A carbon footprint assessment method and apparatus

Abstract

The application relates to the technical field of environmental management, and discloses a carbon footprint evaluation method and device, which comprises the following steps: modeling the probability distribution corresponding to the preprocessed activity data and background data based on a Bayesian statistical model; sampling from the posterior distribution of the Bayesian statistical model by using a Markov chain Monte Carlo method to generate the probability distribution corresponding to the activity data; generating an input variable combination by randomly sampling in the probability distribution through target Monte Carlo simulation to perform carbon footprint calculation; obtaining an uncertainty analysis result by analyzing the distribution of the carbon footprint calculation result; analyzing key variables in the carbon footprint evaluation result by calculating the first-order sensitivity index and the total effect sensitivity index corresponding to the input variables to generate a global sensitivity analysis result; and constructing a carbon footprint evaluation model based on the uncertainty analysis result and the global sensitivity analysis result, and generating a carbon footprint evaluation result based on the carbon footprint evaluation model.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of environmental management, and particularly relates to a carbon footprint assessment method and device. BACKGROUND

[0002] Carbon footprint assessment is a tool to measure the total amount of greenhouse gas emissions in the life cycle of a product, organization or activity. Carbon footprint assessment is widely used in environmental protection, policy making, enterprise management and public awareness raising, and accurate carbon footprint assessment is of great significance for formulating emission reduction strategies and environmental policies.

[0003] Traditional carbon footprint assessment methods usually include the following steps: 1. boundary setting: determining the boundary of the assessment object, including time boundary, space boundary and process boundary; 2. data collection: collecting activity data, emission factors and background data required for assessment; 3. carbon emission calculation: based on the collected data, using relevant formulas to calculate the amount of greenhouse gas emissions; 4. result analysis: analyzing the assessment results to identify the main sources of carbon emissions and influencing factors. However, carbon footprint assessment involves a large amount of input data and model parameters, which have uncertainties that may affect the accuracy and reliability of the assessment results.

[0004] Therefore, how to provide a real-time carbon footprint assessment method to effectively analyze uncertainty and sensitivity has become a technical problem to be solved in carbon footprint assessment. SUMMARY

[0005] Embodiments of the present application provide a carbon footprint assessment method, a carbon footprint assessment device and a computer storage medium, to solve the problem that current carbon footprint assessment cannot analyze uncertainty and sensitivity.

[0006] In a first aspect of the embodiments of the present application, a carbon footprint assessment method is provided, comprising:

[0007] Real-time collection of activity data and background data corresponding to carbon footprint activities of a target object, and preprocessing of the activity data and the background data;

[0008] Modeling the probability distribution corresponding to the activity data and the background data based on a Bayesian statistical model, sampling from the posterior distribution of the Bayesian statistical model using a Markov chain Monte Carlo method to generate the probability distribution corresponding to the activity data, randomly sampling in the probability distribution through target Monte Carlo simulation to generate input variable combinations for carbon footprint calculation, and obtaining uncertainty analysis results by analyzing the distribution of carbon footprint calculation results;

[0009] The first analysis module is configured to model probability distributions corresponding to the activity data and the background data based on a Bayesian statistical model, sample from a posterior distribution of the Bayesian statistical model using a Markov Chain Monte Carlo method to generate a probability distribution corresponding to the activity data, generate input variable combinations for carbon footprint calculation by randomly sampling in the probability distribution through target Monte Carlo simulation, and obtain an uncertainty analysis result by analyzing a distribution of the carbon footprint calculation result.

[0010] The generating module is configured to construct a carbon footprint assessment model based on the uncertainty analysis result and the global sensitivity analysis result, and generate a carbon footprint assessment result based on the carbon footprint assessment model.

[0011] In a second aspect of the embodiments of the present application, a carbon footprint assessment device is provided, which includes:

[0012] The preprocessing module is configured to collect activity data and background data corresponding to carbon footprint activities of a target object in real time, and preprocess the activity data and the background data.

[0013] The first analysis module is configured to model probability distributions corresponding to the activity data and the background data based on a Bayesian statistical model, sample from a posterior distribution of the Bayesian statistical model using a Markov Chain Monte Carlo method to generate a probability distribution corresponding to the activity data, generate input variable combinations for carbon footprint calculation by randomly sampling in the probability distribution through target Monte Carlo simulation, and obtain an uncertainty analysis result by analyzing a distribution of the carbon footprint calculation result.

[0014] The second analysis module is configured to analyze key variables in the carbon footprint assessment result by calculating first-order sensitivity indices and total effect sensitivity indices corresponding to the input variables, and generate a global sensitivity analysis result, wherein the key variables are activity data and background data that have an impact on the assessment result.

[0015] The generating module is configured to construct a carbon footprint assessment model based on the uncertainty analysis result and the global sensitivity analysis result, and generate a carbon footprint assessment result based on the carbon footprint assessment model.

[0016] According to a third aspect of the embodiments of the present application, a computer readable storage medium is provided, which stores computer executable instructions. When the instructions are executed by a processor, the steps of the carbon footprint assessment method described above are implemented.

[0017] The application provides a systematic and universal carbon footprint evaluation method, which can be applied to different evaluation objects and scenes and has wide application value. Specifically, first, activity data and background data corresponding to carbon footprint activities of a target object are collected in real time, and the activity data and the background data are preprocessed; then, the probability distribution corresponding to the activity data and the background data is modeled based on a Bayesian statistical model, a Markov chain Monte Carlo method is used to sample from the posterior distribution of the Bayesian statistical model to generate the probability distribution corresponding to the activity data, input variable combinations are generated by randomly sampling in the probability distribution through target Monte Carlo simulation for carbon footprint calculation, and an uncertainty analysis result is obtained by analyzing the distribution of the carbon footprint calculation result; second, a first-order sensitivity index and a total effect sensitivity index corresponding to the input variables are calculated to analyze key variables in the carbon footprint evaluation result and generate a global sensitivity analysis result, wherein the key variables are activity data and background data that have an impact on the evaluation result; and finally, a carbon footprint evaluation model is constructed based on the uncertainty analysis result and the global sensitivity analysis result, and a carbon footprint evaluation result is generated based on the carbon footprint evaluation model.

[0018] The carbon footprint method provided by the application improves Monte Carlo simulation, Bayesian statistical model, adaptive simulation technology and global sensitivity analysis, solves the problems of high calculation complexity, difficulty in processing large-scale data, insufficient accuracy and reliability of evaluation results, lack of systematicness and universality and the like in the prior art, and can be applied to different evaluation objects and scenes and has wide application value.

[0019] The above description is only a summary of the technical solutions of the application. In order to enable the technical means of the application to be more clearly understood and implemented according to the content of the specification, and in order to enable the above and other purposes, characteristics and advantages of the application to be more apparent and easy to understand, the following specific embodiments of the application are described. BRIEF DESCRIPTION OF DRAWINGS

[0020] Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The accompanying drawings are included to provide a description of the preferred embodiments and are not meant to limit the present application. Moreover, the same reference numerals in the attached drawings indicate the same or similar components. In the drawings:

[0021] Figure 1 A flowchart of a carbon footprint evaluation method provided by an embodiment of the application;

[0022] Figure 2 A structural diagram of a carbon footprint evaluation device provided by an embodiment of the application. DETAILED DESCRIPTION

[0023] Exemplary embodiments of the present disclosure will be described in greater detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it is understood that the present disclosure can be embodied in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided so that the present disclosure can be more thoroughly understood, and so that the full scope of the present disclosure is conveyed to those skilled in the art.

[0024] In traditional carbon footprint assessment methods, input data and model parameters are usually treated as deterministic values, ignoring the impact of their uncertainty on the assessment results. With the increase in data quality and quantity, uncertainty analysis and sensitivity analysis methods are gradually applied to carbon footprint assessment to improve the accuracy and reliability of the results.

[0025] Existing uncertainty and sensitivity analysis methods mainly include the following:

[0026] Monte Carlo simulation, the main implementation steps include: setting probability distribution for each input variable; generating a large number of combinations of input variables by random sampling; performing carbon footprint calculation for each combination to generate a large number of results; statistical analysis of the distribution characteristics of the results to assess uncertainty. But its shortcomings are: high computational complexity, requires a large amount of computing resources and time, especially when dealing with large-scale data, the efficiency is low.

[0027] Global sensitivity analysis, the main implementation steps include: by setting different input variable ranges, multiple simulation calculations are performed; using Sobol' index or variance decomposition techniques to assess the relative importance of each input variable to the output result; identifying key influencing factors for focused analysis and optimization. But its shortcomings are: although it can assess the importance of input variables, when dealing with complex models and large-scale data, it also faces the problem of high computational complexity.

[0028] Scenario analysis, the main implementation steps include: setting different scenarios, changing the values of input variables; performing carbon footprint calculation for each scenario to analyze the changes in the results; comparing the assessment results under different scenarios to identify sensitive variables and sources of uncertainty. But its shortcomings are: scenario analysis relies on the set scenarios, cannot fully quantify uncertainty and sensitivity, has limited quantitative analysis capability for the results.

[0029] In addition, in the Monte Carlo simulation and global sensitivity analysis method, the structural composition of the technical solution mainly includes the following parts: 1. Data input module: collect and preprocess activity data, emission factors and background data; set the probability distribution or range of input variables. 2. Simulation calculation module: generate a large number of input variable combinations through random sampling (Monte Carlo simulation) or setting variable range (global sensitivity analysis); use the carbon footprint calculation model to calculate each combination. 3. Result analysis module: statistically analyze the distribution characteristics of the simulation results, and evaluate the uncertainty (Monte Carlo simulation); use Sobol' index or variance decomposition technology to evaluate the importance of input variables (global sensitivity analysis). 4. Optimization module: based on the result analysis, identify the key influencing factors, optimize the input data and model parameters; propose improvement suggestions and optimization measures.

[0030] Based on the current Monte Carlo simulation and global sensitivity analysis method technical solution, the application proposes an improved carbon footprint evaluation method based on uncertainty and sensitivity analysis, which reduces the computational complexity, improves the accuracy and reliability of the evaluation results, and effectively processes large-scale data by introducing Bayesian statistical model and adaptive simulation technology.

[0031] Referring to Figure 1 , Figure 1 A flowchart of a carbon footprint evaluation method provided by an embodiment of the application is shown in FIG. 1. As shown in FIG. 1, the method comprises the following steps. Figure 1

[0032] Step S102: Real-time collection of activity data and background data corresponding to carbon footprint activities of a target object, and preprocessing of the activity data and the background data.

[0033] It should be noted that the activity data here can be data such as energy consumption and power usage of an enterprise, and the activity data is denoted by AD in the embodiment of the application.

[0034] The background data here can be data such as emission factors, and the background data is denoted by BD in the embodiment of the application.

[0035] ​In the embodiments of the present application, the preprocessing of the activity data and the background data includes: based on the expected value calculated based on the observed data, filling in the missing values of the activity data and / or the background data to obtain first activity data and / or first background data according to the Bayesian multiple imputation method; based on the standard deviation method, removing outliers of the activity data and / or the background data to obtain second activity data and / or second background data; obtaining activity data according to the first activity data and the second activity data, and performing standardization processing on the activity data based on the standard deviation and the mean value corresponding to the activity data; obtaining background data according to the first background data and the second background data, and performing standardization processing on the background data based on the standard deviation and the mean value corresponding to the background data.

[0036] In the carbon footprint assessment, the integrity and accuracy of the data are critical, and the present application proposes a set of data collection and preprocessing methods to ensure data quality.

[0037] Among them, the data preprocessing method proposed in the present application includes missing value filling, outlier removal and data standardization.

[0038] Specifically, for the missing values in the activity data AD and the background data BD, the Bayesian multiple imputation method is used for filling. Taking the activity data AD as an example, the specific method for filling the missing values is as follows:

[0039]

[0040] Among them, IE(ADobserved data) is the expected value calculated based on the observed data.

[0041] For the outliers in the activity data AD and the background data BD, the three standard deviation method is used to remove outliers. Taking the activity data AD as an example, the specific method for removing outliers is as follows:

[0042] AD i is an outlier if|AD i -μ AD |>3σAD;

[0043] Among them, μ AD is the mean value of the activity data; σAD is the standard deviation of the activity data.

[0044] In order to eliminate the influence between different dimensions, the data is standardized, and the method for standardizing the data includes:

[0045] Step S104: modeling the probability distribution corresponding to the activity data and the background data respectively based on a Bayesian statistical model, sampling from the posterior distribution of the Bayesian statistical model by using a Markov Chain Monte Carlo method to generate the probability distribution corresponding to the activity data, randomly sampling in the probability distribution by target Monte Carlo simulation to generate the input variable combination for carbon footprint calculation, and obtaining the uncertainty analysis result by analyzing the distribution of the carbon footprint calculation result.

[0046] It should be noted that the purpose of the uncertainty analysis is to quantify the influence of the uncertainty of the input variables and the model parameters on the carbon footprint calculation result.

[0047] In the embodiments of the present application, on the one hand, a Bayesian statistical model is introduced to model the probability distribution of the input data and the emission factors; on the other hand, a Markov Chain Monte Carlo method is used to sample from the posterior distribution to generate the probability distribution of the input variables; and on the other hand, a large number of input variable combinations are generated by improved Monte Carlo simulation for carbon footprint calculation, and the distribution characteristics of the statistical analysis result are analyzed to comprehensively evaluate the uncertainty.

[0048] Specifically, the probability distribution corresponding to the activity data and the background data is modeled based on a Bayesian statistical model, wherein the Bayesian statistical model comprises:

[0049] π(θX)∝π(θ)·L(Xθ)

[0050] Wherein, θ is a model parameter (here, the model parameter includes the probability distribution parameters of the activity data and the emission factors); π(θ) is a prior distribution (i.e. the initial guess of the model parameter); L(X|θ) is a likelihood function (i.e. the probability of the data given the model parameter); and π(θ|X) is a posterior distribution (i.e. the parameter distribution given the data).

[0051] Modeling the probability distribution of the input data and the emission factors by using the Bayesian statistical model can improve the accuracy of the input data.

[0052] The Markov Chain Monte Carlo method (i.e. MCMC) is used to sample from the posterior distribution of the Bayesian statistical model to generate the probability distribution corresponding to the activity data, wherein the Markov Chain Monte Carlo method comprises:

[0053]

[0054] Wherein, θ (t) is the parameter value of the tth iteration; ∈ is a step size; ▽logπ(θ (t) X) is the gradient of the posterior distribution; and η is a random variable of a standard normal distribution.

[0055] The accuracy of the uncertainty analysis can be enhanced by sampling from the posterior distribution through the MCMC method to generate the probability distribution of the input variables.

[0056] The input variable combination is generated by randomly sampling in the probability distribution through the target Monte Carlo simulation to perform the carbon footprint calculation, wherein the formula of the carbon footprint calculation comprises:

[0057]

[0058] wherein CF i is the carbon footprint result of the i-th simulation; AD ij is the j-th activity data in the i-th simulation; and EF j is the j-th background data.

[0059] The calculation complexity is reduced and the evaluation efficiency is significantly improved by improving the Monte Carlo simulation method, combining the Bayesian statistical model and the adaptive simulation technology; the accuracy and reliability of the evaluation result are improved by introducing the uncertainty analysis to quantify and evaluate the uncertainty of the input data and the model parameters.

[0060] Step S106: The key variables in the carbon footprint evaluation result are analyzed by calculating the first-order sensitivity index and the total effect sensitivity index corresponding to the input variables, and the global sensitivity analysis result is generated, wherein the key variables are the activity data and the background data that have an impact on the evaluation result.

[0061] It should be noted that the global sensitivity analysis is used to identify the key variables that have the greatest impact on the carbon footprint evaluation result.

[0062] In the embodiments of the present application, on the one hand, Sobol' index and variance decomposition technology are used to quantify the relative importance of the input variables to the output result, analyze the key variables in the carbon footprint evaluation, and identify the key variables that have the greatest impact on the evaluation result; on the other hand, the global sensitivity analysis is used to identify the key variables that have the greatest impact on the evaluation result, and the model is optimized based on the key variables.

[0063] Specifically, the first-order sensitivity index corresponding to the input variables is calculated based on the first-order sensitivity index calculation formula, wherein the first-order sensitivity index calculation formula comprises:

[0064]

[0065] wherein S i is the first-order sensitivity index of the input variable X i ; V(Y) is the total variance of the output result Y; and V(X i ) is the variance contribution of the input variable X .

[0066] calculate a total effect sensitivity index corresponding to the input variable based on a total effect sensitivity index calculation formula, wherein the total effect sensitivity index calculation formula comprises:

[0067]

[0068] wherein S Ti is a total effect sensitivity index of an input variable X i ; V(Y) is a total variance of an output result Y; is a variance contribution of other variables except Xi.

[0069] Based on the first-order sensitivity index and the total effect sensitivity index, analyze the key variables in the carbon footprint assessment result, and generate a global sensitivity analysis result by analyzing the initial sample corresponding to the input variable and the perturbed sample corresponding to the key variable.

[0070] Specifically, the Sobol' sequence is used to generate sampling points X and X', and the first-order sensitivity index Y and the total effect sensitivity index Y' are calculated, wherein Y=f(X); Y'=f(X'), wherein X is an initial sample of an input variable; X' is a perturbed sample of the input variable.

[0071] Step S108: Based on the uncertainty analysis result and the global sensitivity analysis result, a carbon footprint assessment model is constructed, and based on the carbon footprint assessment model, a carbon footprint assessment result is generated.

[0072] Based on the uncertainty analysis and global sensitivity analysis results, the carbon footprint assessment model is optimized, and the accuracy and reliability of the evaluation are provided.

[0073] First, the probability distribution and interaction of the key variables are introduced, and the carbon footprint calculation formula is optimized, and the probability distribution and interaction of the key variables are introduced.

[0074] Specifically, based on the probability distribution and mutual relationship of the key variables generated in the carbon footprint assessment result, the carbon footprint calculation formula is optimized, wherein the optimized carbon footprint calculation formula comprises:

[0075]

[0076] wherein CF is the carbon footprint; AD j is the jth activity data; EF j is the jth background data; AD k is the kth activity data; EF i is the ith background data; β kl is the interaction coefficient between the variables AD k and EF l .

[0077] Secondly, the simulation parameters and sampling strategies are dynamically adjusted by using an adaptive step algorithm to improve the calculation efficiency.

[0078] Specifically, the simulation parameters and sampling strategies of the optimized carbon footprint calculation formula are adjusted by using an adaptive step algorithm to generate a carbon footprint evaluation result, wherein the adaptive step algorithm comprises:

[0079] Δt i+1 = Δt i + η · (ε i - ε target )

[0080] wherein Δt i is the step size of the i-th iteration; η is the learning rate; ε i is the current error; and ε target is the target error.

[0081] In another embodiment, the method further comprises generating improvement measures by analyzing the evaluation result and the original evaluation result. That is, based on the carbon footprint evaluation result, specific measures and decisions for reducing carbon emissions are removed, and the carbon footprint evaluation method is applied to actual cases of enterprises. By comparing the carbon emission data before and after the application, the effect of the evaluation method is evaluated.

[0082] Taking the carbon footprint evaluation of an enterprise as an example, activity data AD of the enterprise, such as energy consumption and power usage, is collected, and missing value filling, outlier removal and data standardization processing are performed; Bayesian statistical modeling is performed on the activity data and emission factors of the enterprise, and MCMC method is used to sample from the posterior distribution to generate the probability distribution of the input variables, and carbon footprint calculation and uncertainty quantification are performed; Sobol' index is used to analyze the key variables in the enterprise carbon footprint evaluation, and the activity data and emission factors that have the greatest impact on the evaluation result are identified; based on the uncertainty and sensitivity analysis results, the carbon footprint evaluation model is optimized, the probability distribution and interaction of the key variables are introduced, the simulation parameters and sampling strategies are adjusted, and the accuracy and reliability of the evaluation are improved; based on the optimized carbon footprint evaluation result, specific measures and decision suggestions for reducing carbon emissions are proposed to help enterprises develop practical emission reduction schemes.

[0083] By using the carbon footprint method of the embodiments of the present application, the problems of high calculation complexity, difficulty in processing large-scale data, insufficient accuracy and reliability of the evaluation result, lack of systematicness and universality in the prior art are solved by using the improved Monte Carlo simulation, Bayesian statistical model, adaptive simulation technology and global sensitivity analysis. In addition, the method can be applied to different evaluation objects and scenes, and has wide application value.

[0084] Corresponding to the method embodiments described above, the specification also provides a carbon footprint assessment device embodiment, Figure 2 A structural schematic diagram of a carbon footprint assessment device provided by the embodiments of the present application is shown in FIG. 2. As shown in the figure, the device 200 includes: Figure 2

[0085] A preprocessing module 202 configured to collect activity data and background data corresponding to carbon footprint activities of a target object in real time, and preprocess the activity data and the background data;

[0086] A first analysis module 204 configured to model probability distributions corresponding to the activity data and the background data based on a Bayesian statistical model, sample from a posterior distribution of the Bayesian statistical model using a Markov chain Monte Carlo method to generate a probability distribution corresponding to the activity data, randomly sample in the probability distribution through target Monte Carlo simulation to generate an input variable combination for carbon footprint calculation, and obtain an uncertainty analysis result by analyzing a distribution of carbon footprint calculation results;

[0087] A second analysis module 206 configured to analyze key variables in the carbon footprint assessment result by calculating a first-order sensitivity index and a total effect sensitivity index corresponding to the input variables, and generate a global sensitivity analysis result, wherein the key variables are activity data and background data that have an impact on the assessment result;

[0088] A generation module 208 configured to construct a carbon footprint assessment model based on the uncertainty analysis result and the global sensitivity analysis result, and generate a carbon footprint assessment result based on the carbon footprint assessment model.

[0089] In an optional embodiment, the preprocessing module 202 is further configured to:

[0090] According to a Bayesian multiple imputation method, missing value imputation is performed on the activity data and / or the background data based on an expected value calculated based on observed data to obtain first activity data and / or first background data;

[0091] Based on a standard deviation method, outlier rejection is performed on the activity data and / or the background data to obtain second activity data and / or second background data;

[0092] Based on the first activity data and the second activity data, activity data is obtained, and the activity data is standardized based on a standard deviation and a mean value corresponding to the activity data;

[0093] ​According to the first background data and the second background data, background data is obtained, and the background data is normalized based on a standard deviation and a mean value corresponding to the background data.

[0094] In an optional embodiment, the first analysis module 204 is further configured to:

[0095] modeling probability distributions corresponding to the activity data and the background data respectively based on a Bayesian statistical model, wherein the Bayesian statistical model comprises:

[0096] π(θX)∝π(θ)·L(Xθ)

[0097] wherein θ is a model parameter; π(θ) is a prior distribution; L(X|θ) is a likelihood function; and π(θ|X) is a posterior distribution;

[0098] sampling from the posterior distribution of the Bayesian statistical model to generate the probability distribution corresponding to the activity data by using a Markov chain Monte Carlo method, wherein the Markov chain Monte Carlo method comprises:

[0099]

[0100] wherein θ (t) is a parameter value of the tth iteration; ∈ is a step size; and ▽logπ(θ (t) X) is a gradient of the posterior distribution; and η is a random variable of a standard normal distribution.

[0101] randomly sampling in the probability distribution by target Monte Carlo simulation to generate combinations of input variables for carbon footprint calculation, wherein a calculation formula of the carbon footprint comprises:

[0102]

[0103] wherein CF i is a carbon footprint result of the ith simulation; AD ij is the jth activity data in the ith simulation; and EF j is the jth background data.

[0104] In an optional embodiment, the second analysis module 206 is further configured to:

[0105] calculating a first-order sensitivity index corresponding to the input variable based on a first-order sensitivity index calculation formula, wherein the first-order sensitivity index calculation formula comprises:

[0106]

[0107] wherein V(Y) is a total variance of an output result Y. is the variance contribution of the input variable X i ;

[0108] Based on the total effect sensitivity index calculation formula, the total effect sensitivity index corresponding to the input variable is calculated, wherein the total effect sensitivity index calculation formula comprises:

[0109]

[0110] wherein V(Y) is the total variance of the output result Y; is the variance contribution of other variables except Xi

[0111] Based on the first-order sensitivity index and the total effect sensitivity index, the key variables in the carbon footprint assessment result are analyzed, and the global sensitivity analysis result is generated by analyzing the initial sample corresponding to the input variable and the perturbed sample corresponding to the key variable.

[0112] In an optional embodiment, the generation module 208 is further configured to:

[0113] Based on the probability distribution and the mutual relationship corresponding to the key variables generated in the carbon footprint assessment result, the carbon footprint calculation formula is optimized, wherein the optimized carbon footprint calculation formula comprises:

[0114]

[0115] wherein CF is the carbon footprint; AD j is the jth activity data; EF j is the jth background data; AD k is the kth activity data; EF i is the ith background data; β kl is the interaction coefficient between the variables AD k and EF l ;

[0116] The simulation parameters and the sampling strategy of the optimized carbon footprint calculation formula are adjusted by using an adaptive step algorithm to generate a carbon footprint assessment result, wherein the adaptive step algorithm comprises:

[0117] Δt i+1 = Δt i + η·(∈ i - ∈ target )

[0118] wherein Δt i is the step length of the ith iteration; η is the learning rate; ∈ i is the current error; ∈ target is the target error.

[0119] In an optional embodiment, the device 200 is further configured to:

[0120] The improvement module 210 is configured to generate improvement measures by analyzing the evaluation result and the original evaluation result.

[0121] The carbon footprint device provided by the embodiments of the present application solves the problems of high computational complexity, difficulty in processing large-scale data, insufficient accuracy and reliability of evaluation results, lack of systematicness and universality, and the like in the prior art by using improved Monte Carlo simulation, Bayesian statistical model, adaptive simulation technology, and global sensitivity analysis. In addition, the method can be applied to different evaluation objects and scenarios, and has wide application value.

[0122] Each of the embodiments in the specification is described in a progressive manner, and the same or similar parts of each of the embodiments can be referred to each other. Each of the embodiments mainly describes the differences from other embodiments. In particular, for the carbon footprint evaluation device, since it is basically similar to the carbon footprint evaluation method embodiment, the description is relatively simple, and the relevant parts can be referred to the part of the description of the carbon footprint evaluation method embodiment.

[0123] An embodiment of the specification also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the above-mentioned carbon footprint evaluation method.

[0124] It should be noted that the above describes specific embodiments of the specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in an order different from that in the embodiments and still achieve the desired result. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are possible or can be advantageous. Secondly, those skilled in the art should know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to the embodiments of the specification.

[0125] In the above embodiments, the description of each embodiment has its own focus, and the parts not described in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0126] The preferred embodiments of the present specification disclosed above are only used to help explain the present specification. Alternative embodiments do not describe all the details and limit the present application to the specific embodiments described. Obviously, many modifications and changes can be made according to the content of the embodiments of the present specification. The present specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the embodiments of the present specification, so that those skilled in the art can well understand and utilize the present specification. The present specification is limited only by the claims and their full scope and equivalents.

Claims

1. A carbon footprint assessment method characterized by, The method comprises the following steps: real-time acquisition of activity data and background data corresponding to carbon footprint activities of a target object, and preprocessing of the activity data and the background data; modeling of probability distributions corresponding to the activity data and the background data based on a Bayesian statistical model, sampling of the activity data from a posterior distribution of the Bayesian statistical model by using a Markov chain Monte Carlo method, generation of a probability distribution corresponding to the activity data, random sampling in the probability distribution by target Monte Carlo simulation, generation of an input variable combination for carbon footprint calculation, analysis of a distribution of carbon footprint calculation results, and obtaining of an uncertainty analysis result; calculation of a first-order sensitivity index and a total effect sensitivity index corresponding to the input variables, analysis of key variables in a carbon footprint evaluation result, and generation of a global sensitivity analysis result, wherein the key variables are activity data and background data that have an influence on the evaluation result; construction of a carbon footprint evaluation model based on the uncertainty analysis result and the global sensitivity analysis result, and generation of a carbon footprint evaluation result based on the carbon footprint evaluation model; the modeling of probability distributions corresponding to the activity data and the background data based on the Bayesian statistical model, the sampling of the activity data from the posterior distribution of the Bayesian statistical model by using the Markov chain Monte Carlo method, the generation of the probability distribution corresponding to the activity data, and the random sampling in the probability distribution by the target Monte Carlo simulation to generate the input variable combination for carbon footprint calculation comprise: modeling of probability distributions corresponding to the activity data and the background data based on a Bayesian statistical model, wherein the Bayesian statistical model comprises: π(θ|X)∝π(θ)·L(X|θ) wherein θ is a model parameter, π(θ) is a prior distribution, L(X|θ) is a likelihood function, and π(θ|X) is a posterior distribution; sampling of the activity data from the posterior distribution of the Bayesian statistical model by using a Markov chain Monte Carlo method to generate a probability distribution corresponding to the activity data, wherein the Markov chain Monte Carlo method comprises: where θ (t) is the parameter value at the tth iteration; ∈ is the step size; is the gradient of the posterior distribution; η' is a random variable from a standard normal distribution; random sampling in the probability distribution by target Monte Carlo simulation to generate an input variable combination for carbon footprint calculation, wherein a calculation formula of carbon footprint comprises: where CF i is the carbon footprint result of the i-th simulation; AD ij is the j-th activity data in the i-th simulation; EF j is the j-th background data.

2. The method of claim 1, wherein, the preprocessing of the activity data and the background data comprises: filling of missing values in the activity data and / or the background data based on an expected value calculated based on observed data according to a Bayesian multiple imputation method, to obtain first activity data and / or first background data; removal of outliers in the activity data and / or the background data based on a standard deviation method, to obtain second activity data and / or second background data; obtaining of activity data from the first activity data and the second activity data, and standardization of the activity data based on a standard deviation and a mean value corresponding to the activity data; obtaining of background data from the first background data and the second background data, and standardization of the background data based on a standard deviation and a mean value corresponding to the background data.

3. The method of claim 1, wherein, The first-order sensitivity index corresponding to the input variable is calculated based on a first-order sensitivity index calculation formula, wherein the first-order sensitivity index calculation formula comprises: The total effect sensitivity index corresponding to the input variable is calculated based on a total effect sensitivity index calculation formula, wherein the total effect sensitivity index calculation formula comprises: where V(Y) is the total variance of the output Y; is the variance contribution of the input variable X i . Based on the first-order sensitivity index and the total effect sensitivity index, the key variables in the carbon footprint evaluation result are analyzed, and a global sensitivity analysis result is generated by analyzing the initial sample corresponding to the input variable and the perturbed sample corresponding to the key variable. where V(Y) is the total variance of the output result Y; is the variance contribution of other variables except X i ​ The carbon footprint evaluation model is constructed based on the uncertainty analysis result and the global sensitivity analysis result, and the carbon footprint evaluation result is generated based on the carbon footprint evaluation model, comprising:

4. The method of claim 1, wherein, Based on the probability distribution and the mutual relationship of the key variables corresponding to the carbon footprint evaluation result, the carbon footprint calculation formula is optimized, wherein the optimized carbon footprint calculation formula comprises: The simulation parameters and the sampling strategy of the optimized carbon footprint calculation formula are adjusted by using an adaptive step length algorithm to generate the carbon footprint evaluation result, wherein the adaptive step length algorithm comprises: where CF is the carbon footprint; AD j is the jth activity data; EF j is the jth background data; AD k is the kth activity data; EF i is the ith background data; β kl is the variable AD k and EF i is the interaction coefficient between AD and EF. The method further comprises: Δt i+1 = Δt i + η · (ε i - ε target ) where Δt i is the step size for the i-th iteration; η is the learning rate; ∈ i is the current error; ∈ target is the target error.

5. The method of claim 1, wherein, Improved measures are generated by analyzing the evaluation result and the original evaluation result. Comprise:

6. A carbon footprint assessment apparatus characterized by, The preprocessing module is configured to collect activity data and background data corresponding to carbon footprint activities of a target object in real time, and preprocess the activity data and the background data; The first analysis module is configured to model the probability distribution corresponding to the activity data and the background data based on a Bayesian statistical model, sample from the posterior distribution of the Bayesian statistical model by using a Markov chain Monte Carlo method, generate the probability distribution corresponding to the activity data, randomly sample in the probability distribution by target Monte Carlo simulation to generate input variable combinations for carbon footprint calculation, and obtain an uncertainty analysis result by analyzing the distribution of the carbon footprint calculation result; The second analysis module is configured to analyze key variables in the carbon footprint evaluation result by calculating the first-order sensitivity index and the total effect sensitivity index corresponding to the input variable, and generate a global sensitivity analysis result, wherein the key variables are activity data and background data that have an impact on the evaluation result; The generation module is configured to construct a carbon footprint evaluation model based on the uncertainty analysis result and the global sensitivity analysis result, and generate a carbon footprint evaluation result based on the carbon footprint evaluation model; The probability distribution corresponding to the activity data and the background data is modeled based on a Bayesian statistical model, wherein the Bayesian statistical model comprises: π(θ|X)∝π(θ)·L(X|θ) Wherein, θ is a model parameter; π(θ) is a prior distribution; L(X|θ) is a likelihood function; π(θ|X) is a posterior distribution; ​ sampling from a posterior distribution of the Bayesian statistical model using a Markov Chain Monte Carlo method to generate a probability distribution corresponding to the activity data, wherein the Markov Chain Monte Carlo method comprises: where θ (t) is the parameter value at the tth iteration; ∈ is the step size; is the gradient of the posterior distribution; η' is a random variable from a standard normal distribution; randomly sampling in the probability distribution to generate combinations of input variables for carbon footprint calculation by target Monte Carlo simulation, wherein the formula for carbon footprint calculation comprises: where CF i is the carbon footprint result of the i-th simulation; AD ij is the j-th activity data in the i-th simulation; EF j is the j-th background data.

7. A computer-readable storage medium, characterized in that, The computer readable storage medium stores an information transmission implementation program, and the program is executed by the processor to implement the steps of the method in any one of claims 1-5.

Images