Data multi-dimensional verification method in product carbon footprint evaluation process

Abstract

The invention relates to the technical field of carbon footprint tracking management, in particular to a data multi-dimensional verification method in a product carbon footprint evaluation process. The method comprises the following steps: firstly, obtaining carbon emission of a product at each stage; further obtaining a reference full life cycle group according to the similar distribution of the carbon emissions of the stages before the target stage; further screening out an influence stage in the reference full life cycle group according to the distribution of the proportion of the carbon emission of the target stage and the carbon emission of each stage after the target stage; further obtaining a to-be-analyzed full life cycle according to the change similarity of the carbon emissions of the influence stages in different full life cycles; further according to the distribution difference of the carbon emissions of the target stage and the influence stage, obtaining a correction necessity degree and judging whether the target stage is corrected or not; and finally, when it is judged that correction is carried out, the corrected carbon emission is obtained, multi-dimensional accurate verification of abnormal or conceal data is achieved, and the accuracy and credibility of carbon footprint evaluation are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of carbon footprint tracking management, and in particular to a method for multi-dimensional verification of data in the process of product carbon footprint evaluation. Background Art

[0002] Through verification, accounting loopholes can be identified, the risk of "greenwashing" can be avoided, the credibility of environmental declarations can be improved, and international standards (such as ISO 14067) and supply chain compliance requirements can be met. At the same time, verification can optimize the emission reduction path, assist in carbon asset management, and ultimately promote the actual implementation of emission reduction throughout the life cycle.

[0003] Traditional methods verify the actual carbon emissions of products by comparing the carbon emissions in each stage of the carbon footprint of similar products. It is easily affected by modification operations in some stages of the whole life cycle, resulting in inaccurate actual carbon emissions in subsequent different stages, and ultimately inaccurate identification of the actually misreported stages. Summary of the Invention

[0004] In order to solve the technical problem that the carbon emissions in some stages of the whole life cycle are affected by the previous stages and the identification of the actually misreported stages is inaccurate, the purpose of the present invention is to provide a method for multi-dimensional verification of data in the process of product carbon footprint evaluation. The specific technical solutions adopted are as follows: Obtain the carbon emissions in each stage of the whole life cycle of each product of the product category to be analyzed; select each stage one by one as the target stage; According to the similar distribution of the carbon emissions of different products in the stage before the target stage, obtain the reference whole life cycle group of the target stage; in the reference whole life cycle group, according to the concentration degree of the distribution of the ratio of the carbon emissions of the target stage to each subsequent stage, obtain the influence factor of the target stage on each subsequent stage and screen out the influencing stages based on the influence factor; according to the similarity of the change of the carbon emissions of the influencing stages in different whole life cycles, obtain the whole life cycle to be analyzed; In all the whole life cycles to be analyzed, according to the distribution difference of the carbon emissions of the target stage and the influencing stages, obtain the necessary degree of correction and determine whether the target stage needs to be corrected; when it is determined that correction is required, according to the change of the carbon emissions of the target stage when the carbon emissions of the influencing stage increase, combined with the deviation of the carbon emissions of the target stage of each whole life cycle to be analyzed from the overall carbon emissions in the same stage, and the necessary degree of correction, obtain the corrected carbon emissions.

[0005] Further, the method for obtaining the reference whole life cycle group includes: In each stage before the target stage, group the entire life cycles in which the difference between the carbon emissions and the average value of the carbon emissions of all stages is less than a first preset threshold; take the intersection of the groups of all the entire life cycles corresponding to the stages before the target stage as the reference entire life cycle group of the target stage.

[0006] Further, the method for obtaining the influence factor includes: Select each stage after the target stage one by one as a comparison stage; sort the entire life cycles in the order of the processing time of the entire life cycle, use the entire life cycle number as the horizontal axis, and use the ratio of the carbon emissions of the target stage to the comparison stage in each entire life cycle as the vertical axis to obtain a proportional change curve; Perform vertical axis dimensionality reduction classification on the proportional change curve through the PAA algorithm, and obtain the influence factor of the target stage on the comparison stage according to the proportion of the entire life cycles in the dimensionality reduction classification with the largest number of entire life cycles and the range of the vertical axis data.

[0007] Further, the method for obtaining the entire life cycle to be analyzed includes: For each entire life cycle, sort in the order of the occurrence of the influence stage, use the influence stage as the horizontal axis, and use the corresponding carbon emissions as the vertical axis to obtain a stage change curve for each entire life cycle; Use the DTW algorithm to match the stage change curves in pairs, and obtain the change consistency factor between any two stage change curves according to the sum of the distances of all the connections in the matching result and the number of points whose corresponding points have a number greater than a second preset threshold; Divide the stage change curves into different groups to be analyzed based on the change consistency factor between the stage change curves, and take the entire life cycles corresponding to the group to be analyzed with the largest number of curves as the entire life cycles to be analyzed.

[0008] Further, the method for obtaining the group to be analyzed includes: Take any two ungrouped stage change curves as an initial group. When among the change consistency factors between the other ungrouped stage change curves and the stage change curves in the initial group, the maximum value is greater than a third preset threshold and the minimum value is greater than a fourth preset threshold, include this ungrouped stage change curve in the initial group and update the initial group; traverse all ungrouped stage change curves one by one until no stage change curve meets the inclusion condition, and take the last initial group as a group to be analyzed; Again, take any two of the ungrouped phase change curves as a new initial group, and repeat the process of obtaining the grouping to be analyzed until all the phase change curves are divided.

[0009] Further, the method for obtaining the degree of necessity for correction includes: According to the overall characteristics of the carbon emissions of all the life cycles to be analyzed in each stage, obtain the baseline carbon emissions for each stage; obtain the difference values between the impact stage and the target stage of each life cycle to be analyzed and the corresponding baseline carbon emissions; Within each life cycle to be analyzed, when the ratio of the difference value of the impact stage to the difference value of the target stage is greater than the preset difference ratio threshold, mark the corresponding impact stage as the target impact stage; According to the number of target impact stages of all the life cycles to be analyzed and the maximum number of target impact stages of a single life cycle to be analyzed, obtain the degree of necessity for correction.

[0010] Further, the method for obtaining the corrected carbon emissions includes: Among all the life cycles to be analyzed, screen out all binary combinations that meet the following conditions: the carbon emissions of the impact stage of one life cycle to be analyzed are all greater than or equal to the carbon emissions of the impact stage of another life cycle to be analyzed; According to the main change direction of the carbon emissions of the target stage in all binary combinations, combined with the difference between the carbon emissions of the target stage and the baseline carbon emissions of each life cycle to be analyzed, and the degree of necessity for correction, correct the carbon emissions to obtain the corrected carbon emissions of the target stage of each life cycle to be analyzed; In each life cycle to be analyzed, adopt the opposite correction direction of the carbon emissions of the target stage, and divide the correction amount of the target stage to each impact stage according to the ratio of the impact factors corresponding to different impact stages and the target stage, to obtain the corrected carbon emissions of each impact stage.

[0011] Further, after obtaining the corrected carbon emissions, it also includes: Obtain the output value of each stage in the life cycle of each product of the product category to be analyzed; obtain the ratio of the output value of the target stage in the life cycle to the corrected carbon emissions as the carbon productivity of the target stage; According to the distribution characteristics of the carbon productivity, obtain the likelihood of underreporting of the target stage of each product; Determine whether there is underreporting based on the likelihood of underreporting and display the underreported data.

[0012] Further, the method for obtaining the likelihood of underreporting includes: Sort the entire life cycle according to the carbon productivity, perform ordered sample clustering using the Fisher optimal segmentation method, and in the clustering cluster with the largest sample size, obtain the underreporting manifestation factor based on the range and sample quantity of the carbon productivity. Based on the difference value between the carbon productivity of the target stage of each entire life cycle and the mean value of the carbon productivity in the clustering cluster with the largest sample size, and in combination with the underreporting manifestation factor, obtain the likelihood of underreporting of the target stage of each product.

[0013] Further, the method for determining whether there is underreporting based on the likelihood of underreporting and displaying the underreported data includes: When the likelihood of underreporting is greater than the preset underreporting threshold, it is determined that there is underreporting, and the corresponding product classification, entire life cycle number, problem stage, and the likelihood of underreporting are visually displayed in the form of a table.

[0014] The present invention has the following beneficial effects: The present invention first obtains the carbon emissions of each stage of the product to obtain the basis for data analysis; further obtains the reference entire life cycle group of the target stage to reduce the influence of the previous stages of the target stage, so as to more accurately analyze the influence of the target stage on the subsequent stages; further, according to the degree of concentration of the distribution of the ratio of the carbon emissions of the target stage to each subsequent stage, obtains the influence factor for quantifying the influence degree of the target stage on the subsequent stages, and screens out the influencing stages; further, according to the similarity of the changes in the carbon emissions of the influencing stages in different entire life cycles, screens out the entire life cycle samples mainly affected by the target stage, which is convenient for accurately correcting the carbon emissions subsequently; further, in all the entire life cycles to be analyzed, according to the distribution difference between the carbon emissions of the target stage and the influencing stage, reflects the influence situation of the change in the carbon emissions of the target stage, obtains the necessary degree of correction and determines whether to correct, preventing misjudgment from causing the correction to fail or over-correct; when it is determined to correct, according to the change in the carbon emissions of the target stage when the carbon emissions of the influencing stage increase, in combination with the deviation of the carbon emissions of the target stage of each entire life cycle to be analyzed from the overall carbon emissions of the same stage, and the necessary degree of correction, obtains the corrected carbon emissions. This method analyzes the correlation of carbon emissions in each stage of the entire life cycle, identifies the influence of the target stage on the subsequent stages, and combines the distribution similarity and deviation situation to determine whether it is necessary to correct the carbon emission data, realizing multi-dimensional and accurate verification of abnormal or underreported data, and improving the accuracy and credibility of carbon footprint assessment. Description of the Drawings

[0015] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0016] Figure 1 Flowchart of a method for multi-dimensional verification of data in the evaluation process of product carbon footprint provided by an embodiment of the present invention; Figure 2 Classification schematic diagram of a proportional change curve provided by an embodiment of the present invention; Figure 3 Schematic diagram of the clustering result of carbon productivity provided by an embodiment of the present invention. Detailed implementation manners

[0017] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation manners, structures, features, and effects of a method for multi-dimensional verification of data in the evaluation process of product carbon footprint proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0019] The following specifically describes the specific solution of a method for multi-dimensional verification of data in the evaluation process of product carbon footprint provided by the present invention with reference to the accompanying drawings.

[0020] Please refer to Figure 1 , which shows the flowchart of a method for multi-dimensional verification of data in the evaluation process of product carbon footprint provided by an embodiment of the present invention, specifically including: Step S1: Obtain the carbon emissions of each stage in the whole life cycle of each product of the product category to be analyzed; select each stage as the target stage one by one.

[0021] For similar products, in the whole life cycle, if the carbon emissions change due to the change of operation steps in a relatively early stage, the carbon emissions of some subsequent stages may decrease or increase relatively due to the modification of the operation. If the carbon emissions of different stages are directly used for the analysis of the stage to be improved, it may lead to the omission of the key stage. Therefore, more detailed analysis of the carbon emissions is required.

[0022] In an embodiment of the present invention, similar products are classified through the Nice Classification System. It is required that product manufacturers upload the carbon emissions of each stage in the entire life cycle of the product, and at the same time, transmit the corresponding output value of each stage.

[0023] Select each stage one by one as the target stage. The analysis method for each stage is the same. Here, only one stage is taken as an example for description; the Nice Classification System is an existing classification method and will not be elaborated.

[0024] Step S2: According to the similar distribution of the carbon emissions of different products in the stage before the target stage, obtain the reference entire life cycle group of the target stage; in the reference entire life cycle group, according to the concentration degree of the distribution of the ratio of the carbon emissions of the target stage to each subsequent stage, obtain the influence factor of the target stage on each subsequent stage and screen out the influencing stages based on the influence factor; according to the similarity of the changes in the carbon emissions of the influencing stages in different entire life cycles, obtain the entire life cycle to be analyzed.

[0025] Because any change in the operation mode of each stage before a single stage may affect the carbon emissions of this stage. To reduce the influence of variables, when analyzing the influence of the target stage on the carbon emissions of subsequent stages, first, according to the similar distribution of the carbon emissions of different products in the stage before the target stage, obtain the reference entire life cycle group of the target stage, reduce the influence of the stages before the target stage, and analyze the influence of the target stage on subsequent stages more accurately.

[0026] Preferably, in an embodiment of the present invention, considering that the average value of all the carbon emissions of a stage represents the general carbon emission situation of this stage, so based on the average value of all the carbon emissions of a stage, the entire life cycle with a small difference from the benchmark is regarded as meeting the general situation, and thus the entire life cycle that meets the general situation of all the stages before the target stage; Based on this, in each stage before the target stage, the entire life cycles with the difference between the carbon emissions and the average value of all the carbon emissions of the stage less than the first preset threshold are grouped into one group; the intersection of the groups of the entire life cycles corresponding to all the stages before the target stage is used as the reference entire life cycle group of the target stage.

[0027] As an example, the difference between the carbon emissions is represented by the absolute value of the difference. Since there may be a large difference in the carbon emissions of different stages, the absolute value of the difference between the carbon emissions and the average value of all the carbon emissions of the stage to which it belongs is used as the numerator, and the average value of all the carbon emissions of the stage to which it belongs is used as the denominator, and the fractional ratio is used as the carbon emission difference value. The first preset threshold is set to 30%; each stage before the target stage is analyzed in turn to obtain multiple groups; Take the intersection of all the groups in the full life cycle corresponding to the stages before the target stage as the reference full life cycle group of the target stage; filter out the full life cycles that conform to the general situation of all the stages before the target stage by means of intersection, so as to show the similar distribution of the carbon emissions in the stages before the target stage.

[0028] It should be noted that when the target stage is the first stage, since there is no previous stage and the operations in the first stage may affect each subsequent stage, it can be set not to perform screening and directly take all the full life cycles as the reference full life cycle group.

[0029] In another embodiment of the present invention, considering that 3 Principle (3-sigma principle) can identify the distribution of data, and set the error range of carbon emissions within the range of the mean value, and in each stage before the target stage, divide the full life cycles with carbon emissions within the range of the mean value into a group, where is the standard deviation.

[0030] Considering that the influence degree of the target stage on the subsequent stages is different, and the concentration degree of the distribution of the ratio of carbon emissions between stages reflects the association strength between the target stage and the subsequent stages and shows the influence degree of the target stage on the subsequent stages, so in the reference full life cycle group, according to the concentration degree of the distribution of the ratio of carbon emissions between the target stage and each subsequent stage, obtain the influence factor of the target stage on each subsequent stage and screen out the influence stages based on the influence factor, quantify the influence degree of the target stage on the subsequent stages, and at the same time screen out the stages greatly affected by the target stage, which is convenient for subsequent screening of the full life cycles to be analyzed and obtaining the corrected carbon emissions.

[0031] Preferably, in an embodiment of the present invention, the method for obtaining the influence factor includes: First, select each stage after the target stage as the comparison stage one by one to ensure that each subsequent stage is analyzed; Considering that in different full life cycles, the ratio of carbon emissions between the target stage and the comparison stage is different, in order to facilitate the analysis of the concentration characteristics of the distribution of the carbon emissions ratio, sort the full life cycles in the order of the processing time of the full life cycles, take the full life cycle number as the horizontal axis, and take the ratio of the carbon emissions between the target stage and the comparison stage in each full life cycle as the vertical axis to obtain the ratio change curve; Considering that the PAA algorithm can reduce the dimension and classify data, the vertical axis of the proportional change curve is reduced in dimension and classified by the PAA algorithm. When the proportion of samples in a certain classification in the total samples is larger and the range of data values in the classification is smaller, it indicates that the similarity ratio between the target stage and the comparison stage in the reference full life cycle group is larger and the distribution is more concentrated. Therefore, according to the proportion of the full life cycle in the dimension reduction classification with the most full life cycles and the range of the vertical axis data, the distribution concentration degree of the proportion of carbon emissions between the target stage and the comparison stage is shown, and the influence factor of the target stage on the comparison stage is obtained.

[0032] As an example, please refer to Figure 2 , which shows a classification schematic diagram of a proportional change curve provided by an embodiment of the present invention. Figure 2 In the figure, the horizontal axis is the full life cycle serial number, the vertical axis is the proportion, and the ordinate is the ratio of carbon emissions. Each horizontal dashed line corresponds to the boundary line of a PAA segment, and the area between two adjacent horizontal dashed lines is a dimension reduction classification; the horizontal dashed lines at the edges (the uppermost and lowermost) and the side without adjacent horizontal dashed lines form a dimension reduction classification.

[0033] When fitting the curve, the curve can be fitted by the least squares method, or the adjacent data points can be directly connected to obtain the proportional change curve.

[0034] In the dimension reduction classification with the most full life cycles, the ratio of the number of full life cycles to the total number of full life cycles in all classifications is used as the proportion. The product of the reciprocal of the range of the vertical axis data within the classification and the proportion of the full life cycle is linearly normalized, and the normalized result is used as the influence factor of the target stage on the comparison stage.

[0035] The preset influence factor threshold is 0.7. When the influence factor of the target stage on the comparison stage is greater than the preset influence factor threshold, it indicates that the correlation between the carbon emissions of the target stage and the comparison stage is strong, and the impact of the operation change in the target stage on the carbon emissions of the comparison stage is large. This comparison stage is used as the influence stage.

[0036] Since the full life cycles of actual different products are affected not only by the operation modification of a single stage but also possibly by the operation changes of their own or other certain stages, the change in their carbon emissions will involve multiple aspects, and it is impossible to directly correct the carbon emission impact amount of a certain stage through the carbon emission change situation of the relevant stage.

[0037] For the influencing stage with a greater impact on the target stage, when there are no other stages or very few other stages except the target stage that have an impact on the influencing stage, under the change of the carbon emissions of the target stage in different full life cycles, multiple influencing stages will have similar trends in carbon emissions changes in different full life cycles.

[0038] Therefore, according to the similarity of the carbon emissions changes of the influencing stage in different full life cycles, the full life cycle to be analyzed is obtained, and the full life cycle samples dominated by the target stage are screened out to accurately analyze the impact of the target stage on the carbon emissions of subsequent stages, so as to analyze and correct the abnormal situation of the carbon emissions of the target stage.

[0039] Preferably, in an embodiment of the present invention, considering that the DTW algorithm can analyze the similarity features between time series data, for each full life cycle, it is sorted according to the appearance order of the influencing stage, with the influencing stage as the horizontal axis and the corresponding carbon emissions as the vertical axis to obtain the stage change curve of each full life cycle; The DTW algorithm is used to match the stage change curves in pairs, and the DTW algorithm is used to analyze the similarity of the carbon emissions changes of the influencing stage in different full life cycles; Considering that in the matching results of two stage change curves, when the sum of the distances of the connecting lines of all matching points is smaller, the change trends of the two curves are more consistent; at the same time, when the number of points whose corresponding points on the curve exceed the second preset threshold is smaller, it indicates that the local matching degree is higher and the change is more consistent; Based on this, according to the number of points whose sum of the distances of all connecting lines and the number of corresponding points in the matching results are greater than the second preset threshold, the change consistency factor between any two stage change curves is obtained; As an example, the second preset threshold is 4. In the matching results of two stage change curves, the product of the number of corresponding points on the two curves that exceed 4 and the sum of the distances of all connecting lines is taken, and then 0.1 (a preset non-zero positive parameter) is added and then the reciprocal is taken. After linear normalization of the reciprocal result, it is used as the change consistency factor between these two stage change curves, representing the similarity of the carbon emissions changes of the influencing stage in different full life cycles.

[0040] Wherein the corresponding point of a point on the curve is the number of endpoints where the point has a connecting line with another curve, which is also the number of connecting lines. When the change trends of the two curves are highly similar, the points of the two curves correspond one by one; linear normalization is performed in the corresponding data dimension. The normalization methods adopted in the embodiments of the present invention can all adopt this method. For example, here it is the normalization in the data dimension corresponding to the reciprocal result. It belongs to the well-known technical means of those skilled in the art together with the DTW algorithm and the PAA algorithm, and will not be elaborated here.

[0041] The phase change curves are divided into different groups to be analyzed based on the change consistency factor between the phase change curves, and the entire life cycle corresponding to the group to be analyzed with the largest number of curves is taken as the entire life cycle to be analyzed.

[0042] Preferably, in an embodiment of the present invention, analogous to the multi-round hierarchical growth clustering controlled by the similarity threshold, any two ungrouped phase change curves are used as an initial group. When the change consistency factor between the ungrouped phase change curve and the phase change curves in the initial group satisfies that the maximum value is greater than the third preset threshold and the minimum value is greater than the fourth preset threshold, this ungrouped phase change curve is incorporated into the initial group and the initial group is updated; all ungrouped phase change curves are traversed one by one until no phase change curve meets the incorporation condition, and the last initial group is taken as a group to be analyzed; Again, any two ungrouped phase change curves are used as a new initial group, and the process of obtaining the groups to be analyzed is repeated until all phase change curves are divided.

[0043] As an example, the third preset threshold is 0.8 and the fourth preset threshold is 0.6.

[0044] In another embodiment of the present invention, it can also be specified that when constructing a new initial group, based on the combination of all remaining ungrouped phase change curves, the two phase change curves corresponding to the largest change consistency factor are used to construct the new initial group.

[0045] Step S3: In all the entire life cycles to be analyzed, according to the distribution difference of the carbon emissions in the target phase and the influencing phase, obtain the necessary degree of correction and determine whether the target phase needs to be corrected; when it is determined that correction is required, according to the change in the carbon emissions in the target phase when the carbon emissions in the influencing phase increase, combined with the deviation of the carbon emissions in the target phase of each entire life cycle to be analyzed from the overall carbon emissions in the same phase, and the necessary degree of correction, obtain the corrected carbon emissions.

[0046] In the entire life cycle of a product, it is possible that although the carbon emissions in a previous stage are relatively less than those of similar products, due to the simplified modification of its operation steps, it may become the main reason for the larger carbon emissions in the subsequent processing stage compared to other similar products.

[0047] At this time, although the carbon emissions in the previous stage are small, due to the increase in the carbon emissions in the subsequent stage caused by the modification of its steps, it should bear the responsibility, so its actual carbon emission impact should be greater. The carbon emission impact of a single stage can be corrected by the degree of increase in the carbon emissions in the subsequent stage that is greatly affected when the carbon emissions in this stage change.

[0048] Before the carbon emission correction, it is also necessary to determine whether to perform the correction. Therefore, in all the life cycles to be analyzed, according to the distribution differences of the carbon emissions in the target stage and the impact stage, the impact characteristics of the carbon emission changes in the target stage on the carbon emissions in the subsequent impact stage are reflected, the necessary degree of correction is obtained, and it is determined whether to perform the correction in the target stage, so as to prevent misjudgment from causing the correction to fail or over-correct, prevent misapplying the correction to the target stage with large fluctuations itself, and improve the accuracy of the carbon emission correction.

[0049] Preferably, in an embodiment of the present invention, first, according to the overall characteristics of the carbon emissions in each stage of all the life cycles to be analyzed, the benchmark carbon emissions in each stage are obtained, and a reference benchmark for each stage is obtained, which is convenient for analyzing the distribution characteristics of the carbon emissions.

[0050] As an example, the mean value of the carbon emissions in each stage of all the life cycles to be analyzed is used as the benchmark carbon emissions in each stage.

[0051] Considering that in the life cycles to be analyzed, when the difference value between the carbon emissions in each stage and the corresponding benchmark carbon emissions is smaller, it means that the carbon emissions are closer to the benchmark carbon emissions in this stage. Therefore, the difference values between the impact stage and the target stage of each life cycle to be analyzed and the corresponding benchmark carbon emissions are obtained, and the distribution characteristics of the carbon emissions in the impact stage and the target stage are represented respectively.

[0052] As an example, the absolute value of the difference between the impact stage and the target stage of each life cycle to be analyzed and the corresponding benchmark carbon emissions is used as the corresponding difference value.

[0053] Considering that when the difference value in the impact stage is larger and the difference value in the target stage is smaller, it means that the distribution difference of the carbon emissions between the two is larger. In the form of a ratio, the distribution difference of the carbon emissions is shown by the ratio of the difference value in the impact stage to the difference value in the target stage; Considering that when the distribution difference is larger, it means that the deviation of the carbon emissions in the impact stage from the benchmark carbon emissions in the same stage is more likely to be caused by the target stage; at this time, the more cases where the ratio of the difference values in all the life cycles to be analyzed is larger, and the more cases where the ratio of the difference values in a single life cycle to be analyzed is larger, it means that the impact degree of the target stage on the carbon emissions in the subsequent impact stage is greater, and the necessary degree of correction is greater; Based on this, in each life cycle to be analyzed, when the ratio of the difference value in the impact stage to the difference value in the target stage is greater than the preset difference ratio threshold, the corresponding impact stage is marked as the target impact stage; According to the number of target impact stages in all the life cycles to be analyzed and the maximum number of target impact stages in a single life cycle to be analyzed, the necessary degree of correction is obtained.

[0054] As an example, assuming the preset difference ratio threshold is 0.7, when calculating the ratio of the difference values, the difference value in the target stage is in the denominator position; by comparing the number of target impact stages in different life cycles to be analyzed, the maximum number of target impact stages in a single life cycle to be analyzed is obtained, and the product of this maximum number and the total number of target impact stages in all life cycles to be analyzed is then linearly normalized, and the normalized result is used as the correction necessity degree.

[0055] In an embodiment of the present invention, the correction necessity threshold is set to 0.8. When the correction necessity degree is greater than the correction necessity threshold, it is determined that the target stage and its impact stages need to be corrected.

[0056] When it is determined to perform correction, the change in the carbon emission of the target stage when the carbon emission of the impact stage increases is used to judge the impact on the carbon emission of the subsequent impact stage when the carbon emission of the target stage increases or decreases; the deviation of the carbon emission of the target stage from the overall carbon emission in the same stage reflects whether the carbon emission of the target stage in the life cycle to be analyzed is too large or too small. Combining the change in the carbon emission of the target stage when the carbon emission of the impact stage increases provides the correction direction and correction amplitude; at the same time, the correction necessity degree reflects the influence degree of the target stage on the impact stage, and shows the correction amplitude from the side. Therefore, according to the change in the carbon emission of the target stage when the carbon emission of the impact stage increases, combined with the deviation of the carbon emission of the target stage in each life cycle to be analyzed from the overall carbon emission in the same stage, and the correction necessity degree, the corrected carbon emission is obtained to improve the accuracy of the carbon emission.

[0057] Preferably, in an embodiment of the present invention, first, in all life cycles to be analyzed, all binary combinations that meet the following conditions are screened out: the carbon emission of the impact stage in one life cycle to be analyzed is greater than or equal to the carbon emission of the impact stage in another life cycle to be analyzed, which is convenient for analyzing the change in the carbon emission of the target stage when the carbon emission of the impact stage increases. The change in the carbon emission of the target stage includes two types of changes: increase and decrease. The analysis is based on the change with the larger proportion among the two changes as the main change; considering that in all binary combinations, when the carbon emission of the impact stage increases, if the increase in the carbon emission of the target stage occupies a larger proportion, the larger the carbon emission of the target stage compared to the reference carbon emission, the greater the corrected carbon emission of the target stage should be; on the contrary, if the decrease in the carbon emission of the target stage occupies a larger proportion, it means that it is more likely to be the target stage. At this time, the smaller the carbon emission of the target stage compared to the reference carbon emission, the greater the corrected carbon emission of the target stage should be. Also considering that the greater the degree of correction required, the greater the impact of the target stage on the influencing stage and the greater the correction amplitude of carbon emissions. Based on this, according to the main change direction of the carbon emissions in the target stage among all binary combinations, combined with the difference between the carbon emissions in the target stage and the baseline carbon emissions of each life cycle to be analyzed, and the degree of correction required, the carbon emissions are corrected to obtain the corrected carbon emissions in the target stage of each life cycle to be analyzed; As an example, using the baseline carbon emissions to represent the overall carbon emissions in the target stage, the difference between the carbon emissions in the target stage and the baseline carbon emissions reflects the deviation of the carbon emissions in the target stage from the overall carbon emissions in the same stage; among all binary combinations, when the carbon emissions in the influencing stage increase and the proportion of the increase in the carbon emissions in the target stage is relatively large, the calculation formula for the corrected carbon emissions in the target stage includes: ; where i is the serial number of the life cycle to be analyzed, and a is the serial number of the target stage; represents the carbon emissions in the a-th target stage of the i-th life cycle to be analyzed; represents the corrected carbon emissions in the a-th target stage of the i-th life cycle to be analyzed; represents the tanh function; represents the baseline carbon emissions in the a-th target stage; represents the degree of correction required in the a-th target stage.

[0058] In the formula, by means of the tanh function, the difference between the target stage and the baseline carbon emissions determines the positive and negative, that is, determines the correction direction, and the correction amplitude is adjusted by the degree of correction required; when the carbon emissions in the influencing stage increase and the proportion of the increase in the carbon emissions in the target stage is relatively large, it indicates that the abnormal increase in the carbon emissions in the subsequent influencing stage is caused by the abnormal increase in the carbon emissions in the target stage, the larger, the relatively larger the carbon emissions in the target stage, and the larger the carbon emissions in the target stage should be; at the same time the larger, the greater the correction amplitude and the larger the corrected carbon emissions.

[0059] When the carbon emissions in the influencing stage increase and the proportion of the decrease in the carbon emissions in the target stage is relatively large, the calculation formula for the corrected carbon emissions in the target stage includes: ; When the carbon emissions in the influencing stage increase and the proportion of the decrease in the carbon emissions in the target stage is relatively large, it indicates that the abnormal increase in the carbon emissions in the subsequent influencing stage is caused by the abnormal decrease in the carbon emissions in the target stage, The larger it is, it indicates that the carbon emissions in the target stage are relatively smaller, and the carbon emissions in the target stage should be larger; at the same time The larger it is, the greater the correction amplitude, and the greater the corrected carbon emissions.

[0060] After correcting the carbon emissions in the target stage, it is also necessary to synchronously correct the influencing stage to eliminate the influence on the carbon emissions in the influencing stage; considering that the influencing factors corresponding to different influencing stages and the target stage are different, the larger the influencing factor, the greater the influencing degree, and the corresponding adjustment amplitude should be larger; after the carbon emissions caused by the target stage are borne by its own stage after correction, other influencing stages need to be adjusted in the opposite direction; Based on this, in each life cycle to be analyzed, the opposite correction direction of the carbon emissions in the target stage is adopted, and the correction amount of the target stage is divided among each influencing stage according to the ratio of the influencing factors corresponding to different influencing stages and the target stage, so as to obtain the corrected carbon emissions of each influencing stage.

[0061] As an example, in a life cycle to be analyzed, when the carbon emissions in the target stage are corrected to increase, the other influencing stages are corrected to decrease. Taking the correction amount of the carbon emissions in the target stage as the total correction amount of all influencing stages, and taking the ratio of the influencing factor corresponding to each influencing stage and the target stage to the sum of all influencing factors as the distribution ratio of the total correction amount, the carbon emissions of each influencing stage are corrected to decrease, so as to obtain the corrected carbon emissions of each influencing stage.

[0062] In other embodiments of the present invention, the implementer can also add a sensitive coefficient to adjust the correction sensitivity of the tanh function, such as , , is the sensitive coefficient, and the sensitivity of the tanh function is lower. Under the same independent variable, after being adjusted by the sensitive coefficient, the correction degree is lower.

[0063] It should be noted that when it is determined that no correction is required, the original carbon emissions can be directly regarded as the corrected carbon emissions after correction, skipping the target stage at this time, and analyzing the next target stage.

[0064] It should be noted that when selecting the target stage, according to the chronological order of the stages in the life cycle, start from the second last stage and select in reverse order; when analyzing the subsequent target stages, use the carbon emission data that has been corrected before; for example, after analyzing and correcting the b-th stage, when analyzing the a-th stage, use the latest corrected carbon emission data.

[0065] In an embodiment of the present invention, after obtaining the corrected carbon emissions, it further includes analyzing the corrected carbon emissions after correction, determining whether there is data misreporting, and displaying the misreported data, specifically including: Considering that for some stages of the product's entire life cycle, there is a relatively close relationship between the output value of the stage and the carbon emissions. When the output value is inconsistent with the carbon emissions statistically in the stage, it is more likely to have misreporting of carbon emissions in this stage. Therefore, obtain the output value of each stage in the entire life cycle of each product of the product category to be analyzed; and obtain the ratio of the output value of the target stage in the entire life cycle to the corrected carbon emissions as the carbon productivity of the target stage, associate the output value and the carbon emissions, eliminate the influence of the output value fluctuation of the same type of different products, and facilitate the subsequent accurate analysis of the misreporting possibility of the product in the target stage; Considering that the distribution characteristics of carbon productivity can intuitively reflect the abnormal situation of the carbon productivity in the target stage, therefore, according to the distribution characteristics of carbon productivity, obtain the misreporting possibility of the target stage of each product.

[0066] Preferably, in an embodiment of the present invention, the method for obtaining the misreporting possibility includes: Considering that the Fisher segmentation method is used for cluster analysis of ordered data and can identify the clusters with the normal distribution of carbon productivity, so sort the entire life cycle by carbon productivity and perform ordered sample clustering with the Fisher optimal segmentation method. Please refer to Figure 3 which shows a schematic diagram of the clustering result of a carbon productivity provided by an embodiment of the present invention, Figure 3 where the horizontal axis is the serial number of the entire life cycle, the vertical axis is the ratio of the output value to the carbon emissions, and the carbon emissions here are the corrected carbon emissions after correction. Each ellipse corresponds to a clustering cluster and contains all products of the product category to be analyzed.

[0067] Considering that in the clustering cluster with the largest sample size, the more the sample quantity and the smaller the range of carbon productivity, it indicates that the distribution of carbon productivity within the clustering cluster is more concentrated and stable, reflecting that in the different entire life cycles corresponding to the same type of products in the target stage, the relationship between the output value and the carbon emissions is closer, and it can better reflect the problem of misreporting of carbon emissions. Therefore, in the clustering cluster with the largest sample size, obtain the misreporting manifestation factor according to the range of carbon productivity and the sample quantity.

[0068] As an example, in the clustering cluster with the largest sample size, use the sample quantity as the numerator, the range of carbon productivity as the denominator, and the fractional ratio as the misreporting manifestation factor.

[0069] Further considering that the mean carbon productivity in the cluster with the largest sample size can reflect the normal performance, taking this as a benchmark, the greater the deviation of the carbon productivity in the target stage of a full life cycle from the benchmark, and the higher the concealment manifestation factor, it indicates that the corresponding product is more likely to be concealed in the target stage. Therefore, according to the difference value between the carbon productivity in the target stage of each full life cycle and the mean carbon productivity in the cluster with the largest sample size, combined with the concealment manifestation factor, the concealment possibility of the target stage of each product is obtained.

[0070] As an example, the absolute value of the difference between the carbon productivity in the target stage of each full life cycle and the mean carbon productivity in the cluster with the largest sample size is used as the difference value, and the product of the difference value and the concealment manifestation factor is used as the concealment possibility of the target stage of each product.

[0071] Finally, based on the concealment possibility, it is determined whether there is concealment and the concealed data is displayed.

[0072] As an example, when the concealment possibility of a product in the target stage is greater than 80% of all the concealment possibilities in the target stage, it is determined that there is concealment.

[0073] Preferably, in an embodiment of the present invention, when the concealment possibility is greater than the preset concealment threshold, it is determined that there is concealment, and the corresponding product classification, full life cycle number, problem stage, and concealment possibility are visually displayed in the form of a table. For example, [001, 00101, 2, 87%], which respectively correspond to the product classification, full life cycle number, problem stage, and concealment possibility.

[0074] In other embodiments of the present invention, the implementer can adjust each threshold in the embodiments of the present invention according to actual needs. For example, for more stringent concealment detection, the threshold is adjusted to 75%. When the concealment possibility of a product in the target stage is greater than 75% of all the concealment possibilities in the target stage, it is determined that there is concealment.

[0075] In summary, in view of the technical problem that the carbon emissions in some stages of the whole life cycle are affected by the previous stages and the identification of the actual misreporting stage is not accurate enough, the present invention proposes a multi-dimensional data verification method in the evaluation process of product carbon footprint. The present invention first obtains the carbon emissions of each stage of the product; further obtains a reference whole life cycle group according to the similar distribution of the carbon emissions of the stages before the target stage; further, in the reference whole life cycle group, filters out the influencing stages according to the distribution of the ratio of the carbon emissions of the target stage to each subsequent stage; further obtains the whole life cycle to be analyzed according to the similarity of the changes in the carbon emissions of the influencing stages in different whole life cycles; further obtains the necessary degree of correction and determines whether the target stage needs to be corrected according to the distribution difference between the carbon emissions of the target stage and the influencing stages; finally, when it is determined that correction is required, obtains the corrected carbon emissions. By analyzing the correlation of the carbon emissions of each stage in the whole life cycle, identifying the influence of the target stage on the subsequent stages, and combining the distribution similarity and deviation situation, this method determines whether it is necessary to correct the carbon emission data, realizing multi-dimensional and accurate verification of abnormal or misreported data, and improving the accuracy and credibility of carbon footprint assessment.

[0076] It should be noted that the above sequence of embodiments of the present invention is only for description and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0077] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other. The key points of each embodiment are the differences from other embodiments.

Claims

1. A multi-dimensional data verification method in the evaluation process of product carbon footprint, characterized in that The method includes: Obtaining the carbon emissions of each product in each stage of the entire life cycle of each product category to be analyzed; selecting each stage one by one as the target stage; According to the similar distribution of the carbon emissions of different products in the stage before the target stage, obtaining the reference entire life cycle group of the target stage; in the reference entire life cycle group, according to the degree of concentration of the distribution of the ratio of the carbon emissions of the target stage to each subsequent stage, obtaining the influence factor of the target stage on each subsequent stage and screening out the influencing stages based on the influence factor; according to the similarity of the changes in the carbon emissions of the influencing stages in different entire life cycles, obtaining the entire life cycle to be analyzed; In all the entire life cycles to be analyzed, according to the distribution difference of the carbon emissions of the target stage and the influencing stages, obtaining the necessary degree of correction and determining whether the target stage needs to be corrected; when it is determined that correction is required, according to the change in the carbon emissions of the target stage when the carbon emissions of the influencing stage increase, combined with the deviation of the carbon emissions of the target stage in each entire life cycle to be analyzed from the overall carbon emissions in the same stage, and the necessary degree of correction, obtaining the corrected carbon emissions.

2. The multi-dimensional data verification method in the product carbon footprint evaluation process according to claim 1, characterized in that The method for obtaining the reference entire life cycle group includes: In each stage before the target stage, dividing the entire life cycles with the difference between the carbon emissions and the average value of all the carbon emissions in the stage less than the first preset threshold into one group; taking the intersection of the groups of the entire life cycles corresponding to all the stages before the target stage as the reference entire life cycle group of the target stage.

3. A multi-dimensional data verification method in the evaluation process of product carbon footprint according to claim 1, characterized in that The method for obtaining the influence factor includes: Selecting each stage after the target stage one by one as the comparison stage; sorting the entire life cycles in the order of the processing time of the entire life cycle, taking the entire life cycle number as the horizontal axis, and taking the ratio of the carbon emissions of the target stage and the comparison stage in each entire life cycle as the vertical axis to obtain a ratio change curve; Performing vertical axis dimensionality reduction classification on the ratio change curve through the PAA algorithm, and obtaining the influence factor of the target stage on the comparison stage according to the proportion of the entire life cycles in the dimensionality reduction classification with the most entire life cycles and the range of the vertical axis data.

4. A multi-dimensional data verification method in the product carbon footprint evaluation process according to claim 1, characterized in that The method for obtaining the entire life cycle to be analyzed includes: For each entire life cycle, sorting in the order of the appearance of the influencing stages, taking the influencing stages as the horizontal axis, and taking the corresponding carbon emissions as the vertical axis to obtain the stage change curve of each entire life cycle; Using the DTW algorithm to match the stage change curves pairwise, and obtaining the change consistency factor between any two stage change curves according to the sum of the distances of all the connections in the matching result and the number of points with the number of corresponding points greater than the second preset threshold; Dividing the stage change curves into different groups to be analyzed based on the change consistency factor between the stage change curves, and taking the entire life cycles corresponding to the group to be analyzed with the most curves as the entire life cycles to be analyzed.

5. A multi-dimensional data verification method in the evaluation process of product carbon footprint according to claim 4, characterized in that, The method for obtaining the group to be analyzed includes: Take any two of the ungrouped phase change curves as an initial group. When, among the change consistency factors between the other ungrouped phase change curves and the phase change curves within the initial group, the maximum value is greater than the third preset threshold and the minimum value is greater than the fourth preset threshold, include this ungrouped phase change curve in the initial group and update the initial group; traverse all the ungrouped phase change curves one by one until no phase change curve meets the inclusion condition, and take the last initial group as a grouping to be analyzed. Again, take any two of the ungrouped phase change curves as a new initial group, and repeat the process of obtaining the grouping to be analyzed until all the phase change curves are divided.

6. A multi-dimensional data verification method in the product carbon footprint evaluation process according to claim 1, characterized in that The method for obtaining the necessary degree of correction includes: According to the overall characteristics of the carbon emissions of all the life cycles to be analyzed in each stage, obtain the benchmark carbon emissions for each stage; obtain the difference values between the impact stage and the target stage of each life cycle to be analyzed and the corresponding benchmark carbon emissions. Within each life cycle to be analyzed, when the ratio of the difference value of the impact stage to the difference value of the target stage is greater than the preset difference ratio threshold, mark the corresponding impact stage as the target impact stage. According to the number of target impact stages of all the life cycles to be analyzed and the maximum number of target impact stages of a single life cycle to be analyzed, obtain the necessary degree of correction.

7. A multi-dimensional data verification method in the evaluation process of product carbon footprint according to claim 6, characterized in that The method for obtaining the corrected carbon emissions includes: Among all the life cycles to be analyzed, screen out all binary combinations that meet the following conditions: the carbon emissions of the impact stage of one life cycle to be analyzed are all greater than or equal to the carbon emissions of the impact stage of another life cycle to be analyzed. According to the main change direction of the carbon emissions of the target stage in all binary combinations, combine the difference between the carbon emissions of the target stage and the benchmark carbon emissions of each life cycle to be analyzed, and the necessary degree of correction to correct the carbon emissions, and obtain the corrected carbon emissions of the target stage of each life cycle to be analyzed. In each life cycle to be analyzed, adopt the opposite correction direction of the carbon emissions of the target stage, and divide the correction amount of the target stage to each impact stage according to the ratio of the impact factors corresponding to different impact stages and the target stage, to obtain the corrected carbon emissions of each impact stage.

8. A multi-dimensional data verification method in the evaluation process of product carbon footprint according to claim 1, characterized in that After obtaining the corrected carbon emissions, it further includes: Obtain the output value of each stage in the life cycle of each product of the product category to be analyzed; obtain the ratio of the output value of the target stage in the life cycle to the corrected carbon emissions as the carbon productivity of the target stage. According to the distribution characteristics of the carbon productivity, obtain the likelihood of underreporting of the target stage of each product. Based on the likelihood of underreporting, determine whether there is underreporting and display the underreported data.

9. A method for multi-dimensional verification of data in the evaluation process of product carbon footprint according to claim 8, characterized in that, The method for obtaining the likelihood of underreporting includes: Sort the entire life cycle according to the carbon productivity, and perform ordered sample clustering using the Fisher optimal segmentation method. In the clustering cluster with the largest sample size, obtain the underreporting manifestation factor based on the range of the carbon productivity and the number of samples. Based on the difference value between the carbon productivity of the target stage of each entire life cycle and the mean value of the carbon productivity in the clustering cluster with the largest sample size, and in combination with the underreporting manifestation factor, obtain the underreporting probability of the target stage of each product.

10. A multi-dimensional data verification method in the evaluation process of product carbon footprint according to claim 8, characterized in that The method for determining whether there is underreporting based on the underreporting probability and displaying the underreporting data includes: When the underreporting probability is greater than the preset underreporting threshold, it is determined that there is underreporting, and the corresponding product classification, entire life cycle number, problem stage, and the underreporting probability are visually displayed in the form of a table.

Images