Material supplier supply potential analysis method based on carbon level prediction

Abstract

The invention discloses a material supplier supply potential analysis method based on carbon level prediction, and belongs to the technical field of material supplier potential analysis. An enterprise index set is acquired according to enterprise historical data, and a power grid material supplier is evaluated according to the enterprise index set; and constructing a T-S fuzzy neural network prediction model to predict supplier data of a power grid material supplier, obtaining a prediction score according to the prediction supplier data and the evaluation method, analyzing data of which the prediction score is lower than a score threshold, and obtaining the supply potential of the power grid material supplier according to an analysis result. A power grid material supplier can be helped to clearly know own defects, and the supply potential of the supplier is pertinently improved based on the defects, so that the purchase bid winning probability of the supplier is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of material supplier potential analysis, and in particular to a material supplier supply potential analysis method based on carbon level prediction. Background Art

[0002] In order to effectively promote the virtuous cycle of upstream, midstream and downstream industries of typical power grid materials and achieve the purpose of consolidating, strengthening and supplementing the chain, each downstream industry has carried out different technical research and development. Taking upstream industry suppliers as the research subject, when purchasing and applying for power grid enterprises, how to predict their own comprehensive scores without knowing the evaluation details is an important way to increase the probability of suppliers winning the bid. In the field of typical power grid material research, various technologies to improve material management have been comprehensively applied, providing effective support for improving the efficiency and accuracy of material management. However, the relevant technologies of material management alone cannot increase the probability of suppliers winning the bid. It is necessary to realize the energy transformation of suppliers, adopt more environmentally friendly materials and manufacturing processes, and improve the green carbon emission level of products throughout their life cycle.

[0003] Chinese patent, publication number: CN117892856A, publication date: April 16, 2023, discloses an evaluation and prediction method and related equipment based on the green procurement of typical power grid materials, which performs preprocessing by obtaining indicator data; calculates the green carbon emission evaluation value of typical power grid materials in different procurement cycles through a preset algorithm based on the obtained standard indicator data and divides it into a training set and a prediction set; constructs an Elman neural network prediction model based on a particle swarm algorithm, calculates the fitness function through the particle swarm algorithm, and determines the optimal weight coefficient and the optimal threshold; assigns the optimal weight coefficient and the optimal threshold to the neural network; trains and predicts the assigned model based on the training set and the prediction set. Although it has a predictive effect on the green procurement of typical power grid materials, it cannot solve the predicted problems based on the prediction results, and the probability of suppliers winning the bid is not significantly improved. Summary of the invention

[0004] The present invention aims at the problem that although the existing technology predicts the carbon emission level of power grid material suppliers, it is difficult to accurately judge the supply potential of each supplier through the prediction results; a material supplier supply potential analysis method based on carbon level prediction is proposed, by obtaining an enterprise indicator set based on the enterprise historical data, evaluating the power grid material suppliers based on the enterprise indicator set, and then constructing a TS fuzzy neural network prediction model to predict the supplier data of the power grid material suppliers, and obtaining a prediction score based on the predicted supplier data and the above-mentioned evaluation method, analyzing the data with a predicted score lower than the score threshold, and obtaining the supply potential of the power grid material supplier based on the analysis results, which can help the power grid material supplier to identify its own shortcomings and improve its own supply potential in a targeted manner based on the shortcomings, thereby increasing the probability of winning the supplier procurement bid.

[0005] In a first aspect, a technical solution provided in an embodiment of the present invention is: a method for analyzing the supply potential of a material supplier based on carbon level prediction, comprising the following steps: S1. Obtain the carbon footprint data of the enterprise's historical data and analyze it to obtain the enterprise indicator set; S2. Power grid material suppliers report their own historical supplier data, and obtain the historical comprehensive score of each supplier based on the historical supplier data and enterprise indicator set; S3, constructing a TS fuzzy neural network prediction model, using historical supplier data and enterprise indicator set as inputs of the TS fuzzy neural network prediction model to obtain predicted supplier data, and obtaining a predicted score based on the predicted supplier data and enterprise indicator set; S4, constructing a prediction region of the predicted score in the time domain and an evaluation region of the score threshold in the time domain; segmenting the prediction region and / or the evaluation region based on the principle of numerical intersection to obtain segmented sub-regions, and determining a target segmented sub-region in the segmented sub-region based on a numerical comparison criterion; S5. Analyze the predicted supplier data in the target segmented sub-area to obtain problem indicators, and the power grid material suppliers obtain the supply potential of the material suppliers based on the problem indicators.

[0006] In this scheme, in order to predict the carbon emission level of power grid material suppliers and adjust their own deficiencies according to the prediction results, the historical data of power grid enterprises are first analyzed to obtain enterprise indicators that can measure the carbon emission level of power grid material suppliers, and these enterprise indicators are integrated to obtain an enterprise indicator set. Based on the enterprise indicator set, the historical supplier data of power grid material suppliers are analyzed to obtain the historical comprehensive score of each power grid material supplier. The TS fuzzy neural network is trained based on the historical comprehensive score and the enterprise indicator set to obtain the TS fuzzy neural network prediction model. The supplier data of power grid material suppliers is predicted based on the TS fuzzy neural network prediction model to obtain predicted supplier data. Then, the prediction score is obtained according to the evaluation method of S2 and the prediction score is compared with the score threshold. The predicted supplier data below the score threshold is analyzed to obtain the specific occurrence time and the weight of the enterprise indicator set indicators in the data segment. The higher the weight, the lower the prediction score. Therefore, the indicators exceeding the weight threshold are defined as problem indicators. The supply potential of material suppliers is obtained according to the problem indicators. The fewer the problem indicators and the smaller the weight of the problem indicators, the greater the supply potential. The greater the supply potential, the higher the probability of suppliers winning the bid. It can help power grid material suppliers to clarify their own shortcomings and improve their own supply potential based on the shortcomings, so as to increase the probability of suppliers winning the bid.

[0007] Preferably, in S1, obtaining carbon footprint data of historical data of an enterprise and analyzing the data to obtain an enterprise indicator set includes the following steps: Obtaining historical data of the enterprise, which at least includes carbon footprint data during the acquisition of raw and auxiliary materials, transportation of raw and auxiliary materials, product production, product transportation, and product use; Integrate the carbon footprint data to build a carbon emission index evaluation system, and obtain the enterprise indicator set based on the carbon emission index evaluation system.

[0008] In this plan, in order to obtain the enterprise indicator set, the enterprise historical data is analyzed, and the data is divided into carbon footprint data in the processes of raw and auxiliary material acquisition, raw and auxiliary material transportation, product production, product transportation, and product use according to the data type. The data generated in these processes are analyzed and calculated, and a carbon emission indicator evaluation system is constructed. The various indicators in the carbon emission indicator evaluation system are extracted and integrated into the enterprise indicator set, which provides an evaluation basis for the prediction of the future carbon emission levels of power grid material suppliers.

[0009] As a preferred embodiment, in S2, the power grid material supplier reports its own historical supplier data, and obtains the historical comprehensive score of each supplier based on the historical supplier data and the enterprise indicator set, including the following steps: Standardize the enterprise indicator set to obtain the indicator matrix and define the optimal and worst values ​​for each indicator; Analyze the historical supplier data of each power grid material supplier and each indicator in the indicator matrix in turn to obtain the indicator value corresponding to each indicator of each power grid material supplier; Based on the indicator values ​​corresponding to each indicator of each power grid material supplier, the optimal distance and the worst distance between the indicator value and the optimal value and the worst value of the corresponding indicator are obtained, and the historical comprehensive score of each supplier is obtained based on the optimal distance and the worst distance corresponding to each indicator.

[0010] In this scheme, in order to make a quantitative assessment of the carbon emission level of power grid material suppliers, the entropy weight-TOPSIS method is used to calculate the comprehensive score of the carbon emission index evaluation system. First, the enterprise indicator values ​​are standardized to eliminate the dimension effect, and a standardized matrix, namely the indicator matrix, is formed. The optimal value and the worst value corresponding to each indicator are defined based on this. The maximum and minimum data of each column in the indicator matrix are obtained to determine the impact of each indicator. Based on this, the weight calculation is performed to obtain the probability matrix, and the indicator information entropy is calculated to calculate the indicator weight. Based on the indicator weight, the optimal distance and the worst distance between the indicator value corresponding to each indicator of each power grid material supplier and the optimal value and the worst value of the corresponding indicator are obtained. In this way, the comprehensive score of each power grid material supplier is calculated, and the carbon emission level of the power grid material supplier is quantified, so that the power grid material supplier has a clearer understanding of its own carbon emission level, which is convenient for the subsequent targeted improvement of its own carbon emission level through predictive scoring.

[0011] Preferably, in S3, a TS fuzzy neural network prediction model is constructed, and the historical supplier data and the enterprise indicator set are used as inputs of the TS fuzzy neural network prediction model to obtain a prediction score, including the following steps: The input dimension is obtained based on the enterprise indicator set, the output dimension is obtained based on the historical comprehensive score, the number of membership functions, the center, width and coefficient of the fuzzy membership function are obtained through random initialization, and the TS fuzzy neural network prediction model is constructed based on the above data; Perform feature analysis on historical supplier data to obtain feature values, use the feature values ​​and enterprise indicator sets as input to the TS fuzzy neural network prediction model, output predicted supplier data, and obtain prediction scores based on the predicted supplier data and enterprise indicator sets.

[0012] In this scheme, in order to construct the TS fuzzy neural network prediction model, the TS fuzzy neural network data is first initialized. The data initialization includes determining the input and output dimensions of the neural network training samples, determining the number of membership functions, the center, width and coefficient of the fuzzy membership function. The input dimension is released by the power grid company. This scheme integrates it into an enterprise indicator set. The input dimension is determined based on the number of indicators in the enterprise indicator set, and the final output value is the predicted score. Therefore, the output dimension can be set to 1. The number of membership functions, the center, width and coefficient of the fuzzy membership function are obtained by random initialization. The fuzzy neural network is trained by sorting the training samples of the historical comprehensive scores, setting the maximum number of iterations and the number of samples, and testing with the trained data. The training data is compared with the actual data to determine whether the error is less than the specified value. If it is less than, the result is output. If not, the iteration continues until the error is less than the specified value. This means that the fuzzy neural network training is completed and a reliable TS fuzzy neural network prediction model is obtained. By predicting the comprehensive score based on the TS fuzzy neural network, the accuracy and reliability of the prediction results can be improved.

[0013] Preferably, in S4, constructing a prediction region of a prediction score in the time domain and constructing an evaluation region of a score threshold in the time domain include the following steps: A Cartesian coordinate system is established with time as the horizontal axis and score as the vertical axis. The area formed by the predicted score in the time dimension is the prediction area; the area formed by the score threshold in the time dimension is the evaluation area.

[0014] In this scheme, in order to facilitate the analysis of the predicted score, by constructing a Cartesian coordinate system, the predicted score and the score threshold are placed in the same time domain for analysis, which can clearly and intuitively reflect the size relationship between the predicted score and the score threshold, and facilitate the separation of the time domain where the predicted score is lower than the score threshold for separate analysis.

[0015] Preferably, in S4, segmenting the prediction region and / or the evaluation region based on the numerical intersection principle to obtain segmented sub-regions, and determining the target segmented sub-region in the segmented sub-regions according to the numerical comparison criterion, comprises the following steps: The prediction region and / or the evaluation region is segmented based on the intersection of the prediction score and the score threshold to obtain segmented sub-regions, and the segmented sub-regions with the prediction score less than the score threshold are taken as target segmented sub-regions.

[0016] In this solution, in order to extract and process part of the data that affects the supply potential in the predicted supplier data, data in some areas where the predicted score that is positively correlated with the supply potential is less than the score threshold is extracted, thereby reducing the difficulty of analyzing a large amount of data in sequence.

[0017] Preferably, segmenting the prediction region and / or the evaluation region based on the intersection of the prediction score and the score threshold to obtain segmented subregions, and taking the segmented subregions whose prediction scores are less than the score threshold as target segmented subregions, comprises the following steps: Get the intersection points of the predicted score and the score threshold and number the intersection points according to the time sequence; Based on the time points at which the intersections of adjacent numbers are located, the prediction region and / or the evaluation region are cut to obtain segmented sub-regions; When the predicted score in a segmentation sub-region is less than the score threshold, the corresponding segmentation sub-region is used as the target segmentation sub-region, and the target segmentation sub-region is marked by the intersection number.

[0018] In this scheme, in order to facilitate the extraction of predicted supplier data with lower carbon emission levels, a Cartesian coordinate system is constructed to segment the prediction area and / or evaluation area at the intersection of the scoring threshold and the predicted score. The area with a predicted score lower than the scoring threshold is extracted as the target segmented sub-area, and the predicted supplier data in the time area corresponding to the target segmented sub-area is analyzed. This allows for a clearer understanding of the carbon emission levels of power grid material suppliers in the prediction time period, which facilitates the subsequent reverse deduction of the reasons for the low prediction scores for the prediction time period. Adjustments and optimizations can be made based on the reasons found, thereby improving the company's own carbon emission levels in the future.

[0019] Preferably, in S5, analyzing the predicted supplier data in the target segmented sub-region to obtain the problem index includes the following steps: Describe the indicator characteristics of the predicted supplier data based on the indicators in the enterprise indicator set to obtain the indicator factors, standardize each indicator factor, and obtain the indicator value of each indicator factor; Determine the weight of each indicator factor based on the indicator value of each indicator factor and the overall indicator value of the overall indicator factor; The indicators corresponding to the indicator factors whose weights exceed the weight threshold are judged as problematic indicators.

[0020] In this scheme, in order to analyze the predicted supplier data with low prediction scores and find the reasons for the low scores, the predicted supplier data in the target segmentation sub-area is extracted, and the predicted supplier data indicator weights for the future prediction time period are obtained based on the enterprise indicator set. If the weight of an indicator in the enterprise indicator set is high, it means that the indicator has a large impact on the low prediction score, and it is necessary to prioritize this indicator to improve the supplier's carbon emission level. For example, the first and second indicators of the predicted supplier data for the future prediction time period have the highest weights, and the weights of other indicators are all below the weight threshold. This means that the first and second indicators, such as transportation distance and emissions, are not stable enough in the production process, resulting in a low prediction score. Therefore, the supplier needs to reduce transportation distance and emissions in the future prediction time period to improve its own carbon emission level. The future improvement direction can be clearly obtained, thereby significantly improving its own carbon emission level, which is more targeted.

[0021] Preferably, in S5, the power grid material supplier obtains the material supplier supply potential based on the problem indicator, including the following steps: Based on the indicator value of the problem indicator, the optimal distance and the worst distance between the indicator value of the problem indicator and the optimal value and the worst value of the corresponding indicator are obtained, and the supply potential of the material supplier is obtained based on the optimal distance and the worst distance corresponding to the problem indicator.

[0022] In this scheme, in order to obtain the supply potential of material suppliers, a method similar to that of obtaining the historical comprehensive scores of material suppliers is adopted to calculate the problem indicators separately, and the problem indicators and the corresponding indicators in the indicator matrix are analyzed in turn to obtain the optimal distance and the worst distance between the indicator value of the problem indicator and the optimal value and the worst value of the corresponding indicator. The larger the optimal distance of the problem indicator, the smaller the worst distance, and the greater the negative impact of the problem indicator on the supply potential of the material supplier. The problem indicators can be solved in order according to the order of the negative impact of each problem indicator on the supply potential to ensure the fastest improvement of the supply potential.

[0023] Preferably, the carbon emission index evaluation system at least includes: Functional unit consumables, material emissions and designed functional unit usage when obtaining raw and auxiliary materials; Transportation distance and carbon emissions of raw and auxiliary materials; Greenhouse gas emissions, fuel consumption, and electricity usage during product production; The distance travelled and the emissions generated when products are transported.

[0024] In this scheme, since power grid material suppliers need to obtain raw materials when supplying materials, when obtaining raw materials, the smaller the supplier's functional unit consumables, material emissions and design functional unit usage, the more efficient the supplier's material acquisition and the higher the utilization rate of its own resources; after obtaining the raw materials, the materials need to be transported. The smaller the transportation distance and the smaller the carbon emissions on the way during transportation, the stronger the supplier's transportation and route planning capabilities; when producing products, the smaller the supplier's greenhouse gas fugitive emissions, fuel consumption and electricity emissions, the higher the proportion of clean energy used by the supplier and the higher the energy utilization rate; the smaller the transportation distance and emissions during product transportation, the smaller the impact of the transportation vehicle selected by the supplier on the environment. The carbon emission index evaluation system constructed with the above indicators can comprehensively reflect the carbon emission level of power grid material suppliers in each link of the supply, making the evaluation of the supplier's carbon emission level more real and effective.

[0025] The beneficial effects of the present invention are as follows: (1) The present invention constructs a carbon emission index evaluation system by analyzing the historical data of power grid enterprises to comprehensively evaluate the carbon emission level of power grid material suppliers, and quantifies the manifestation of carbon emission levels, so that power grid material suppliers can have a clear understanding of their own carbon emission levels; (2) The present invention predicts the future comprehensive scores of power grid material suppliers by constructing a TS fuzzy neural network prediction model, and obtains problem indicators that will appear in the future, so that power grid material suppliers can optimize and improve according to the problem indicators, enhance their own supply potential, and thus increase the probability of suppliers winning the bid.

[0026] The above invention content is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Other features, objects and advantages of the present invention will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings. The drawings are only for the purpose of illustrating preferred embodiments and are not to be considered as limiting the present invention. Also, the same reference symbols are used throughout the drawings to represent the same parts.

[0028] Figure 1 This is a flow chart of a method for analyzing the supply potential of material suppliers based on carbon level prediction according to the present invention; Figure 2 A fuzzy neural network algorithm flow chart in an embodiment of the present invention; Figure 3This is a test data error diagram in an embodiment of the present invention; Figure 4 is a graph of predicted data and comprehensive score in an embodiment of the present invention; Figure 5 A schematic diagram of the weights of the prediction scores of the present invention. DETAILED DESCRIPTION

[0029] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific implementation method described herein is only an optimal embodiment of the present invention, which is only used to explain the present invention and does not limit the scope of protection 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.

[0030] Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flow charts. Although the flow charts describe the operations (or steps) as sequential processes, many of the operations (or steps) therein can be implemented in parallel, concurrently, or simultaneously. In addition, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but can also have additional steps not included in the drawings; the process can correspond to a method, function, procedure, subroutine, subprogram, etc.

[0031] Example: Figure 1 As shown, in order to solve the problem that although the existing technology predicts the carbon emission level of power grid material suppliers, it is difficult to accurately judge the supply potential of each supplier through the prediction results, this embodiment provides a material supplier supply potential analysis method based on carbon level prediction, including the following steps: S1: Obtain carbon footprint data from the historical data of the power grid company, and analyze the historical data to obtain the enterprise indicator set.

[0032] In this embodiment, the carbon footprint data of the enterprise's historical data is obtained from the power grid enterprise, and the carbon footprint data of the enterprise's historical data is analyzed to obtain the enterprise indicator set, including the following steps: Obtaining historical data of the enterprise, which at least includes carbon footprint data during the acquisition of raw and auxiliary materials, transportation of raw and auxiliary materials, product production, product transportation, and product use; Integrate the carbon footprint data to build a carbon emission index evaluation system, and obtain the enterprise indicator set based on the carbon emission index evaluation system.

[0033] In order to obtain the enterprise indicator set, this embodiment analyzes the enterprise historical data, divides the data into carbon footprint data in the processes of raw and auxiliary material acquisition, raw and auxiliary material transportation, product production, product transportation and product use according to the data type, analyzes and calculates the data generated in these processes, constructs a carbon emission indicator evaluation system, extracts various indicators in the carbon emission indicator evaluation system and integrates them into the enterprise indicator set, which provides an evaluation basis for predicting the future carbon emission level of power grid material suppliers.

[0034] S2: Power grid material suppliers report their own historical supplier data, and obtain the historical comprehensive score of each supplier based on the historical supplier data and enterprise indicator set.

[0035] In this embodiment, the power grid material supplier reports its own historical supplier data, and obtains the historical comprehensive score of each supplier based on the historical supplier data and the enterprise indicator set, including the following steps: Standardize the enterprise indicator set to obtain the indicator matrix and define the optimal and worst values ​​for each indicator; Analyze the historical supplier data of each power grid material supplier and each indicator in the indicator matrix in turn to obtain the indicator value corresponding to each indicator of each power grid material supplier; Based on the indicator values ​​corresponding to each indicator of each power grid material supplier, the optimal distance and the worst distance between the indicator value and the optimal value and the worst value of the corresponding indicator are obtained, and the historical comprehensive score of each supplier is obtained based on the optimal distance and the worst distance corresponding to each indicator.

[0036] In order to quantitatively evaluate the carbon emission level of power grid material suppliers, this embodiment adopts the entropy weight-TOPSIS method to calculate the comprehensive score of the carbon emission index evaluation system. First, the enterprise index values ​​are standardized to eliminate the dimension effect, and a standardized matrix, namely, an index matrix, is formed. The formula is as follows: The optimal and worst values ​​corresponding to each indicator are defined based on this, and the maximum and minimum data of each column in the indicator matrix are obtained to determine the impact of each indicator. The formula is as follows: Based on this, the weight calculation is performed to obtain the probability matrix, and the formula is as follows: And calculate the indicator information entropy, the formula is as follows; The indicator weight is calculated and the formula is as follows: Based on the indicator weight, the optimal distance and the worst distance between the indicator value corresponding to each indicator of each power grid material supplier and the optimal value and the worst value of the corresponding indicator are obtained. The formula is as follows: In this way, the comprehensive score of each power grid material supplier is calculated. The formula is as follows: Among them, Y ij is the indicator matrix, x ij is the data in the i-th row and j-th column, that is, the value of the j-th indicator under the i-th power grid material supplier, where i∈n, j∈m; and are the maximum and minimum data of the corresponding j columns of the indicator matrix respectively; based on this, the weight calculation is performed, P ij is the probability matrix, e j is the information entropy of the indicator, w j is the indicator weight; The optimal distance is the worst distance, S i For the comprehensive score.

[0037] The carbon emission levels of power grid material suppliers have been quantified, so that power grid material suppliers have a clearer understanding of their own carbon emission levels, which will facilitate targeted improvement of their own carbon emission levels through predictive scoring in the future.

[0038] S3: Build a TS fuzzy neural network prediction model, use historical supplier data and enterprise indicator set as input of the TS fuzzy neural network prediction model to obtain predicted supplier data, and obtain prediction scores based on the predicted supplier data and enterprise indicator set.

[0039] In this embodiment, a TS fuzzy neural network prediction model is constructed, and the historical supplier data and the enterprise indicator set are used as inputs of the TS fuzzy neural network prediction model to obtain a prediction score, including the following steps: The input dimension is obtained based on the enterprise indicator set, the output dimension is obtained based on the historical comprehensive score, the number of membership functions, the center, width and coefficient of the fuzzy membership function are obtained through random initialization, and the TS fuzzy neural network prediction model is constructed based on the above data; Perform feature analysis on historical supplier data to obtain feature values, use the feature values ​​and enterprise indicator sets as input to the TS fuzzy neural network prediction model, output predicted supplier data, and obtain prediction scores based on the predicted supplier data and enterprise indicator sets.

[0040] In this embodiment, a TS fuzzy neural network prediction model is constructed. Figure 2As shown, first, the TS fuzzy neural network data is initialized. The data initialization includes determining the input and output dimensions of the neural network training samples, determining the number of membership functions, the center, width and coefficient of the fuzzy membership function. The input dimension is released by the power grid company. This solution integrates it into an enterprise indicator set, including but not limited to the acquisition of raw and auxiliary materials (functional unit consumables, material emissions, design functional unit usage), transportation of raw and auxiliary materials (transportation distance, carbon emissions), product production (greenhouse gas fugitive emissions, fuel consumption, electricity emissions), product transportation (transportation distance, emissions) and other 4 first-level indicators and 10 second-level indicators. The input dimension is determined based on the number of indicators in the enterprise indicator set. The input dimension of this embodiment is 10, and the final output value is the predicted score, so the output dimension can be set to 1. The number of membership functions, the center, width and coefficient of the fuzzy membership function are obtained by random initialization, where the membership function formula is expressed as follows: In the formula, is a fuzzy set, x j is the input variable; are width and center respectively; k is the number of input parameters; n is the number of fuzzy subsets.

[0041] By sorting out the training samples of historical comprehensive scores, setting the maximum number of iterations and the number of samples to train the fuzzy neural network, firstly, the membership function is used to fuzzify the parameters, and the formula is expressed as follows: In the formula, ω i is the product of the membership of the input parameters; The output calculation of the above formula is as follows: In the formula, are the neural network coefficients.

[0042] Use the trained data for testing, compare the training data with the actual data, and determine whether the error is less than the specified value. If it is, output the result. If not, continue to iterate until the error is less than the specified value. This means that the fuzzy neural network training is completed and a reliable TS fuzzy neural network prediction model is obtained. By predicting the comprehensive score based on the TS fuzzy neural network, the accuracy and reliability of the prediction results can be improved.

[0043] This embodiment uses the proposed prediction algorithm to first divide the green carbon emission scores published in previous years into data (including training data and test data), and conducts a comprehensive score prediction for the current year through learning and training with a large amount of data from previous years; secondly, the proposed prediction algorithm uses fuzzy mathematics theory to construct a TS fuzzy neural network prediction model, which can realize automatic updating and continuous correction of membership functions, calculate the weights between each layer of the network through fuzzy multiplication, and check the final prediction results through three steps: error calculation, coefficient correction, and parameter correction; finally, by combining the bidding grade standards of power grid enterprises, the strengths and weaknesses of the suppliers themselves are confirmed, and optimization and improvement are carried out accordingly.

[0044] In this embodiment, the TS fuzzy neural network can process uncertain and imprecise data through fuzzification processing, thereby improving the robustness and adaptability of the system; at the same time, due to the easy implementation of neural networks, the TS fuzzy neural network has been modified on the basis of traditional neural networks to become more convenient, and through its powerful approximation ability, it can effectively approximate any nonlinear function. That is, the TS fuzzy neural network has certain advantages in processing uncertain and imprecise data, which can be applied in the context of suppliers not understanding the evaluation process of power grid companies.

[0045] S4: construct a prediction region for the prediction score in the time domain and an evaluation region for the score threshold in the time domain; segment the prediction region and / or the evaluation region based on the principle of numerical intersection to obtain segmentation sub-regions, and determine the target segmentation sub-region in the segmentation sub-region according to the numerical comparison criterion.

[0046] In this embodiment, constructing a prediction region of a prediction score in the time domain and constructing an evaluation region of a score threshold in the time domain include the following steps: A Cartesian coordinate system is established with time as the horizontal axis and score as the vertical axis. The area formed by the predicted score in the time dimension is the prediction area; the area formed by the score threshold in the time dimension is the evaluation area.

[0047] In order to facilitate the analysis of the predicted score, this embodiment constructs a Cartesian coordinate system and places the predicted score and the score threshold on the same time domain for analysis. This can clearly and intuitively reflect the size relationship between the predicted score and the score threshold, and facilitates the separation of the time domain where the predicted score is lower than the score threshold for separate analysis.

[0048] In this embodiment, the prediction region and / or the evaluation region are segmented based on the numerical intersection principle to obtain segmented sub-regions, and the target segmented sub-regions in the segmented sub-regions are determined according to the numerical comparison criterion, including the following steps: The prediction region and / or the evaluation region is segmented based on the intersection of the prediction score and the score threshold to obtain segmented sub-regions, and the segmented sub-regions with the prediction score less than the score threshold are taken as target segmented sub-regions.

[0049] In order to extract and process part of the data that affects the supply potential in the predicted supplier data, this embodiment extracts data from some areas where the predicted score that is positively correlated with the supply potential is less than the score threshold, thereby reducing the difficulty of analyzing a large amount of data in sequence.

[0050] In this embodiment, the prediction region and / or the evaluation region is segmented based on the intersection of the prediction score and the score threshold to obtain segmented sub-regions, and the segmented sub-regions whose prediction scores are less than the score threshold are used as target segmented sub-regions, including the following steps: obtaining the intersection of the prediction score and the score threshold and numbering the intersections according to the time sequence; Based on the time points at which the intersections of adjacent numbers are located, the prediction region and / or the evaluation region are cut to obtain segmented sub-regions; When the predicted score in a segmentation sub-region is less than the score threshold, the corresponding segmentation sub-region is used as the target segmentation sub-region, and the target segmentation sub-region is marked by the intersection number.

[0051] In order to facilitate the extraction of predicted supplier data with low carbon emission levels, this embodiment constructs a Cartesian coordinate system, divides the prediction area and / or evaluation area according to the intersection of the scoring threshold and the predicted score, extracts the area with a predicted score lower than the scoring threshold as the target segmented sub-area, and analyzes the predicted supplier data in the time area corresponding to the target segmented sub-area. This allows a clearer understanding of the carbon emission levels of power grid material suppliers in the prediction time period, facilitates the subsequent reverse deduction of the reasons for the low prediction scores for the prediction time period, and allows the system to adjust and optimize itself based on the reasons found, thereby improving its own carbon emission levels in the future.

[0052] S5: Analyze the predicted supplier data in the target segmented sub-area to obtain problem indicators, and the power grid material supplier obtains the material supplier supply potential based on the problem indicators.

[0053] In this embodiment, analyzing the predicted supplier data in the target segmentation sub-region to obtain the problem index includes the following steps: Describe the indicator characteristics of the predicted supplier data based on the indicators in the enterprise indicator set to obtain the indicator factors, standardize each indicator factor, and obtain the indicator value of each indicator factor; Determine the weight of each indicator factor based on the indicator value of each indicator factor and the overall indicator value of the overall indicator factor; The indicators corresponding to the indicator factors whose weights exceed the weight threshold are judged as problematic indicators.

[0054] In order to analyze the predicted supplier data with low prediction scores and find out the reasons for the low scores, this embodiment extracts the predicted supplier data in the target segmentation sub-area, and obtains the predicted supplier data indicator weights for the future prediction time period based on the enterprise indicator set. If the weight of an indicator in the enterprise indicator set is high, it means that the indicator has a large impact on the low prediction score, and it is necessary to prioritize the indicator to improve the supplier's carbon emission level. For example, the first indicator and the second indicator of the predicted supplier data for the future prediction time period have the highest weights, and the weights of other indicators are all below the weight threshold. This means that the first indicator and the second indicator, such as transportation distance and emissions, are not stable enough in the production link, resulting in a low prediction score. Therefore, the supplier needs to reduce transportation distance and emissions in the future prediction time period to improve its carbon emission level. The future improvement direction can be clearly obtained, thereby significantly improving its own carbon emission level, which is more targeted.

[0055] In this embodiment, the power grid material supplier obtains the supply potential of the material supplier based on the problem indicator, including the following steps: obtaining the optimal distance and the worst distance between the indicator value of the problem indicator and the optimal value and the worst value of the corresponding indicator based on the indicator value of the problem indicator, and obtaining the supply potential of the material supplier based on the optimal distance and the worst distance corresponding to the problem indicator.

[0056] In order to obtain the supply potential of the material supplier, this embodiment adopts a method similar to obtaining the historical comprehensive score of the material supplier, calculates the problem index separately, analyzes the problem index and the corresponding index in the index matrix in turn, and obtains the index value d of the problem index t The optimal distance d between the optimal and worst values ​​of the corresponding indicator + and the worst distance d - , the supply potential calculation formula is expressed as: d t ∈[d max ,d min ] d + =d max -d t d - =|d t -d min | d + =|d t -d max | where d max is the optimal value of the indicator corresponding to the problem indicator, d min is the worst value of the indicator corresponding to the problem indicator, and s is the supply potential.

[0057] The larger the optimal distance of the problem indicator, the smaller the worst distance. The greater the negative impact of the problem indicator on the supply potential of the material supplier, the smaller the supply potential. The problems can be solved in order according to the order of the negative impact of each problem indicator on the supply potential to ensure the fastest improvement of the supply potential.

[0058] In this embodiment, the carbon emission index evaluation system at least includes: Functional unit consumables, material emissions and designed functional unit usage when obtaining raw and auxiliary materials; Transportation distance and carbon emissions of raw and auxiliary materials; Greenhouse gas emissions, fuel consumption, and electricity usage during product production; The distance travelled and the emissions generated when products are transported.

[0059] In this embodiment, since the power grid material supplier needs to obtain raw materials when supplying materials, when obtaining raw materials, the smaller the functional unit consumables, material emissions and design functional unit usage of the supplier are, the higher the efficiency of the supplier's material acquisition and the utilization rate of its own resources can be reflected; after obtaining the raw materials, the materials need to be transported. During the transportation process, the smaller the transportation distance and the smaller the carbon emissions on the way, the stronger the transportation and route planning capabilities of the supplier can be reflected; when producing products, the smaller the greenhouse gas fugitive emissions, fuel consumption and electricity emissions of the supplier, the higher the proportion of clean energy used by the supplier and the higher the energy utilization rate; the smaller the transportation distance and emissions during product transportation, the smaller the impact of the transportation vehicle selected by the supplier on the environment can be reflected. The carbon emission index evaluation system constructed with the above indicators can comprehensively reflect the carbon emission level of the power grid material supplier in each link of the supply, making the evaluation of the supplier's carbon emission level more real and effective.

[0060] In order to predict the carbon emission level of power grid material suppliers and adjust their own deficiencies according to the prediction results, this embodiment first analyzes the historical data of power grid enterprises to obtain enterprise indicators that can measure the carbon emission level of power grid material suppliers, integrates these enterprise indicators to obtain an enterprise indicator set, analyzes the historical supplier data of power grid material suppliers based on the enterprise indicator set, obtains the historical comprehensive score of each power grid material supplier, trains the TS fuzzy neural network based on the historical comprehensive score and the enterprise indicator set to obtain a TS fuzzy neural network prediction model, predicts the supplier data of power grid material suppliers based on the TS fuzzy neural network prediction model to obtain predicted supplier data, and then obtains the predicted score according to the evaluation method of S2 and compares the predicted score with the score threshold, analyzes the predicted supplier data below the score threshold, obtains the specific occurrence time and the weight of the enterprise indicator set in the data, the higher the weight, the lower the predicted score, so the indicator exceeding the weight threshold is defined as a problem indicator, and obtains the supply potential of the material supplier according to the problem indicator. The fewer the problem indicators and the smaller the weight of the problem indicators, the greater the supply potential, and the greater the supply potential, the higher the probability of the supplier winning the bid, which can help power grid material suppliers to clarify their own shortcomings and improve their own supply potential in a targeted manner based on the shortcomings, thereby increasing the probability of supplier procurement winning the bid.

[0061] As a further supplement to this embodiment, the following scenario is taken as an example to further illustrate this solution: This embodiment selects the enterprise historical data and historical supplier data of each month from October 2018 to September 2023 as experimental data, analyzes the enterprise historical data to obtain the enterprise indicator set, obtains the historical comprehensive score of the historical supplier data based on the enterprise indicator set, and then obtains the input dimension based on the enterprise indicator set, obtains the output dimension based on the historical comprehensive score, obtains the number of membership functions, the center, width and coefficient of the fuzzy membership function through random initialization, and constructs a TS fuzzy neural network prediction model based on the above data; performs feature analysis on the historical supplier data to obtain feature values, uses the feature values ​​and the enterprise indicator set as inputs of the TS fuzzy neural network prediction model, and outputs predicted supplier data, such as Figure 3 As shown in the figure, the predicted supplier data and actual supplier data for each month from October 2018 to September 2023, as well as the error between them, Figure 3 It can be seen that the error is basically in a low and range, indicating that the predicted results are close to the actual results and have a certain degree of reliability.

[0062] Based on the forecast supplier data and enterprise indicator set, the forecast score is obtained, and the data from January to September 2023 is selected for forecasting. Figure 4 As shown in the figure, it is a curve chart of predicted data and actual comprehensive score. Figure 4The forecast data is basically consistent with the actual comprehensive score. Only the forecast data for August and September 2023 have a difference of about 0.001 from the actual comprehensive score, which can be basically ignored. This shows that the error between the actual output and the forecast output fluctuates within the allowable range, which can be used for comprehensive score prediction of typical power grid material suppliers. Figure 4 The overall forecast level from 2023 to September is in line with expectations, and the supplier's overall comprehensive score is on an upward trend, which meets the procurement needs of power grid companies. However, the lowest comprehensive score appears in April 2023. At this time, the power grid company can make public the weight of all data provided by the supplier, and the supplier can use this to make targeted improvements. The forecast supplier data for April 2023 is analyzed to obtain the indicator weight of the data, such as Figure 5 As shown, Figure 5 Among them, indicators 4 and 10 have the highest weights, corresponding to transportation distance and emissions respectively, indicating that these two indicators of suppliers in the production process are not stable enough. Therefore, targeted adjustments are needed to reduce or stabilize transportation distance and emissions.

[0063] The prediction method proposed in this embodiment can help suppliers in the power grid material supply chain to predict this year's scores by providing previous green carbon emission comprehensive scores and the weights of each indicator in the green indicator system, without understanding how the power grid enterprise conducts green carbon emission evaluation during the bidding and procurement process. This can clarify the supplier's own green carbon emission level in advance. If the predicted score is at the lowest level, the supplier who should actively improve his green carbon emission level can be identified based on the ranking of indicator weights (suppliers corresponding to indicators with high weights occupy a relatively dominant position in the typical green carbon emission evaluation scores of power grid materials), so that suppliers can promptly clarify their own position in procurement, improve their own green carbon emission levels, and increase their procurement winning rate.

[0064] It can be seen from the above embodiments that at least the following substantial effects are achieved: (1) The present invention analyzes the historical data of power grid enterprises and constructs a carbon emission index evaluation system to comprehensively evaluate the carbon emission level of power grid material suppliers, and quantifies the manifestation of carbon emission levels, so that power grid material suppliers can have a clear understanding of their own carbon emission levels; (2) The present invention predicts the future comprehensive scores of power grid material suppliers by constructing a TS fuzzy neural network prediction model, and obtains problem indicators that will appear in the future, so that power grid material suppliers can optimize and improve according to the problem indicators, enhance their own supply potential, and thus increase the probability of suppliers winning the bid.

[0065] The specific implementation described above is a preferred implementation of the material supplier supply potential analysis method for carbon level prediction of the present invention, and is not intended to limit the specific implementation scope of the present invention. The scope of the present invention includes but is not limited to this specific implementation. All equivalent changes made in accordance with the shape and structure of the present invention are within the protection scope of the present invention.

Claims

1. A material supplier supply potential analysis method based on carbon level forecasting, characterized by: The following steps are involved: S1. Obtain the carbon footprint data of the enterprise's historical data and analyze it to obtain the enterprise indicator set; S2. Power grid material suppliers report their own historical supplier data, and obtain the historical comprehensive score of each supplier based on the historical supplier data and enterprise indicator set; S3, constructing a TS fuzzy neural network prediction model, using historical supplier data and enterprise indicator set as inputs of the TS fuzzy neural network prediction model to obtain predicted supplier data, and obtaining a prediction score based on the predicted supplier data and enterprise indicator set; S4, constructing a prediction region of a prediction score in the time domain and an evaluation region of a score threshold in the time domain; Segment the prediction region and / or the evaluation region based on the numerical intersection principle to obtain segmentation sub-regions, and determine the target segmentation sub-region in the segmentation sub-regions based on the numerical comparison criterion; S5. Analyze the predicted supplier data in the target segmented sub-area to obtain problem indicators, and the power grid material suppliers obtain the supply potential of the material suppliers based on the problem indicators.

2. The material supplier supply potential analysis method based on carbon level forecasting according to claim 1 is characterized by: In S1, the carbon footprint data of the enterprise's historical data is obtained and analyzed to obtain the enterprise indicator set, including the following steps: obtaining the enterprise's historical data, the enterprise's historical data at least includes the carbon footprint data of the raw and auxiliary material acquisition, raw and auxiliary material transportation, product production, product transportation and product use process; Integrate the carbon footprint data to build a carbon emission index evaluation system, and obtain the enterprise indicator set based on the carbon emission index evaluation system.

3. The material supplier supply potential analysis method based on carbon level forecasting according to claim 1 is characterized by: In S2, power grid material suppliers report their own historical supplier data, and obtain the historical comprehensive score of each supplier based on the historical supplier data and enterprise indicator set, including the following steps: Standardize the enterprise indicator set to obtain the indicator matrix and define the optimal and worst values ​​for each indicator; Analyze the historical supplier data of each power grid material supplier and each indicator in the indicator matrix in turn to obtain the indicator value corresponding to each indicator of each power grid material supplier; Based on the indicator values ​​corresponding to each indicator of each power grid material supplier, the optimal distance and the worst distance between the indicator value and the optimal value and the worst value of the corresponding indicator are obtained, and the historical comprehensive score of each supplier is obtained based on the optimal distance and the worst distance corresponding to each indicator.

4. The material supplier supply potential analysis method based on carbon level forecasting according to claim 1 is characterized by: In S3, a TS fuzzy neural network prediction model is constructed, and the historical supplier data and enterprise indicator set are used as inputs of the TS fuzzy neural network prediction model to obtain prediction scores, including the following steps: The input dimension is obtained based on the enterprise indicator set, the output dimension is obtained based on the historical comprehensive score, the number of membership functions, the center, width and coefficient of the fuzzy membership function are obtained through random initialization, and the TS fuzzy neural network prediction model is constructed based on the above data; Perform feature analysis on historical supplier data to obtain feature values, use the feature values ​​and enterprise indicator sets as input to the TS fuzzy neural network prediction model, output predicted supplier data, and obtain prediction scores based on the predicted supplier data and enterprise indicator sets.

5. The material supplier supply potential analysis method based on carbon level forecasting according to claim 1 is characterized by: In S4, constructing a prediction region of a prediction score in the time domain and an evaluation region of a score threshold in the time domain includes the following steps: A Cartesian coordinate system is established with time as the horizontal axis and score as the vertical axis. The area formed by the predicted score in the time dimension is the prediction area; the area formed by the score threshold in the time dimension is the evaluation area.

6. The material supplier supply potential analysis method based on carbon level forecasting according to claim 1 is characterized by: In S4, the prediction region and / or the evaluation region are segmented based on the numerical intersection principle to obtain segmented sub-regions, and a target segmented sub-region in the segmented sub-regions is determined based on a numerical comparison criterion, including the following steps: The prediction region and / or the evaluation region is segmented based on the intersection of the prediction score and the score threshold to obtain segmented sub-regions, and the segmented sub-regions with the prediction score less than the score threshold are taken as target segmented sub-regions.

7. The material supplier supply potential analysis method based on carbon level forecasting according to claim 6 is characterized by: The prediction region and / or the evaluation region is segmented based on the intersection of the prediction score and the score threshold to obtain segmented subregions, and the segmented subregions whose prediction scores are less than the score threshold are used as target segmented subregions, including the following steps: Get the intersection points of the predicted score and the score threshold and number the intersection points according to the time sequence; Based on the time points at which the intersections of adjacent numbers are located, the prediction region and / or the evaluation region are cut to obtain segmented sub-regions; When the predicted score in a segmentation sub-region is less than the score threshold, the corresponding segmentation sub-region is used as the target segmentation sub-region, and the target segmentation sub-region is marked by the intersection number.

8. The material supplier supply potential analysis method based on carbon level forecasting according to claim 1 is characterized by: In S5, the predicted supplier data in the target segmentation sub-region is analyzed to obtain the problem index, including the following steps: Describe the indicator characteristics of the predicted supplier data based on the indicators in the enterprise indicator set to obtain the indicator factors, standardize each indicator factor, and obtain the indicator value of each indicator factor; Determine the weight of each indicator factor based on the indicator value of each indicator factor and the overall indicator value of the overall indicator factor; The indicators corresponding to the indicator factors whose weights exceed the weight threshold are judged as problematic indicators.

9. The material supplier supply potential analysis method based on carbon level forecasting according to claim 1 is characterized by: In S5, the power grid material supplier obtains the material supplier supply potential based on the problem indicators, including the following steps: Based on the indicator value of the problem indicator, the optimal distance and the worst distance between the indicator value of the problem indicator and the optimal value and the worst value of the corresponding indicator are obtained, and the supply potential of the material supplier is obtained based on the optimal distance and the worst distance corresponding to the problem indicator.

10. The material supplier supply potential analysis method based on carbon level forecasting according to claim 2 is characterized by: The carbon emission index evaluation system at least includes: Functional unit consumables, material emissions and designed functional unit usage when obtaining raw and auxiliary materials; Transportation distance and carbon emissions of raw and auxiliary materials; Greenhouse gas emissions, fuel consumption, and electricity usage during product production; The distance travelled and the emissions generated when products are transported.

Images