Carbon footprint prediction method and device based on multi-angle similarity matching and medium

Abstract

The invention discloses a carbon footprint prediction method and device based on multi-angle similarity matching and a medium, belongs to the technical field of carbon footprint prediction, and is used for solving the problems that in an existing method for predicting carbon footprints based on case reasoning, input product design information is lack of comprehensiveness, retrieved similar cases are lack of accuracy, and the probability of carbon footprint prediction is low. And the determined carbon emission prediction result is lack of accuracy. The method comprises the following steps: constructing a full-life-cycle information model based on a design scene; determining target design scene information of the target case and historical design scene information of the historical case based on the full life cycle information model; determining a historical case most similar to the target case based on the scene similarity between the two cases; and determining a carbon emission prediction result of the target case based on the most similar historical case. The historical case which is most similar to the target case and retrieved through the method is more accurate, and the accuracy of a carbon emission prediction result is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of carbon footprint prediction, and in particular to a carbon footprint prediction method, device and medium based on multi-angle similarity matching. Background Art

[0002] As global climate change continues to intensify, the coordinated development of economy and low carbon has become a global issue, and the "dual carbon" goal has become a common call of countries around the world. Carbon emissions from my country's manufacturing industry account for more than 80% of the country's total carbon emissions. As an indispensable and important part of modern manufacturing, electromechanical products have the characteristics of many types, large total amount, large number of manufacturing consumables, and high energy consumption. They shoulder the heavy responsibility of being a pioneer in the green economy, and it is urgent to accelerate the green and low-carbon transformation of electromechanical products. According to statistics, nearly 70% of the performance of the product and its impact on the ecological environment have been determined in the design and development stage. Therefore, conducting carbon footprint estimation of electromechanical products in the early design stage can reduce environmental problems from the source of design and achieve twice the result with half the effort.

[0003] However, in the early design stage, the lack of life cycle inventory data and the difficulty in collecting data greatly hindered data collection and made it difficult to evaluate the carbon footprint of products. Generally speaking, product design plans are updated and improved on their original plans, and there is a large amount of historical data that can support low-carbon assessment work. Case-based reasoning methods provide ideas for new problems by reusing past solutions and experiences, showing superior flexibility and providing a basis for carbon footprint estimation. For example, Guo et al. proposed a new semantic retrieval method based on information content and node distance weighting to retrieve the most suitable cases for reuse. Jiang et al. used the nearest neighbor algorithm to calculate the similarity between new cases and retrieved cases, and used a comprehensive evaluation method to obtain the closest matching case for remanufacturing process planning. From the above analysis, it can be seen that the determination of "input information" and a reasonable similarity matching method are necessary prerequisites for case-based reasoning, which ultimately affect the reliability and accuracy of product carbon emission estimation.

[0004] In case-based reasoning, the "input information" to be retrieved in the target case must include all the design information of the product to ensure its comprehensiveness and the accuracy of the retrieval results. Generally, the expression of design scheme information mainly relies on design modeling, which is divided into unit process-oriented models and life cycle-oriented models. The unit process-oriented model is essentially a low-carbon design method that targets specific design goals or focuses on a single life cycle stage. However, this method can easily form information islands. Therefore, it is crucial to build a life cycle model to uniformly describe the design information of the entire life cycle of the product. The life cycle modeling method includes a variety of technologies, including function-behavior-structure (FBS), feature technology, graph theory, etc. For example, Fu et al. proposed a constraint-driven product concept design model based on FBS, and considered the constraint evaluation criteria in the FBS reasoning process to obtain the optimal concept solution. Kong et al. combined the FBS model with feature technology to establish a life cycle-oriented model that includes function, structure, design features, and processing features. The model clearly expresses the association mechanism of design information and efficiently generates low-carbon design solutions through heuristic algorithms.

[0005] In addition, similarity matching plays a key role in the complex process of case-based reasoning. This method can quickly and accurately retrieve the most relevant cases from existing cases by quantifying the similarity between cases. At present, the most widely used research is similarity matching based on text (such as strings and sentences). For example, He et al. proposed an ontology-based rapid knowledge modeling method for remanufacturing process planning, which obtained the most similar situation by calculating the local and global similarity between attributes, thereby supporting the planning of remanufacturing processes. Ren et al. used fuzzy set theory to represent the uncertain design attributes of product cases, and constructed an efficient index structure based on the improved discretized maximum similarity method to generate an index set of similar cases. Belwal et al. developed a new graph-based matching technology that not only considers the similarity between individual sentences, but also the similarity between sentences and the overall document.

[0006] In the process of retrieving input information, the design model of product design schemes generally focuses on the design information that directly affects the carbon emissions of products. However, some other factors have a significant impact on carbon emissions at various stages of the product life cycle, such as energy type and geographical location, which are ignored in the modeling process. Umeda et al. pointed out that regional differences may lead to differences in environmental regulations, resource requirements, labor costs, etc., which may affect the design of product life cycles. Kong et al. also analyzed the environmental impact of the same design scheme under different energy types. The results show that the impact of hydropower on the environment is about 49.62% lower than that of thermal power, indicating that the impact of differences in energy types on the environment cannot be ignored. In order to describe this information, Fukushige et al. proposed the concept of scenario, which summarized all design information at the end of the life cycle in the form of 5W1H, that is, product design information in the recycling and treatment scenario. The description of the scenario concept can systematically describe the connection and interaction between the product and the external environment. In the design process, the definition, association, change and assumption of scenarios throughout the product life cycle have an important impact on the evaluation and decision-making of product design schemes. Therefore, the "input information" based on case reasoning should pay more attention to the dynamics and accuracy of the scenario, and there is currently a lack of such research.

[0007] To summarize, the existing method of predicting carbon footprint based on case reasoning determines the similar cases corresponding to the target case through the product design information of the target case ("input information"), and determines the carbon footprint prediction result of the target case based on the product carbon emissions of the similar cases. However, the input product design information lacks comprehensiveness, the similar cases retrieved based on the input product design information lack accuracy, and the carbon emission prediction results determined based on the similar cases lack accuracy. Summary of the invention

[0008] The present invention provides a carbon footprint prediction method, device and medium based on multi-angle similarity matching, which are used to solve the technical problems in the existing method of predicting carbon footprint based on case reasoning, that is, the input product design information lacks comprehensiveness, the retrieved similar cases lack accuracy, and the determined carbon emission prediction results lack accuracy.

[0009] In a first aspect, the present invention provides a carbon footprint prediction method based on multi-angle similarity matching, comprising:

[0010] Build a full life cycle information model based on design scenarios;

[0011] Determine target design scenario information of a target case and historical design scenario information of historical cases based on the constructed full life cycle information model based on design scenarios;

[0012] Determining the scene similarity between the target design scene information and the historical design scene information;

[0013] Based on the scenario similarity, determining the historical case that is most similar to the target case;

[0014] Based on the carbon emissions of the most similar historical case, a carbon emissions prediction result of the target case is determined.

[0015] In a feasible implementation manner, the design scenario-based full life cycle information model includes a design feature layer, and determining target design scenario information of a target case and historical design scenario information of a historical case includes:

[0016] Determine product design information in a raw material acquisition scenario of a target case and product design information in a raw material acquisition scenario of a historical case, wherein the product design information in the raw material acquisition scenario includes at least one of material type, part size, material quality, part shape, geographic location, and energy type.

[0017] In a feasible implementation manner, the design scenario-based full life cycle information model includes a processing technology layer, and determines the target design scenario information of the target case and the historical design scenario information of the historical case, including:

[0018] Determine product design information in a processing and manufacturing scenario of a target case and product design information in a processing and manufacturing scenario of a historical case, wherein the product design information in the processing and manufacturing scenario includes at least one of process type, processing power, processing time, energy type, processing equipment type and geographical location.

[0019] In a feasible implementation manner, the design scenario-based full life cycle information model includes a functional layer, and determining target design scenario information of a target case and historical design scenario information of a historical case includes:

[0020] Determine product design information in a usage scenario of a target case and product design information in a usage scenario of a historical case, wherein the product design information in the usage scenario includes at least one of function type, service life, usage condition, energy type, operating power and geographic location.

[0021] In a feasible implementation manner, the design scenario-based full life cycle information model includes a structure layer, and the target design scenario information of the target case and the historical design scenario information of the historical case are determined, including:

[0022] Determine product design information in a recycling scenario of a target case and product design information in a recycling scenario of a historical case, wherein the product design information in the recycling scenario includes at least one of a structure type, recycling quality, recycling time, recycling power, energy type, and geographic location.

[0023] In a feasible implementation manner, determining the scene similarity between the target design scene information and the historical design scene information includes:

[0024] Determining attribute similarity between the target design scene information and the historical design scene information based on the attribute of the target design scene information and the attribute of the historical design scene information;

[0025] The attribute similarity is determined as the scene similarity between the target design scene information and the historical design scene information.

[0026] In a feasible implementation manner, determining the attribute similarity between the target design scene information and the historical design scene information based on the attribute of the target design scene information and the attribute of the historical design scene information includes:

[0027] If the attribute of the target design scene information and the attribute of the historical design scene information are text, determining the text similarity between the attribute of the target design scene information and the attribute of the historical design scene information;

[0028] If the attribute of the target design scene information and the attribute of the historical design scene information are numerical values, determining the numerical similarity between the attribute of the target design scene information and the attribute of the historical design scene information;

[0029] If the attribute of the target design scene information and the attribute of the historical design scene information are intervals, determining the interval similarity between the attribute of the target design scene information and the attribute of the historical design scene information;

[0030] According to the formula Determine the attribute similarity between the target design scene information R and the historical design scene information D Among them, r i Design the i-th attribute in the scene information for the target, d i is the i-th attribute in the historical design scene information, w i is the weight of the i-th attribute, and n is the number of attributes in the target design scenario information.

[0031] In a feasible implementation manner, after determining the attribute similarity between the target design scene information and the historical design scene information, the method further includes:

[0032] Determine the structural similarity between the target design scenario information and the historical design scenario information based on the structure of the target design scenario information in the design scenario-based full life cycle information model and the structure of the historical design scenario information in the design scenario-based full life cycle information model;

[0033] Based on the attribute similarity and the structure similarity, the scene similarity between the target design scene information and the historical design scene information is determined.

[0034] In a second aspect, the present invention provides a carbon footprint prediction device based on multi-angle similarity matching, the device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor so that the at least one processor can execute the carbon footprint prediction method based on multi-angle similarity matching described in any of the above embodiments.

[0035] In a third aspect, the present invention provides a non-volatile computer storage medium, which is a non-volatile computer-readable storage medium, and the non-volatile computer-readable storage medium stores at least one program, each of which includes instructions, and when the instructions are executed by a terminal, the terminal executes the carbon footprint prediction method based on multi-angle similarity matching described in any of the above embodiments.

[0036] Compared with the prior art, the carbon footprint prediction method, device and medium based on multi-angle similarity matching provided by the present invention have the following beneficial technical effects:

[0037] (1) All historical cases in the present invention can be expressed through a full life cycle information model based on design scenarios, effectively capturing the entire life cycle information and its complex interrelationships, thereby ensuring the accuracy of subsequent similarity matching. The input information of the target case is also expressed through this model, which can comprehensively express the design information of each stage of the product life cycle. In this way, the target scenario information of the target case and the historical scenario information of the historical case expressed by the model, the historical case most similar to the target case retrieved is more accurate, and the carbon emission prediction result determined based on the most similar historical case is also more accurate.

[0038] (2) The present invention comprehensively considers the attribute information of the design scenarios and the complex structural associations between the design scenarios, and proposes a multi-angle similarity matching method based on the design scenarios. With the help of the idea of ​​knowledge reuse, the most similar cases are comprehensively retrieved in the historical case library through attribute similarity and structural similarity matching, supporting efficient and accurate estimation of carbon emissions in the early design stage of products. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0040] Figure 1 A flow chart of a carbon footprint prediction method based on multi-angle similarity matching provided by the present invention;

[0041] Figure 2 A schematic diagram of a full life cycle information model based on a design scenario provided by the present invention;

[0042] Figure 3 A schematic diagram of the classification of a design scenario provided by the present invention;

[0043] Figure 4 A schematic diagram of a multi-angle similarity matching process based on design scene attributes and structures provided by the present invention;

[0044] Figure 5 A schematic diagram of the structural association of design scenarios in a target case and a historical case provided by the present invention;

[0045] Figure 6 This is a schematic diagram comparing the actual carbon emission results with the results obtained by the multi-angle similarity matching method;

[0046] Figure 7 A schematic diagram of the structure of a carbon footprint prediction device based on multi-angle similarity matching provided by the present invention. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the specific embodiments of the present invention and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0048] Estimating carbon emissions of design solutions in the early stages of design can comprehensively consider environmental issues from the source, which is of great significance to the realization of product carbon reduction and emission reduction and dual carbon goals. However, the lack of life cycle inventory data, coupled with the complex data collection process, has brought great challenges to carbon emission assessment. To address this problem, this application uses the idea of ​​knowledge reuse based on case reasoning to propose a method for estimating carbon emissions of products in the early design stage based on multi-angle similarity matching of design scenarios. Specifically, this application defines the concept of design scenarios for product design solutions, and comprehensively describes all life cycle activities that affect product carbon emissions. A multi-domain association mechanism of design domain-life cycle domain-scenario domain is proposed to construct a product life cycle information model based on design scenarios, express product life cycle design information and its intricate interrelationships, and use it as input information for case retrieval. Subsequently, a multi-perspective similarity matching strategy that integrates design scenario attribute information and structural information is proposed, which can accurately identify the most similar cases, realize efficient reuse of historical data, and support accurate estimation of product carbon footprint. This work provides designers with a targeted strategy to obtain carbon emissions in the early stages of design, thereby promoting the optimization decision of low-carbon design.

[0049] The technical solution proposed by the present invention is described in detail below with reference to the accompanying drawings.

[0050] Figure 1 The flowchart of the carbon footprint prediction method based on multi-angle similarity matching provided by the present invention is as follows. Figure 1 As shown, the method includes the following execution steps:

[0051] S101. Build a full life cycle information model based on design scenarios.

[0052] The full life cycle information model based on design scenarios in the present invention includes product design information at each stage of the product's entire life cycle. All historical cases can be expressed through the full life cycle information model based on design scenarios, effectively capturing the entire life cycle information and its complex interrelationships, thereby ensuring the accuracy of subsequent similarity matching. The input information of the target case is also expressed through this model, which can comprehensively express the design information at each stage of the product life cycle, thereby improving the accuracy of carbon emission prediction.

[0053] For example, Figure 2 A schematic diagram of a full life cycle information model based on a design scenario provided by the present invention is shown in FIG. Figure 2As shown in the figure, a full life cycle information model based on design scenarios of function-structure-design feature-process feature is constructed to uniformly express product design information. Among them, the design feature mainly describes the shape, position, material type, quality and other information corresponding to the structure. The process feature mainly describes the processing technology that forms the design feature, including: process type, processing power, processing time, etc.

[0054] On this basis, a multi-domain mapping framework of "design domain-lifecycle domain-scenario domain" is proposed. This framework system integrates the design nodes and design scenario interfaces of the lifecycle information model and constructs a full lifecycle information model based on the design scenario, such as Figure 2 As shown. The complex associations between design scenarios are mainly expressed through a tree structure, and the design scheme can be identified as a product tree. All historical cases can be expressed through a life cycle information model based on design scenarios, effectively capturing the entire life cycle information and its complex interrelationships, thereby ensuring the accuracy of subsequent similarity matching. The input information of the target case is also expressed through this model, which can fully express the design information of each stage of the product life cycle, thereby improving the accuracy of carbon emission acquisition.

[0055] S102: Based on the constructed design scenario-based full life cycle information model, determine the target design scenario information of the target case and the historical design scenario information of the historical case.

[0056] Generally speaking, the accuracy and comprehensiveness of the "input information" are crucial to the retrieval results of similar cases. At each stage of the product life cycle, the activities and interactions between the product and its external environment are different. These changes will significantly affect the carbon emission results of the product, and ultimately challenge the accuracy of the product's low-carbon design. For example, even if the same processing technology is used in the manufacturing stage, the use of different equipment will result in different carbon emissions. Based on this, the present invention introduces the concept of design scenarios, which systematically describes all life cycle activities related to product carbon emissions, including the status and external environment of each life cycle stage during the product design process; Figure 3 A classification diagram of a design scenario provided by the present invention, such as Figure 3 As shown, according to the scenario characteristics of each stage of the life cycle, the design scenarios are divided into four categories: raw material acquisition scenarios, processing and manufacturing scenarios, usage scenarios and recycling and disposal scenarios.

[0057] As a possible implementation method, the full life cycle information model based on the design scenario includes a design feature layer, which determines the target design scenario information of the target case and the historical design scenario information of the historical case, including: determining the product design information in the raw material acquisition scenario of the target case, and the product design information in the raw material acquisition scenario of the historical case, the product design information in the raw material acquisition scenario includes at least one of material type, part size, material quality, part shape, geographical location and energy type.

[0058] As a possible implementation method, the full life cycle information model based on the design scenario includes a processing technology layer, and the target design scenario information of the target case and the historical design scenario information of the historical case are determined, including: determining the product design information in the processing and manufacturing scenario of the target case, and the product design information in the processing and manufacturing scenario of the historical case, and the product design information in the processing and manufacturing scenario includes at least one of process type, processing power, processing time, energy type, processing equipment type and geographical location.

[0059] As a possible implementation method, the full life cycle information model based on the design scenario includes a functional layer, and the target design scenario information of the target case and the historical design scenario information of the historical case are determined, including: determining the product design information in the usage scenario of the target case and the product design information in the usage scenario of the historical case, and the product design information in the usage scenario includes at least one of the function type, service life, usage condition, energy type, operating power and geographical location.

[0060] As a possible implementation method, the full life cycle information model based on the design scenario includes a structural layer, which determines the target design scenario information of the target case and the historical design scenario information of the historical case, including determining the product design information in the recycling and processing scenario of the target case, and the product design information in the recycling and processing scenario of the historical case, and the product design information in the recycling and processing scenario includes at least one of the structure type, recycling quality, recycling time, recycling power, energy type and geographical location.

[0061] S103: Determine the scene similarity between the target design scene information and the historical design scene information.

[0062] As a possible implementation method, S103 determines the scene similarity between the target design scene information and the historical design scene information, including: determining the attribute similarity between the target design scene information and the historical design scene information based on the attributes of the target design scene information and the attributes of the historical design scene information; determining the attribute similarity as the scene similarity between the target design scene information and the historical design scene information.

[0063] The four types of design scenarios defined in this invention include a variety of influencing factors, each of which has unique attributes, such as text, number, and interval attributes. For example, the lifespan of the use scenario is a number attribute, while the energy type in the manufacturing scenario belongs to a text attribute. In order to effectively retrieve the most similar historical cases from the historical case library, the different attributes of the design scenarios must be considered. Therefore, an attribute similarity matching method based on the design scenario is proposed, and the similarity of each attribute is calculated as shown below.

[0064] As a possible implementation manner, based on the attributes of the target design scene information and the attributes of the historical design scene information, determining the attribute similarity between the target design scene information and the historical design scene information includes:

[0065] (1) Text Similarity

[0066] Text similarity is a metric that quantifies the degree of similarity or dissimilarity between different text strings. There are multiple text factors in the defined design scenarios, such as energy type in the usage scenario, recycling type in the recycling scenario, geographic location in the raw material acquisition scenario, and processing technology in the manufacturing scenario. Assume that the text of the attribute in the target design scenario information is represented as r i , the corresponding information text in the historical case is represented as d i , the text similarity between two attributes can be determined mathematically as follows:

[0067]

[0068] Among them, num(r i ) and num(d i ) represent r i and d i The number of different strings, N is r i and d i The sum of the numbers of two strings.

[0069] (2) Numerical similarity

[0070] Design scenarios contain a large number of attributes that carry numerical information, and numerical similarity becomes a key indicator to quantify the degree of similarity between two entities. By calculating the numerical similarity, the design information in the case library that is closest to the numerical data can be identified. The numerical attributes of the design scenario include the operating power of the usage scenario, the recycling quality of the recycling scenario, the size of the material scenario, the processing time of the manufacturing scenario, etc. The definition of the numerical similarity between the attributes of the target design scenario information and the attributes of the historical design scenario information is summarized as follows:

[0071]

[0072] In the formula, max(ri ,d i ) is r i With d i The maximum value between .

[0073] (3) Interval similarity

[0074] Interval similarity is to quantify the similarity between two interval-valued attributes. Usually, in the design process, the attributes in the design scenario are not defined by precise, single values. They usually contain a large amount of fuzzy, interval-based data. The values ​​of the attributes of the target design scenario information and the values ​​of the attributes of the historical design scenario information are [r1, r2] and [d1, d2] respectively, then the interval similarity between the two is:

[0075]

[0076] In the formula, r max and r min Design the maximum and minimum values ​​of the lower bound of the scene information for the target; d max ,d min Indicates the maximum and minimum values ​​of the upper bound of the historical design scenario information.

[0077] Based on the above method, the attribute similarity between the historical design scene information and the target design scene information can be calculated by formula (4).

[0078]

[0079] Among them, r i Design the i-th attribute in the scene information for the target, d i is the i-th attribute in the historical design scene information, w i is the weight of the i-th attribute, and n is the number of attributes in the target design scenario information.

[0080] When similarity matching is performed, the focus is currently on aspects such as text. However, a product design scheme can be viewed as an intricate structure tree, containing multiple levels and components, and there are intricate relationships between design information. However, current similarity matching methods fail to fully explore the structural relationships between these design information, thus hindering the accurate retrieval of similar cases.

[0081] As a possible implementation method, this application systematically considers the properties of design scenarios and the complex structures between them, and proposes a multi-perspective similarity matching method based on design scenarios to retrieve the most similar cases from historical cases and allocate carbon emissions to the target cases. Figure 4A schematic diagram of a multi-angle similarity matching process based on design scene attributes and structure provided by the present invention. This method consists of two parts: 1) similarity matching based on design scene attributes; 2) similarity matching based on design scene structure.

[0082] It is worth noting that throughout the life cycle of a product design solution, there are intricate interdependencies between various design solutions, and this interdependency is mainly expressed through a tree structure. In other words, the design information of the life cycle information model based on the design scenario is interconnected through a tree structure. However, the existing similarity matching methods ignore the complex connections between various design scenarios, affecting the accuracy of the results. Therefore, in the pursuit of accurate carbon emission estimates, not only the properties of the design scenario should be considered, but also the complex structure. In view of this, this paper proposes a structural similarity matching method based on design scenarios, which uses the Levenshtein distance, i.e., the edit distance, to evaluate the structural similarity between two case trees. The details are as follows:

[0083] The target case is expressed through a lifecycle information model based on design scenarios, which interconnects the design information through a hierarchical tree structure composed of multiple interrelated graphs. The case tree consists of nodes and edges. Assume that u and v are two nodes in the same graph, and the directed edge between the two nodes is represented as (u→v). The editing operations in the graph are described by node replacement operations (u→v), node insertion operations (ε→v), and node deletion operations (u→ε). g1 Convert to tree g2 The required editing operation sequence set is denoted as k = {e1, e2, ..., e j , …e n}, where e j represents the jth edit operation. The replacement, insertion and deletion of nodes constitute a complete edit path λ. In addition, a cost function is precisely defined for each step of the edit operation to quantify the edit distance. The edit distance between two case trees is determined by the minimum sum of the cost function values ​​along a given path. The cost value of each edit operation is obtained by the cost function, and the smaller the cost value, the higher the similarity between the two objects. The edit operation c(tree g1 -tree g2 )(e j ) can be expressed as Equation (5), and the graph editing operation itself is defined as Equation (6).

[0084]

[0085] In the formula, y(tree g1 →tree g2 ) indicates that from tree g1to tree g2 All complete edit paths that can be converted; c(tree g1 -tree g2 )(e j ) indicates an edit operation j The editing cost of min represents the path with the minimum cost among all complete edit distances. Set the cost of substitution, insertion, and delegation to S respectively. j ,I j , D j , the number of edit operations is greater than or equal to 1.

[0086] The structural similarity of design scenarios between case trees can be calculated as follows.

[0087]

[0088] Where the coefficient ε is a constant used to adjust the similarity value. The cost is mapped to the interval [0,1] using the function exp(-x). exp(-x) is a monotonically decreasing function. The smaller the cost, the greater the structural similarity.

[0089] In summary, attribute similarity matching is conducive to analyzing the design scenario information itself, while structural similarity matching is conducive to analyzing the correlation structure between design information. Based on the above factors, a multi-angle similarity matching method based on design scenarios is proposed. Determining scenario similarity based on attribute similarity and structural similarity achieves more accurate information matching and effectively estimates the carbon emissions of the target case. In some embodiments, the scenario similarity function is defined as shown in formula (8).

[0090]

[0091] Among them, α and β represent the weight of attribute similarity and the weight of structural similarity respectively.

[0092] S104: Based on the scene similarity, determine the historical case that is most similar to the target case.

[0093] The scene similarity between the historical design scene information and the target design scene information in the present invention represents the similarity between the target case and the historical case. Therefore, based on the scene similarity, the historical case most similar to the target case can be determined.

[0094] In some embodiments, the historical case with the highest similarity to the target case scenario may be determined as the most similar historical case.

[0095] S105. Determine the carbon emission prediction result of the target case based on the carbon emission of the most similar historical case.

[0096] After the present invention determines the historical case most similar to the target case, the historical scenario design information of the most similar historical case is most similar to the target design scenario information of the target case. Therefore, it can be considered that the carbon emissions of the target case are closest to the carbon emissions of the similar historical case.

[0097] In some embodiments, for ease of implementation, the carbon emissions of the most similar historical case may be directly determined as the predicted result of the carbon emissions of the target case, thereby achieving carbon footprint prediction of the target case.

[0098] The present invention has the following beneficial effects:

[0099] (1) Effectiveness of multi-angle similarity matching compared with other matching methods

[0100] The proposed multi-view similarity matching strategy can reasonably and accurately identify the most similar cases. Relying solely on a single viewpoint, whether it is the attributes of the design scene or the structure between the design scenes, will limit the matching similarity results.

[0101] (2) Accuracy of multi-angle similarity matching

[0102] In order to further verify the accuracy of the proposed method, the accuracy of the estimated results was verified by collecting actual emission data. The actual carbon emission results were compared with the results obtained by the multi-view similarity matching method. The results showed that compared with other methods, the results obtained by the proposed multi-view similarity matching method were closest to the actual carbon emission results, with the smallest error of only 2.75%, thus clarifying its effectiveness and reliability.

[0103] The key points of the present invention include the following two aspects.

[0104] (1) Define the concept of design scenario to describe all life cycle activities that affect product carbon emissions, propose a multi-domain association mechanism, integrate it with the life cycle information model system, build a life cycle information model based on design scenarios, and express the design information of the product throughout its life cycle in a structured manner.

[0105] (2) Taking into account the attribute information of design scenarios and the complex structural relationships between design scenarios, a multi-angle similarity matching method based on design scenarios is proposed. With the help of the idea of ​​knowledge reuse, the most similar cases are comprehensively retrieved to support the efficient and accurate estimation of carbon emissions in the early design stage of products.

[0106] The following example uses a wind turbine gearbox with a power of 2.5MW and a service life of 20 years to verify the effectiveness of the proposed method. All data are from CRRC Shandong Wind Power Co., Ltd.

[0107] (1) Product design scenario expression

[0108] This application takes the gearbox as the “target case”. Table 1 is the gearbox design scenario information, which provides the design scenario information for each component or design node.

[0109] Table 1 Gearbox design scenario information

[0110]

[0111] The historical case data of the gearbox, including the design scenario information of the design scheme and its carbon emission data, are shown in Tables 2 to 5.

[0112] Table 2 Raw material acquisition scenario information of historical cases

[0113]

[0114] Table 3 Processing and manufacturing scenario information of historical cases

[0115]

[0116] Table 4 Usage scenario information of historical cases

[0117]

[0118] Table 5 Recycling and treatment scenario information of historical cases

[0119]

[0120] Figure 5 A schematic diagram of the structural association of design scenarios in a target case and a historical case provided by the present invention. Based on the target case and the historical case, a multi-angle similarity matching method based on the design scenario is adopted to obtain the most similar case of the gearbox. Among them, the weights of each influencing factor in the attribute similarity matching are equal, and the cost values ​​of the three operations of deletion, replacement, and insertion are set to 0.3, 0.4, and 0.3 respectively.

[0121] (2) Gearbox carbon emission estimation results

[0122] Based on the above information, the attribute similarity and structural similarity of the design scenarios between the target case and the historical cases are considered simultaneously to identify the most similar cases in the case library, as shown in Table 6.

[0123] Table 6 Multi-angle similarity matching results based on design scenarios

[0124]

[0125] Through analysis, the carbon emission estimation result of the gearbox (target case) is 171881.37kgCO2e, and the design nodes of the most similar cases retrieved are: F13, S13, DF12, DF21, MF12, MF22, MF33, MF42. It can be seen that the proposed multi-angle similarity matching strategy can reasonably and accurately identify the most similar cases. Relying solely on a single perspective, it is difficult to obtain the best results, whether it is the attributes of the design scene or the structural associations between design scenes. For example, when only considering attribute similarity, DF11 is most similar to DF1. However, their structures are very different. The method proposed in this application integrates these two perspectives, providing a comprehensive matching strategy that captures the attributes and complex multi-layer structures of product design scenarios, and accurately identifies DF12 as the most similar case to DF1. In summary, the multi-perspective similarity matching strategy proposed in this application transcends the limitations of a single perspective, and the results are more objective and realistic.

[0126] In order to further verify the accuracy of the proposed method, the accuracy of the estimated results was verified by collecting actual emission data. The energy consumption and emissions were collected using a dynamic data collection tool, and the actual carbon emissions were calculated based on the life cycle assessment method. Figure 6 This is a schematic diagram comparing the actual carbon emission results with the results obtained by the multi-angle similarity matching method. Figure 6 As shown, the results show that compared with other methods, the proposed multi-angle similarity matching method is closest to the actual conditions and has the smallest error of only 2.75%, thus clarifying its effectiveness and reliability.

[0127] Corresponding to the above-mentioned embodiment, the present invention also provides a carbon footprint prediction device based on multi-angle similarity matching. Figure 7 A structural schematic diagram of a carbon footprint prediction device based on multi-angle similarity matching provided in an embodiment of the present invention, the device may include: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, so that the at least one processor can execute the carbon footprint prediction method based on multi-angle similarity matching as described in the above embodiment.

[0128] In a possible implementation of the present invention, the aforementioned at least one processor is capable of executing: constructing a full life cycle information model based on design scenarios; determining the target design scenario information of the target case and the historical design scenario information of the historical case based on the constructed full life cycle information model based on design scenarios; determining the scenario similarity between the target design scenario information and the historical design scenario information; based on the scenario similarity, determining the historical case that is most similar to the target case; and determining the carbon emission prediction result of the target case based on the carbon emissions of the most similar historical case.

[0129] In a specific implementation, the present invention can also be provided as a non-volatile computer storage medium, wherein the storage medium is a non-volatile computer-readable storage medium, and the non-volatile computer-readable storage medium stores at least one program, each program includes instructions, and when the instructions are executed by the terminal, the terminal executes the carbon footprint prediction method based on multi-angle similarity matching as described in the above embodiment.

[0130] In a possible implementation of the present invention, the aforementioned terminal executes, constructing a full life cycle information model based on the design scenario; determining the target design scenario information of the target case and the historical design scenario information of the historical case based on the constructed full life cycle information model based on the design scenario; determining the scenario similarity between the target design scenario information and the historical design scenario information; based on the scenario similarity, determining the historical case that is most similar to the target case; and determining the carbon emission prediction result of the target case based on the carbon emissions of the most similar historical case.

[0131] The various embodiments of the present invention are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0132] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0133] The above description is only an embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A carbon footprint prediction method based on multi-angle similarity matching, characterized in that: The method comprises: Build a full life cycle information model based on design scenarios; Determine target design scenario information of a target case and historical design scenario information of historical cases based on the constructed full life cycle information model based on design scenarios; Determining the scene similarity between the target design scene information and the historical design scene information; Based on the scenario similarity, determining the historical case that is most similar to the target case; Based on the carbon emissions of the most similar historical case, a carbon emissions prediction result of the target case is determined.

2. The carbon footprint prediction method based on multi-angle similarity matching according to claim 1 is characterized in that: The design scenario-based full life cycle information model includes a design feature layer, which determines the target design scenario information of the target case and the historical design scenario information of the historical case, including: Determine product design information in a raw material acquisition scenario of a target case and product design information in a raw material acquisition scenario of a historical case, wherein the product design information in the raw material acquisition scenario includes at least one of material type, part size, material quality, part shape, geographic location, and energy type.

3. The carbon footprint prediction method based on multi-angle similarity matching according to claim 1 is characterized in that: The full life cycle information model based on design scenarios includes a processing technology layer, which determines the target design scenario information of the target case and the historical design scenario information of the historical case, including: Determine product design information in a processing and manufacturing scenario of a target case and product design information in a processing and manufacturing scenario of a historical case, wherein the product design information in the processing and manufacturing scenario includes at least one of process type, processing power, processing time, energy type, processing equipment type and geographical location.

4. The carbon footprint prediction method based on multi-angle similarity matching according to claim 1 is characterized in that: The full life cycle information model based on design scenarios includes a functional layer, which determines the target design scenario information of the target case and the historical design scenario information of the historical case, including: Determine product design information in a usage scenario of a target case and product design information in a usage scenario of a historical case, wherein the product design information in the usage scenario includes at least one of function type, service life, usage condition, energy type, operating power and geographic location.

5. The carbon footprint prediction method based on multi-angle similarity matching according to claim 1 is characterized in that: The design scenario-based full life cycle information model includes a structure layer, which determines the target design scenario information of the target case and the historical design scenario information of the historical case, including: Determine product design information in a recycling scenario of a target case and product design information in a recycling scenario of a historical case, wherein the product design information in the recycling scenario includes at least one of a structure type, recycling quality, recycling time, recycling power, energy type, and geographic location.

6. The carbon footprint prediction method based on multi-angle similarity matching according to claim 1 is characterized in that: Determining the scene similarity between the target design scene information and the historical design scene information includes: Determining attribute similarity between the target design scene information and the historical design scene information based on the attribute of the target design scene information and the attribute of the historical design scene information; The attribute similarity is determined as the scene similarity between the target design scene information and the historical design scene information.

7. The carbon footprint prediction method based on multi-angle similarity matching according to claim 6 is characterized in that: Determining the attribute similarity between the target design scene information and the historical design scene information based on the attribute of the target design scene information and the attribute of the historical design scene information includes: If the attribute of the target design scene information and the attribute of the historical design scene information are text, determining the text similarity between the attribute of the target design scene information and the attribute of the historical design scene information; If the attribute of the target design scene information and the attribute of the historical design scene information are numerical values, determining the numerical similarity between the attribute of the target design scene information and the attribute of the historical design scene information; If the attribute of the target design scene information and the attribute of the historical design scene information are intervals, determining the interval similarity between the attribute of the target design scene information and the attribute of the historical design scene information; According to the formula Determine the attribute similarity between the target design scene information R and the historical design scene information D Among them, r i Design the i-th attribute in the scene information for the target, d i is the i-th attribute in the historical design scene information, w i is the weight of the i-th attribute, and n is the number of attributes in the target design scenario information.

8. The carbon footprint prediction method based on multi-angle similarity matching according to claim 6 is characterized in that: After determining the attribute similarity between the target design scenario information and the historical design scenario information, the method further includes: Determine the structural similarity between the target design scenario information and the historical design scenario information based on the structure of the target design scenario information in the design scenario-based full life cycle information model and the structure of the historical design scenario information in the design scenario-based full life cycle information model; Based on the attribute similarity and the structure similarity, the scene similarity between the target design scene information and the historical design scene information is determined.

9. A carbon footprint prediction device based on multi-angle similarity matching, characterized in that: The device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, so that the at least one processor can execute the carbon footprint prediction method based on multi-angle similarity matching according to any one of claims 1-8.

10. A non-volatile computer storage medium, characterized in that: The storage medium is a non-volatile computer-readable storage medium, and the non-volatile computer-readable storage medium stores at least one program, each of which includes instructions, and when the instructions are executed by the terminal, the terminal executes the carbon footprint prediction method based on multi-angle similarity matching according to any one of claims 1-8.

Images