and the corresponding input and output lists are combined into a process unit. The carbon emissions of the process unit are obtained and summarized to obtain the carbon footprint accounting result of the target industrial product.

[0088] In a specific embodiment, the manufacturing process of a process-oriented product can be divided into multiple process steps, each with a corresponding input / output list. Since some process-oriented products from different companies within the industry may lack standardized manufacturing processes, even products manufactured using similar processes by different companies may exhibit subtle differences in their manufacturing processes. Therefore, this embodiment allows for adaptive process editing, including modifying process names or editing the input / output lists corresponding to process units based on the actual process characteristics of the process-oriented product. Furthermore, if the carbon accounting model for a process-oriented product lacks certain process steps or has unique characteristics, process units can be added and edited based on these steps, broadening the system's applicability and practicality while reducing usage costs. For products with special processes and / or the companies to which these products belong, a process unit library can be established. When carbon accounting for a process-oriented product is required, the corresponding process unit can be selected from the process unit library, and its corresponding process steps and input / output lists can be retrieved.

[0089] Furthermore, product industry classifications can be added to the process-oriented product carbon accounting model. For example, when calculating the carbon emissions of a gold product, the process-oriented product carbon accounting model classifies it as a precious metal processing industry. Then, it obtains the manufacturing process of the gold product according to the precious metal industry. Gold product processing processes include CNC machining, casting, and traditional methods. Finally, each corresponding process step and its corresponding input / output list are combined into a process unit to obtain the carbon footprint accounting result of the target industrial product.

[0090] Furthermore, the input and output lists corresponding to each process step can be entered into the process-type product carbon accounting model through system integration and / or manual input. The input and output list data of the entire product life cycle to be carbon accounted for includes the input and output list data of five stages: raw material acquisition, production and manufacturing, product transportation and distribution, product use, and product recycling and disposal.

[0091] In one embodiment, step S105 further includes:

[0092] The carbon emissions of each component, the assembly carbon emissions, and the process unit carbon emissions are calculated using the emission factor method.

[0093] The emission-factor approach is the most widely applicable and prevalent carbon accounting method. The basic idea of ​​the emission-factor approach is to construct activity data and an emission factor for each emission source (i.e., the components and process units in this embodiment) based on the carbon emission inventory list. The carbon emission value is then calculated using the product of the activity data and the emission factor, according to the following formula:

[0094] GHG = AD × EF

[0095] In the formula, GHG represents greenhouse gas emissions, AD represents activity data, and EF represents the emission factor. The emission factor can be calculated based on representative measurement data. When multiple emission factors are available, they can be selected according to a preset emission factor ranking. For example, the ranking from high to low can be set as the emission factor obtained by measurement / material balance method, the emission factor obtained by the empirical coefficient of the same process / equipment, the emission factor provided by the equipment manufacturer, the regional emission factor, the national emission factor, and the international emission factor. In specific application scenarios, these can be changed according to actual needs. Specifically, the activity data for discrete products is the type and / or quantity of components; the activity data for process products is the input and output list corresponding to the process steps. The input and output list can include input energy (e.g., water, electricity, gas, etc.), product raw materials, packaging materials, and output waste (wastewater, waste gas, and solid waste).

[0096] Figure 3 A schematic block diagram of a carbon accounting device 300 based on product type and process provided in an embodiment of the present invention, the device 300 comprising:

[0097] The data preprocessing unit 301 is used to acquire industrial product data and to label and preprocess the product types of the industrial product data; wherein, the product types include discrete products and process products.

[0098] The prediction training unit 302 is used to perform classification prediction training on the industrial product data using a logistic regression algorithm to build a product classification prediction model.

[0099] The model classification unit 303 is used to classify the pre-set product carbon accounting model into discrete product carbon accounting model and process product carbon accounting model based on product classification.

[0100] The classification prediction unit 304 is used to use the product classification prediction model to classify and predict the target industrial product to be carbon accounted for, and to obtain the classification result of the target industrial product.

[0101] The carbon footprint accounting unit 305 is used to select either a discrete product carbon accounting model or a process product carbon accounting model as the target product carbon accounting model based on the classification results, and to perform carbon footprint accounting on the target industrial product through the target product carbon accounting model.

[0102] In one embodiment, the data preprocessing unit 301 includes:

[0103] The data cleaning unit is used to acquire the industrial product data through web crawling technology and to clean the industrial product data.

[0104] The data processing unit is used to perform word segmentation, encoding and normalization processing on the cleaned industrial product data in sequence.

[0105] The product attribute unit is used to record the product attributes corresponding to the industrial product data. The product attributes include product name, product type, product specifications, product unit, product material, and product manufacturing materials.

[0106] The feature engineering unit performs feature engineering processing on the product attributes, transforming each product attribute into a product feature, wherein the product feature includes numerical features, category features, and / or word vector features.

[0107] In one embodiment, the prediction training unit 302 includes:

[0108] Input feature units are used to construct the input feature vector of the product classification prediction model according to the following formula:

[0109] P = [p] 1， p2, ..., p n ] T

[0110] In the formula, P is the input feature vector, p1, p2, ..., p n All of these are the product features described above;

[0111] The weight vector unit is used to construct the weight vector of the product classification prediction model according to the following formula:

[0112] W = [w1, w2, ..., w n ] T

[0113] In the formula, W is the weight vector, w1, w2, ..., w n All of these are weight vectors corresponding to the product features;

[0114] The model building unit is used to input the input feature vector and weight vector into the logistic regression algorithm according to the following formula, thereby constructing the product classification prediction model:

[0115] y=σ(W T p+b)

[0116] in:

[0117]

[0118] In the formula, b is the deviation vector, z is the variable, e is the constant, and y is the output of the product classification prediction model.

[0119] Combination Figure 4 As shown, in one embodiment, the carbon footprint accounting unit 305 includes:

[0120] Discrete product unit 401 is used to select the discrete product carbon accounting model as the target product carbon accounting model when the classification result is a discrete product.

[0121] The component carbon emission unit 402 is used to divide the discrete product into several components based on the discrete product carbon accounting model, obtain the carbon emission of each component, and combine them to obtain the component carbon emission of the discrete product.

[0122] Assembly carbon emission unit 403 is used to obtain the assembly carbon emissions of the discrete product assembly process;

[0123] The discrete carbon emission unit 404 is used to summarize the carbon emissions of the components and the carbon emissions of the assembly into the carbon emissions of the discrete product, and output the carbon emissions of the discrete product as the carbon footprint accounting result of the target industrial product.

[0124] In one embodiment, the carbon footprint accounting unit 305 further includes:

[0125] A process-type product unit is used to select the process-type product carbon accounting model as the target product carbon accounting model when the classification result is a process-type product.

[0126] A process unit is used to acquire the manufacturing process of the process product and perform adaptive process editing on the manufacturing process to obtain the process technology of the process product.

[0127] A process step unit is used to assemble each process step in the flow-type process and the corresponding input and output list of each process step into a process unit.

[0128] A process-type carbon emission unit is used to acquire the carbon emissions of each process unit, summarize the carbon emissions of all process units into the carbon emissions of the process-type product, and output the carbon emissions of the process-type product as the carbon footprint accounting result of the target industrial product.

[0129] In one embodiment, the carbon footprint accounting unit 305 further includes:

[0130] A carbon emission calculation unit is used to calculate the carbon emissions of each of the components, the assembly carbon emissions, and the process unit carbon emissions using the emission factor method.

[0131] In one embodiment, the carbon accounting device 300 based on product type and process further includes:

[0132] The verification and evaluation unit is used to generate a model verification set using the industrial product data, and to verify and evaluate the product classification prediction model using the model verification set.

[0133] The threshold judgment unit is used to statistically analyze the prediction results of the product classification prediction model for the model validation set and determine whether the accuracy of the prediction results is higher than a preset threshold.

[0134] The model qualification unit is used to determine that the product classification prediction model is qualified if the accuracy of the prediction result is higher than a preset threshold.

[0135] The optimization and update unit is used to determine that the product classification prediction model is unqualified if the accuracy of the prediction result is not higher than a preset threshold, and to optimize and update the product classification prediction model until the accuracy of the prediction result is higher than the preset threshold.

[0136] Since the embodiments of the apparatus and the embodiments of the method correspond to each other, please refer to the description of the embodiments of the method for the embodiments of the apparatus, which will not be repeated here.

[0137] This invention also provides a computer-readable storage medium storing a computer program thereon, which, when executed, can perform the steps provided in the above embodiments. The storage medium may include various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0138] This invention also provides a computer device, which may include a memory and a processor. The memory stores a computer program, and when the processor calls the computer program in the memory, it can implement the steps provided in the above embodiments. Of course, the computer device may also include various network interfaces, power supplies, and other components.

[0139] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to in the method section. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from the principles of this application, and these improvements and modifications also fall within the protection scope of the claims of this application.

[0140] It should also be noted that, in this specification, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

Claims

1. A carbon accounting method based on product type and process, characterized in that, include: Acquire industrial product data, and label and preprocess the product types of the industrial product data; wherein, product types include discrete products and process products; The industrial product data is trained using a logistic regression algorithm for classification and prediction to construct a product classification prediction model. Based on product classification, the pre-set product carbon accounting models are divided into discrete product carbon accounting models and process-type product carbon accounting models. The product classification prediction model is used to classify and predict the target industrial products to be carbon accounted for, and the classification results of the target industrial products are obtained. Based on the classification results, either a discrete product carbon accounting model or a process product carbon accounting model is selected as the target product carbon accounting model, and the carbon footprint of the target industrial product is calculated using the target product carbon accounting model. The process of acquiring industrial product data and labeling and preprocessing the product types of the industrial product data includes: acquiring the industrial product data through web crawling technology and cleaning the industrial product data; performing word segmentation, encoding, and normalization processing on the cleaned industrial product data in sequence; recording the product attributes corresponding to the industrial product data, the product attributes including product name, product type, product specifications, product unit, and product material; and performing feature engineering processing on the product attributes to transform each product attribute into product features, wherein the product features include numerical features, category features, and / or word vector features. The step of training the industrial product data for classification and prediction using a logistic regression algorithm to construct a product classification prediction model includes: constructing the input feature vector of the product classification prediction model according to the following formula: In the formula, P is the input feature vector, p1, p2, ..., p n All of these are the product features described above; The weight vector of the product classification prediction model is constructed according to the following formula: In the formula, W is the weight vector, w1, w2, ..., w n All of these are weight vectors corresponding to the product features; The input feature vector and the weight vector are input into the logistic regression algorithm according to the following formula to construct the product classification prediction model: in: In the formula, b is the deviation vector, z is the variable, e is the constant, and y is the output of the product classification prediction model.

2. The carbon accounting method based on product type and process according to claim 1, characterized in that, The step of selecting either a discrete product carbon accounting model or a process product carbon accounting model as the target product carbon accounting model based on the classification results, and then calculating the carbon footprint of the target industrial product using the target product carbon accounting model, includes: When the classification result is a discrete product, the discrete product carbon accounting model is selected as the target product carbon accounting model. Based on the discrete product carbon accounting model, the discrete product is divided into several components, the carbon emissions of each component are obtained, and the component carbon emissions of the discrete product are obtained by combining them. Obtain the carbon emissions from the assembly process of the discrete product; The carbon emissions from the components and the carbon emissions from the assembly are aggregated into the carbon emissions of the discrete product, and the carbon emissions of the discrete product are output as the carbon footprint accounting result of the target industrial product.

3. The carbon accounting method based on product type and process according to claim 2, characterized in that, The step of selecting either a discrete product carbon accounting model or a process product carbon accounting model as the target product carbon accounting model based on the classification results, and then calculating the carbon footprint of the target industrial product using the target product carbon accounting model, further includes: When the classification result is a process-type product, the carbon accounting model for process-type products is selected as the carbon accounting model for the target product. Obtain the manufacturing process of the process-type product, and perform adaptive process editing on the manufacturing process to obtain the process-type process of the process-type product; Each process step in the process flow and the corresponding input / output list of each process step are combined to form a process unit. The carbon emissions of each process unit are obtained, and the carbon emissions of all process units are aggregated into the carbon emissions of the process product. The carbon emissions of the process product are then output as the carbon footprint accounting result of the target industrial product.

4. The carbon accounting method based on product type and process according to claim 3, characterized in that, Also includes: The carbon emissions of each component, the assembly carbon emissions, and the process unit carbon emissions are calculated using the emission factor method.

5. The carbon accounting method based on product type and process according to claim 1, characterized in that, Also includes: A model validation set is generated using the industrial product data, and the product classification prediction model is validated and evaluated using the model validation set. The prediction results of the product classification prediction model on the model validation set are statistically analyzed, and it is determined whether the accuracy of the prediction results is higher than a preset threshold. If the accuracy of the prediction result is higher than a preset threshold, the product classification prediction model is deemed qualified. If the accuracy of the prediction result is not higher than a preset threshold, the product classification prediction model is deemed unqualified, and the product classification prediction model is optimized and updated until the accuracy of the prediction result is higher than the preset threshold.

6. A carbon accounting device based on product type and process, characterized in that, include: The data preprocessing unit is used to acquire industrial product data and to label and preprocess the product types of the industrial product data; wherein, the product types include discrete products and process products; The prediction training unit is used to perform classification prediction training on the industrial product data using a logistic regression algorithm to build a product classification prediction model. The model classification unit is used to classify the pre-set product carbon accounting models into discrete product carbon accounting models and process-type product carbon accounting models based on product classification. The classification prediction unit is used to use the product classification prediction model to classify and predict the target industrial products to be carbon accounted for, and to obtain the classification results of the target industrial products. The carbon footprint accounting unit is used to select either a discrete product carbon accounting model or a process product carbon accounting model as the target product carbon accounting model based on the classification results, and to perform carbon footprint accounting on the target industrial product through the target product carbon accounting model. The data preprocessing unit is specifically used for: acquiring the industrial product data through web crawling technology and cleaning the industrial product data; performing word segmentation, encoding, and normalization processing on the cleaned industrial product data in sequence; recording the product attributes corresponding to the industrial product data, the product attributes including product name, product type, product specifications, product unit, and product material; and performing feature engineering processing on the product attributes to convert each product attribute into product features, wherein the product features include numerical features, category features, and / or word vector features. The prediction training unit is specifically used to: construct the input feature vector of the product classification prediction model according to the following formula: In the formula, P is the input feature vector, p1, p2, ..., p n All of these are the product features described above; The weight vector of the product classification prediction model is constructed according to the following formula: In the formula, W is the weight vector, w1, w2, ..., w n All of these are weight vectors corresponding to the product features; The input feature vector and the weight vector are input into the logistic regression algorithm according to the following formula to construct the product classification prediction model: in: In the formula, b is the deviation vector, z is the variable, e is the constant, and y is the output of the product classification prediction model.

7. A computer device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement a carbon accounting method based on product type and process as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements a carbon accounting method based on product type and process as described in any one of claims 1 to 5.

Images