A warehouse carbon emission factor dynamic calibration and adaptive allocation method and system

Abstract

The present application relates to the technical field of carbon footprint accounting and low-carbon management, and specifically provides a warehouse carbon emission factor dynamic calibration and adaptive allocation method and system, which comprises the following steps: collecting multi-dimensional data affecting warehouse carbon emission in a target warehouse in real time and quantifying the data to obtain a dynamic warehouse feature vector; generating an initial carbon emission factor from a pre-constructed carbon emission base factor library, and performing Bayesian calibration on the initial carbon emission factor to obtain a precise carbon emission factor; allocating carbon emission at the cargo level according to the data availability and business scenarios of the target warehouse; and then updating the precise carbon emission factor to the carbon emission base factor library based on the preset update trigger condition. The present application can realize scenario-based precise generation, dynamic calibration and cargo-level adaptive allocation of warehouse carbon emission factors, and significantly improve the accuracy and flexibility of warehouse carbon accounting.

Description

Technical Field

[0001] This invention belongs to the field of carbon footprint accounting and low-carbon management technology, specifically involving a method and system for dynamic calibration and adaptive allocation of carbon emission factors in warehousing, which can be applied to accurate carbon footprint accounting and refined carbon management scenarios in industries such as power, logistics, retail and manufacturing. Background Technology

[0002] With the ongoing pursuit of "dual carbon" goals, accurate carbon footprint accounting for warehousing, a crucial link in the supply chain, has become a core requirement for enterprises undergoing green transformation. However, existing warehousing carbon emission accounting technologies have significant limitations in addressing the demands for refined and dynamic management.

[0003] Specifically, existing technologies generally adopt static carbon emission factors based on industry averages. This generic approach ignores the non-linear impact of multiple dimensions of characteristics on actual carbon emissions, such as building structure (e.g., the difference between single-story and multi-story warehouses), equipment energy efficiency, operating mode (automation or manual), and cargo storage characteristics (e.g., refrigerated or hazardous chemicals). This leads to a discrepancy of 30% to 50% between the calculated results and actual emissions, severely weakening the data's decision-making value. Furthermore, the lack of effective automatic factor calibration and iteration mechanisms relies primarily on periodic manual surveys to update the factor database, resulting in long cycles, high costs, and an inability to adjust factors in real time based on dynamic changes such as warehouse equipment upgrades and operational optimizations, causing the calculated results to lag behind reality.

[0004] Meanwhile, at the carbon emission allocation level, existing systems mostly use a single dimension such as weight or volume for allocation, which cannot flexibly adapt to the data availability of different warehouses (for example, some warehouses lack accurate weight measurement data), nor does it consider the impact of storage time and cargo turnover rate on the allocation results, resulting in the distortion of cargo-level carbon emission accounting.

[0005] Especially with the continuous emergence of new warehousing scenarios such as zero-carbon warehouses and integrated photovoltaic storage, existing technologies have failed to construct targeted factor accounting methods, making it difficult to accurately quantify the carbon emission reduction benefits brought about by self-generated and self-consumed green electricity consumption and energy storage peak shaving.

[0006] Although existing logistics carbon footprint accounting systems and green supply chain management platforms have basic warehousing carbon accounting functions, their underlying logic treats carbon emission factors as constant preset input parameters, failing to fundamentally solve the problems of accurate matching, dynamic updating, and adaptive allocation of the factors themselves. This results in existing technologies being unable to meet the development needs of refined and scenario-based warehousing carbon accounting. Summary of the Invention

[0007] This invention aims to address the problems of existing static and universal warehousing carbon emission factors, single allocation methods, lack of automatic calibration, and poor adaptability to new scenarios. It provides a dynamic calibration and adaptive allocation method and system for warehousing carbon emission factors, enabling scenario-based accurate generation, real-time dynamic calibration, and cargo-level adaptive allocation of warehousing carbon emission factors, significantly improving the accuracy and flexibility of warehousing carbon accounting.

[0008] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0009] In a first aspect, the present invention provides a method for dynamic calibration and adaptive allocation of carbon emission factors in warehousing, comprising:

[0010] Real-time collection and quantification of multi-dimensional data affecting carbon emissions in the target warehouse to obtain a dynamic warehousing feature vector of the target warehouse; wherein, the warehousing feature vector includes building features, equipment features, operational features, cargo features and energy features;

[0011] Based on the storage feature vector of the target warehouse, an initial carbon emission factor is generated by matching from a pre-constructed carbon emission basic factor library; wherein, each factor entry in the carbon emission basic factor library includes a factor value, applicable feature range, data source, DQI data quality score, and update time;

[0012] The data quality of the energy data of the target warehouse is quantitatively scored; the initial carbon emission factor is dynamically calibrated using a Bayesian calibration algorithm to generate an accurate carbon emission factor; wherein, when the target warehouse is a zero-carbon warehouse, the carbon emissions corresponding to the self-generated and self-consumed green electricity of the zero-carbon warehouse are deducted during the execution of the Bayesian calibration algorithm.

[0013] Initiate the allocation decision, and based on the data availability and business scenario of the target warehouse, select the optimal carbon emission allocation method from multiple allocation methods to generate cargo-level carbon emission allocation coefficients;

[0014] By comparing the precise carbon emission factor with the initial carbon emission factor, and based on meeting the preset update trigger conditions, the precise carbon emission factor is updated to the carbon emission basic factor library.

[0015] A further approach: Generating initial carbon emission factors by matching from the aforementioned carbon emission basic factor library includes:

[0016] An improved cosine similarity algorithm is used to calculate the similarity between the feature terms in the storage feature vector of the target warehouse and the applicable feature terms in each factor entry of the carbon emission basic factor library. The formula is as follows:

[0017] ;

[0018] Wherein, V represents the storage feature vector of the target warehouse; n represents the total number of feature terms in the storage feature vector; This represents the quantized score of the k-th feature term in the warehouse feature vector; This represents the weight of the k-th feature term in the warehouse feature vector; This represents the i-th factor in the aforementioned carbon emission basic factor library; This represents the quantized score of the k-th feature term in the entry for the i-th factor;

[0019] The calculated similarities are sorted from highest to lowest; the top few basic factors in the similarity ranking are selected and the initial carbon emission factor of the target warehouse is generated by weighted averaging.

[0020] A further approach: Dynamically calibrating the initial carbon emission factor includes:

[0021] The energy data of the target warehouse is scored based on four dimensions: activity data credibility, factor data credibility, time representativeness, and geographical representativeness, to obtain the data quality score of the energy data.

[0022] Based on the data quality score, determine the data quality weight corresponding to the energy data;

[0023] Using the initial carbon emission factor as the prior distribution and the daily carbon emission per unit area as the likelihood function, Bayesian calibration is performed in conjunction with the data quality weights to obtain the mean of the posterior distribution, which is then used as the calibrated precise carbon emission factor.

[0024] A further proposed solution: The formula for calculating the daily carbon emissions per unit area is as follows:

[0025] ;

[0026] Among them, the This represents the daily carbon emissions per unit area; This represents the consumption of the i-th type of energy; the... The carbon emission factor of the i-th energy source is represented; S represents the warehouse floor area; and D represents the number of days in a year.

[0027] The formula used for the Bayesian calibration is:

[0028] ;

[0029] Among them, the Indicates precise carbon emission factors; the Indicates data quality weights; the This represents the initial carbon emission factor.

[0030] A further solution: When the target warehouse is a zero-carbon warehouse equipped with photovoltaic or energy storage systems, the carbon emissions corresponding to self-generated and self-consumed green electricity are deducted when calculating the daily carbon emissions per unit area.

[0031] A further solution: The process for the cost-sharing decision includes:

[0032] The available data types in the target warehouse are detected, and multiple usable allocation methods are generated.

[0033] Based on the business scenario of the target warehouse and the DQI data quality score of the precise carbon emission factor, the available allocation methods are ranked by priority, the allocation method with the highest priority is selected, and the carbon emission allocation coefficient for each item is calculated.

[0034] The generated allocation methods include: allocation by floor area, allocation by weight, allocation by volume, allocation by value, and weighted allocation by storage duration.

[0035] A further proposed solution is as follows: the preset update trigger condition is that the deviation between the accurate carbon emission factor and the initial carbon emission factor is greater than 15%, and the data quality score of the energy data of the target warehouse is greater than or equal to 2.

[0036] A further solution: In the aforementioned warehouse feature vector,

[0037] The architectural features include building structure, building area, thermal insulation performance, and lighting design;

[0038] The equipment features include the type, power, energy efficiency rating, and quantity of equipment used for heating, cooling, ventilation, lighting, and handling;

[0039] The operational characteristics include storage methods, work processes, staffing, and inventory turnover rate;

[0040] The cargo characteristics include cargo type, temperature requirements, packaging form, and average storage time;

[0041] The energy characteristics include consumption data for electric energy, natural gas energy, and diesel energy, the proportion of green electricity self-generation and self-consumption, and energy storage charging and discharging data;

[0042] The expression for the storage feature vector of the target warehouse is:

[0043] ;

[0044] Wherein, V represents the warehouse feature vector; n is a label index, representing the nth feature term; This represents the quantized score corresponding to the nth feature item.

[0045] A further solution: The carbon emission basic factor library adopts a multi-layer architecture; wherein, the first layer is the industry general factor layer, which stores the average warehousing emission factors of different industries; the second layer is the warehouse type factor layer, which stores the basic factors of different types of warehouses; the third layer is the sub-scenario factor layer, which stores the scenario factors under different equipment configurations and different operating modes.

[0046] Secondly, the present invention provides a dynamic calibration and adaptive allocation system for carbon emission factors in warehousing, comprising:

[0047] The warehouse multi-dimensional feature acquisition module is used to collect and quantify multi-dimensional data affecting warehouse carbon emissions in real time, and build a dynamic warehouse feature vector library for the target warehouse.

[0048] A carbon emission basic factor library is used to store basic carbon emission factors for warehousing at different levels and of different types.

[0049] The initial carbon emission factor matching and generation module is used to match and generate initial carbon emission factors from the carbon emission basic factor library based on the warehouse feature vector of the target warehouse.

[0050] The calibration module is used to dynamically calibrate the initial carbon emission factor based on the energy data of the target warehouse and the corresponding data quality evaluation results, so as to generate a scenario-based accurate carbon emission factor.

[0051] The adaptive allocation decision module is used to select the optimal carbon emission allocation method from multiple allocation methods based on the data availability of the target warehouse and the business scenario, and generate cargo-level carbon emission allocation coefficients.

[0052] The factor iteration update and sharing module is used to update the precise carbon emission factor to the carbon emission basic factor library based on the satisfaction of preset update trigger conditions; the factor iteration update and sharing module is also used to provide a sharing interface to support upstream and downstream enterprises in the supply chain to share the precise carbon emission factor.

[0053] In addition, a visualization interface module is provided to offer visualization functionality; the visualization interface module is also used to provide an interface for interfacing with the carbon accounting platform.

[0054] The beneficial effects of this invention are as follows:

[0055] 1. Significantly improves the accuracy of carbon emission factors in warehousing: Initial carbon emission factors are generated through multi-dimensional feature matching, and Bayesian dynamic calibration is performed by combining real-time collected energy data and corresponding data quality evaluation to obtain accurate carbon emission factors after calibration. This solves the problem of large errors in static general factors and improves the accuracy of carbon accounting.

[0056] 2. Achieve adaptive carbon emission allocation: Automatically select the optimal allocation method based on data availability and business scenario, solving the problem of poor applicability of a single allocation method, especially suitable for warehouse scenarios without cargo weight data.

[0057] 3. Establish an automatic factor iteration and update mechanism: The calibrated accurate carbon emission factors are automatically fed back to the carbon emission basic factor library to achieve continuous factor optimization without the need for regular manual surveys, which greatly reduces maintenance costs.

[0058] 4. Fully adaptable to new warehousing scenarios: A special factor accounting method has been designed for new scenarios such as zero-carbon warehouses to accurately deduct the carbon emission reduction brought about by green electricity self-generation and self-consumption and energy storage peak shaving.

[0059] 5. Excellent compatibility and scalability: The dynamic calibration and adaptive allocation system for carbon emission factors in warehousing of this invention can operate as an independent system and is equipped with an interface, enabling the system to connect with existing carbon accounting platforms and green supply chain management systems, providing warehousing carbon accounting support for enterprises of different industries and sizes. Attached Figure Description

[0060] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0061] Figure 1 This is a flowchart of a dynamic calibration and adaptive allocation method for carbon emission factors in warehousing according to the present invention;

[0062] Figure 2 This is a flowchart of the steps for matching and generating the initial carbon emission factor in this invention;

[0063] Figure 3 This is a flowchart of the steps for dynamically calibrating the initial carbon emission factor in this invention. Detailed Implementation

[0064] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the protection scope of the present invention.

[0065] Example 1:

[0066] like Figure 1 As shown, this embodiment provides a method for dynamic calibration and adaptive allocation of carbon emission factors in warehousing, characterized by including:

[0067] S100. Collect and quantify multi-dimensional data affecting carbon emissions in the target warehouse in real time to obtain the dynamic storage feature vector of the target warehouse.

[0068] S200. Based on the storage feature vector of the target warehouse, generate an initial carbon emission factor by matching from the pre-constructed carbon emission basic factor library;

[0069] S300: Quantitatively score the data quality of the energy data of the target warehouse; dynamically calibrate the initial carbon emission factor using a Bayesian calibration algorithm to generate an accurate carbon emission factor;

[0070] S400. Initiate the allocation decision. Based on the data availability of the target warehouse and the business scenario, select the optimal carbon emission allocation method from multiple allocation methods and generate a cargo-level carbon emission allocation coefficient.

[0071] S500. Compare the precise carbon emission factor with the initial carbon emission factor, and update the precise carbon emission factor to the carbon emission basic factor library based on the satisfaction of the preset update triggering conditions.

[0072] It should be noted that the aforementioned carbon emission basic factor library is pre-built. The carbon emission basic factor library adopts a three-layer architecture of "industry-type-sub-scenario": the first layer is the industry-general factor layer, storing average warehousing emission factors for different industries such as power, retail, and manufacturing; the second layer is the warehouse type factor layer, storing basic factors for different types of warehouses such as ordinary warehouses, cold storage warehouses, frozen warehouses, automated storage and retrieval systems, hazardous materials warehouses, and zero-carbon warehouses; the third layer is the sub-scenario factor layer, storing scenario factors under different equipment configurations and operating modes.

[0073] Each factor entry in the carbon emission basic factor library includes a factor value, applicable characteristic range, data source, DQI data quality score, and update time.

[0074] A more specific example of step S100 is as follows:

[0075] The warehouse feature vector includes building features, equipment features, operational features, cargo features, and energy features; specifically:

[0076] The architectural features include building structure, building area, thermal insulation performance, and lighting design;

[0077] The equipment features include the type, power, energy efficiency rating, and quantity of equipment used for heating, cooling, ventilation, lighting, and handling;

[0078] The operational characteristics include storage methods, work processes, staffing, and inventory turnover rate;

[0079] The cargo characteristics include cargo type, temperature requirements, packaging form, and average storage time;

[0080] The energy characteristics include consumption data for electric energy, natural gas energy, and diesel energy, the proportion of green electricity self-generation and self-consumption, and energy storage charging and discharging data.

[0081] The collected multidimensional feature data is standardized and quantized to obtain the storage feature vector of the target warehouse; the expression of the storage feature vector is:

[0082] ;

[0083] Wherein, V represents the warehouse feature vector; n is a label index, representing the nth feature term; This represents the quantized score corresponding to the nth feature item.

[0084] A more specific example of step S200 is as follows:

[0085] like Figure 2 As shown, the steps for generating initial carbon emission factors from the carbon emission basic factor library include:

[0086] S201. Using an improved cosine similarity algorithm, calculate the similarity between the feature items in the storage feature vector of the target warehouse and the applicable feature items in each factor entry of the carbon emission basic factor library, using the following formula:

[0087] ;

[0088] Wherein, V represents the storage feature vector of the target warehouse; n represents the total number of feature terms in the storage feature vector; This represents the quantized score of the k-th feature term in the warehouse feature vector; This represents the weight of the k-th feature term in the warehouse feature vector; This represents the i-th factor in the aforementioned carbon emission basic factor library; This represents the quantized score of the k-th feature term in the entry for the i-th factor;

[0089] S202. Sort the calculated similarities from highest to lowest.

[0090] S203. Select the top several basic factors in the similarity ranking and generate the initial carbon emission factor of the target warehouse by weighted averaging. In this embodiment, during the similarity ranking, the first three basic factors can be selected, and then the initial carbon emission factor is generated by weighted averaging. .

[0091] A more specific example of step S300 is as follows:

[0092] like Figure 3 As shown, the steps for dynamically calibrating the initial carbon emission factor include:

[0093] S301. The energy data of the target warehouse is scored based on four dimensions: activity data credibility, factor data credibility, time representativeness, and geographical representativeness, to obtain the data quality score of the energy data.

[0094] The credibility of the activity data refers to the accuracy of the energy data such as electricity, gas, and oil actually consumed by the target warehouse (e.g., meter accuracy and meter reading frequency).

[0095] Factor data credibility refers to the reliability of the source of the carbon emission factors used in the data quality assessment of energy data;

[0096] Geographical representativeness refers to the degree of matching between the climate or power grid structure of the source region of the cited carbon emission factors and the location of the target warehouse.

[0097] The time representativeness refers to whether the time period for which the energy data is collected can represent the typical operating cycle of the target warehouse;

[0098] Since the core sources of error in carbon emission accounting are "data accuracy" (credibility category) and "data suitability" (representative category), the above four dimensions cover the key elements of carbon emission accounting reliability and are consistent with the data quality assessment framework recommended in the Intergovernmental Panel on Climate Change (IPCC) National Greenhouse Gas Inventory Guidelines.

[0099] S302. Determine the data quality weight corresponding to the energy data based on the data quality score;

[0100] Specifically, the data quality weights are obtained by normalizing the data quality scores using the following formula:

[0101] ;

[0102] Among them, the Indicates data quality weights, and The Q represents the data quality score; This indicates the preset lower limit of the rating (e.g.) =1); the This indicates the preset lower limit of the rating (e.g.) =5); the above This represents the set minimum weight threshold (e.g.) =0.05), to ensure that the Always greater than 0.

[0103] S303. Using the initial carbon emission factor as the prior distribution and the daily carbon emission per unit area as the likelihood function, Bayesian calibration is performed in conjunction with the data quality weights to obtain the mean of the posterior distribution as the calibrated precise carbon emission factor.

[0104] Specifically, the formula for calculating the daily carbon emissions per unit area is as follows:

[0105] ;

[0106] Among them, the This represents the daily carbon emissions per unit area; This represents the consumption of the i-th type of energy; the... The carbon emission factor of the i-th energy source is represented; S represents the warehouse floor area; and D represents the number of days in a year.

[0107] The formula used for the Bayesian calibration is:

[0108] ;

[0109] Among them, the Indicates precise carbon emission factors; the Indicates data quality weights; the This represents the initial carbon emission factor.

[0110] As another alternative case: Figure 1 As shown, when the target warehouse is a zero-carbon warehouse equipped with photovoltaic or energy storage systems, the daily carbon emissions per unit area are calculated during the execution of the Bayesian calibration algorithm. The carbon emissions corresponding to self-generated and self-consumed green electricity are deducted, and the impact of energy storage charging and discharging efficiency is taken into account.

[0111] Daily carbon emissions per unit area of ​​the zero-carbon warehouse The calculation formula is:

[0112] ;

[0113] Among them, the This indicates the annual electricity purchased from outside the country; the aforementioned This represents the annual self-generated and self-consumed green electricity; the aforementioned This represents the carbon emission factor of the power grid.

[0114] A more specific example of step S400 is as follows:

[0115] The allocation decision process includes: detecting the available data types of the target warehouse and generating multiple available allocation methods; ranking the available allocation methods according to the business scenario of the target warehouse and the DQI data quality score of the precise carbon emission factor, selecting the allocation method with the highest priority, and calculating the carbon emission allocation coefficient for each item.

[0116] The detection of the available data types of the target warehouse refers to probing the data source of the target warehouse to determine the data types of cargo attributes that the warehouse can provide. The data types include at least one or a combination of weight, volume, storage duration, and cargo turnover rate.

[0117] For the business scenario of the target warehouse: the registration attribute information of the warehouse and the associated equipment and goods data can be obtained; based on the warehouse structure type, temperature control level and energy configuration information in the registration attribute information, and combined with the attribute distribution of the goods stored in the warehouse, the business scenario tag corresponding to the target warehouse can be determined.

[0118] The generated allocation methods include: allocation by floor area, allocation by weight, allocation by volume, allocation by value, and weighted allocation by storage duration.

[0119] Preferably, for power material storage, the allocation should be based on the land area, and for bulk commodity storage, the allocation should be based on the weight.

[0120] A more specific example of step S500 is as follows:

[0121] The preset update trigger condition is: the deviation between the accurate carbon emission factor and the initial carbon emission factor is greater than 15%, and the data quality score of the energy data of the target warehouse is greater than or equal to 2. Here, the deviation is the absolute value of the difference between the accurate carbon emission factor and the initial carbon emission factor, as a percentage relative to either the initial carbon emission factor or the accurate carbon emission factor.

[0122] When the preset update triggering conditions are met, namely, the deviation between the accurate carbon emission factor and the initial carbon emission factor is greater than 15%, and the data quality score of the energy data is greater than or equal to 2, the accurate carbon emission factor and the corresponding storage characteristics are updated to the carbon emission basic factor library.

[0123] If the preset update triggering conditions are not met, the precise carbon emission factor does not need to be updated to the carbon emission basic factor library.

[0124] As an optional approach, the factors in the carbon emission basic factor library can be periodically clustered to merge similar factors and delete outdated factors, in order to maintain the simplicity and accuracy of the carbon emission basic factor library.

[0125] The specific embodiments of the present invention will be described in detail below with reference to two specific examples:

[0126] It should be noted that both Case 1 and Case 2 use a pre-built library of basic carbon emission factors.

[0127] Case 1 involves the calibration and allocation of carbon factor emissions from a typical warehouse in a certain location.

[0128] The characteristic data of the ordinary warehouse were collected: the warehouse building structure is a single-story warehouse with an area of ​​9072m². 2 The wall insulation performance is average, and LED lighting is used; all equipment is electrified, including electric forklifts, lighting equipment, and office equipment; the operation mode is manual-assisted mechanization, and the inventory turnover rate is low; the goods type is electrical materials, with no weight data but available floor space data; energy consumption is solely electricity, with an annual power consumption of 45,000 kWh. These characteristics are quantified into a warehouse feature vector V.

[0129] The storage feature vector V is matched with a pre-built database of basic carbon emission factors. The top three factors in the similarity ranking are selected: 0.0082 kg CO2e / (m²·d), 0.0071 kg CO2e / (m²·d), and 0.0079 kg CO2e / (m²·d). A weighted average is then used to obtain the initial carbon emission factor. =0.0077 kgCO2e / (m²·d).

[0130] Calculate the daily carbon emissions per unit area of ​​the aforementioned ordinary warehouse. =0.00757 kgCO2e, the data quality evaluation score of the energy consumption data is 2 points (industry statistical data), data quality weight =0.7. The calibrated precise carbon emission factor was obtained through Bayesian calibration. =0.7×0.00757+0.3×0.0077=0.00761kgCO2e / (m²·d).

[0131] The system detected no cargo weight data for this ordinary warehouse, but it has data on floor area and storage duration. The business scenario is power material storage, so the allocation method based on floor area × storage duration is selected. The carbon emissions of an oil-immersed transformer with a rated voltage of 10kV (floor area of ​​2m², stored for 120 days) are calculated to be: 2 × 120 × 0.00761 = 1.826 kgCO2e.

[0132] Since the deviation between the precise carbon emission factor 0.00761 kgCO2e / (m²·d) and the initial carbon emission factor 0.0077 kgCO2e / (m²·d) is less than 15%, it is not necessary to update the precise carbon emission factor into the carbon emission basic factor library.

[0133] Case 2 involves the calibration and allocation of carbon factor emissions from a zero-carbon warehouse in a certain location:

[0134] The following characteristic data about the zero-carbon warehouse were collected: the warehouse building structure is a high-standard single-story warehouse with an area of ​​17,000 m². 2 It boasts excellent thermal insulation performance; equipped with a microgrid system, 590kW of photovoltaic capacity, and 466kWh of energy storage; all equipment is electrified, including electric forklifts and automated handling equipment; the operation mode is semi-automated, with a high inventory turnover rate; the goods are electrical materials; energy consumption is mainly from self-generated and self-consumed green electricity, with an annual photovoltaic power generation of 475,505kWh and purchased electricity of 1,057,516kWh. These characteristics are quantified into a warehousing feature vector V.

[0135] The storage feature vector V is matched with a pre-built library of basic carbon emission factors to obtain the initial carbon emission factors. =0.0012 kgCO2e / (m²·d).

[0136] When calculating daily carbon emissions per unit area, deduct the carbon emissions corresponding to self-generated and self-consumed green electricity: Total annual carbon emissions = (1057516 - 475505) × 0.5568 = 323725 kg CO2e, daily carbon emissions per unit area =323725 / (17000×365)=0.00523kgCO2e. The data quality evaluation score of the energy consumption data is 1 point (real-world production data), and the data quality weight is... =0.9. The calibrated precise carbon emission factor was obtained through Bayesian calibration. =0.9×0.00523+0.1×0.0012=0.00483kgCO2e / (m²·d).

[0137] The zero-carbon warehouse was found to have no cargo weight data. The business scenario is power material storage. The preferred allocation method is based on floor area × storage duration. (Case 2 mainly illustrates the factor calibration process in a zero-carbon scenario, i.e., how to calculate daily carbon emissions per unit area.) After deducting self-generated and self-consumed green electricity to complete Bayesian calibration, the allocation decision process can be referred to in Case 1).

[0138] If the deviation between the precise carbon emission factor 0.00483 kgCO2e / (m²·d) and the initial carbon emission factor 0.0012 kgCO2e / (m²·d) is greater than 15%, the data quality score of the energy data is 1 point, which does not meet the update trigger condition of a data quality score ≥ 2 points. Therefore, the precise carbon emission factor cannot be updated to the carbon emission basic factor library.

[0139] Example 2:

[0140] This invention provides a dynamic calibration and adaptive allocation system for carbon emission factors in warehousing, comprising: a multi-dimensional feature acquisition module for warehousing, a basic carbon emission factor library, an initial carbon emission factor matching and generation module, a calibration module, an adaptive allocation decision module, a factor iterative update and sharing module, and a visualization interface module;

[0141] The warehouse multi-dimensional feature acquisition module is used to collect and quantify multi-dimensional data affecting warehouse carbon emissions in real time, and construct a dynamic warehouse feature vector library for the target warehouse; the warehouse multi-dimensional feature acquisition module includes:

[0142] The building feature acquisition unit is used to collect data on features such as building structure, area, thermal insulation performance, and lighting design.

[0143] The equipment feature acquisition unit is used to collect data on the type, power, energy efficiency rating, and quantity of equipment used for heating, cooling, ventilation, lighting, and handling.

[0144] The operational feature acquisition unit is used to collect data on features such as storage method, work process, personnel configuration, and inventory turnover rate;

[0145] The cargo feature acquisition unit is used to collect data on features such as cargo type, temperature requirements, packaging form, and average storage time.

[0146] The energy characteristic acquisition unit is used to collect data on the consumption of electric energy, natural gas energy and diesel energy, the proportion of green electricity self-generation and self-consumption, and energy storage charging and discharging data, among other characteristic items.

[0147] In addition, a feature quantization unit is used to quantize multi-dimensional data collected in real time to obtain a dynamic warehouse feature vector library. Wherein, V represents the warehouse feature vector; n is a label index, representing the nth feature term; This represents the quantized score corresponding to the nth feature item.

[0148] The carbon emission basic factor library is used to store basic carbon emission factors for storage at different levels and of different types.

[0149] The initial carbon emission factor matching and generation module is used to match and generate initial carbon emission factors from the carbon emission basic factor library based on the warehouse feature vector of the target warehouse.

[0150] The calibration module is used to dynamically calibrate the initial carbon emission factor based on the energy data of the target warehouse and the corresponding data quality evaluation results, generating a scenario-specific, accurate carbon emission factor; the calibration module includes:

[0151] The data quality evaluation unit scores the energy characteristic data of the target warehouse based on four dimensions: activity data credibility, factor data credibility, time representativeness, and geographical representativeness, to obtain a data quality score for the energy characteristic data; the data quality score is then normalized to obtain a data quality weight.

[0152] The Bayesian calibration unit is used to perform Bayesian calibration with the initial carbon emission factor as the prior distribution, the daily carbon emission per unit area as the likelihood function, and the data quality weights, to obtain the mean of the posterior distribution as the calibrated accurate carbon emission factor.

[0153] A special processing unit for zero-carbon scenarios is used to deduct the carbon emissions corresponding to self-generated and self-consumed green electricity when calculating the daily carbon emissions per unit area for zero-carbon warehouses equipped with photovoltaic or energy storage systems.

[0154] The adaptive allocation decision module is used to select the optimal carbon emission allocation method from multiple allocation methods based on the data availability of the target warehouse and the business scenario, and generate cargo-level carbon emission allocation coefficients.

[0155] The factor iteration update and sharing module is used to update the precise carbon emission factors to the carbon emission basic factor library based on preset update trigger conditions. It periodically performs cluster analysis on the factors in the carbon emission basic factor library, merging similar factors and deleting outdated factors to maintain the simplicity and accuracy of the carbon emission basic factor library. In addition, the factor iteration update and sharing module also provides a sharing interface to support upstream and downstream enterprises in the supply chain to share the precise carbon emission factors.

[0156] The visualization interface module provides visualization functions, including a comparison chart of factors before and after calibration (initial carbon emission factors versus precise carbon emission factors), a weighted chart of the impact of each feature data collected by the warehouse multi-dimensional feature acquisition module on carbon emissions, detailed cargo carbon emission allocation, and a data quality scoring report. Additionally, the visualization interface module also provides interfaces for connecting to the carbon accounting platform, including a carbon emission factor query interface, a cargo carbon emission accounting interface, and a factor data reporting interface.

[0157] This invention is not limited to the above-described optional embodiments. Anyone can derive other various forms of products under the guidance of this invention. However, regardless of any changes made in their shape or structure, any technical solution that falls within the scope of the claims of this invention shall be protected by this invention.

Claims

1. A method for dynamic calibration and adaptive allocation of carbon emission factors in warehousing, characterized in that, include: Real-time collection and quantification of multi-dimensional data affecting carbon emissions in the target warehouse to obtain a dynamic warehousing feature vector of the target warehouse; wherein, the warehousing feature vector includes building features, equipment features, operational features, cargo features and energy features; Based on the storage feature vector of the target warehouse, an initial carbon emission factor is generated by matching from a pre-constructed carbon emission basic factor library; wherein, each factor entry in the carbon emission basic factor library includes a factor value, applicable feature range, data source, DQI data quality score, and update time; The data quality of the energy data of the target warehouse is quantitatively scored; the initial carbon emission factor is dynamically calibrated using a Bayesian calibration algorithm to generate an accurate carbon emission factor; wherein, when the target warehouse is a zero-carbon warehouse, the carbon emissions corresponding to the self-generated and self-consumed green electricity of the zero-carbon warehouse are deducted during the execution of the Bayesian calibration algorithm. Initiate the allocation decision, and based on the data availability and business scenario of the target warehouse, select the optimal carbon emission allocation method from multiple allocation methods to generate cargo-level carbon emission allocation coefficients; By comparing the precise carbon emission factor with the initial carbon emission factor, and based on meeting the preset update trigger conditions, the precise carbon emission factor is updated to the carbon emission basic factor library.

2. The method for dynamic calibration and adaptive allocation of carbon emission factors in warehousing according to claim 1, characterized in that, Generating initial carbon emission factors by matching from the aforementioned carbon emission basic factor library includes: An improved cosine similarity algorithm is used to calculate the similarity between the feature terms in the storage feature vector of the target warehouse and the applicable feature terms in each factor entry of the carbon emission basic factor library. The formula is as follows: ； Wherein, V represents the storage feature vector of the target warehouse; n represents the total number of feature terms in the storage feature vector; This represents the quantized score of the k-th feature term in the warehouse feature vector; This represents the weight of the k-th feature term in the warehouse feature vector; This represents the i-th factor in the aforementioned carbon emission basic factor library; This represents the quantized score of the k-th feature term in the entry for the i-th factor; The calculated similarities are sorted from highest to lowest; the top few basic factors in the similarity ranking are selected and the initial carbon emission factor of the target warehouse is generated by weighted averaging.

3. The method for dynamic calibration and adaptive allocation of carbon emission factors in warehousing according to claim 2, characterized in that, Dynamic calibration of the initial carbon emission factor includes: The energy data of the target warehouse is scored based on four dimensions: activity data credibility, factor data credibility, time representativeness, and geographical representativeness, to obtain the data quality score of the energy data. Based on the data quality score, determine the data quality weight corresponding to the energy data; Using the initial carbon emission factor as the prior distribution and the daily carbon emission per unit area as the likelihood function, Bayesian calibration is performed in conjunction with the data quality weights to obtain the mean of the posterior distribution, which is then used as the calibrated precise carbon emission factor.

4. The method for dynamic calibration and adaptive allocation of carbon emission factors in warehousing according to claim 3, characterized in that, The formula for calculating the daily carbon emissions per unit area is: ； Among them, the This represents the daily carbon emissions per unit area; This represents the consumption of the i-th type of energy; the... The carbon emission factor of the i-th energy source is represented; S represents the warehouse floor area; and D represents the number of days in a year. The formula used for the Bayesian calibration is: ； Among them, the Indicates precise carbon emission factors; the Indicates data quality weights; the This represents the initial carbon emission factor.

5. The method for dynamic calibration and adaptive allocation of carbon emission factors in warehousing according to claim 3, characterized in that, When the target warehouse is a zero-carbon warehouse equipped with photovoltaic or energy storage systems, the carbon emissions corresponding to self-generated and self-consumed green electricity are deducted when calculating the daily carbon emissions per unit area.

6. The method for dynamic calibration and adaptive allocation of carbon emission factors in warehousing according to claim 1, characterized in that, The process for making the allocation decision includes: The available data types in the target warehouse are detected, and multiple usable allocation methods are generated. Based on the business scenario of the target warehouse and the DQI data quality score of the precise carbon emission factor, the available allocation methods are ranked by priority, the allocation method with the highest priority is selected, and the carbon emission allocation coefficient for each item is calculated. The generated allocation methods include: allocation by floor area, allocation by weight, allocation by volume, allocation by value, and weighted allocation by storage duration.

7. The method for dynamic calibration and adaptive allocation of carbon emission factors in warehousing according to claim 1, characterized in that, The preset update trigger condition is: the deviation between the accurate carbon emission factor and the initial carbon emission factor is greater than 15%, and the data quality score of the energy data of the target warehouse is greater than or equal to 2.

8. The method for dynamic calibration and adaptive allocation of carbon emission factors in warehousing according to claim 1, characterized in that, In the storage feature vector, The architectural features include building structure, building area, thermal insulation performance, and lighting design; The equipment features include the type, power, energy efficiency rating, and quantity of equipment used for heating, cooling, ventilation, lighting, and handling; The operational characteristics include storage methods, work processes, staffing, and inventory turnover rate; The cargo characteristics include cargo type, temperature requirements, packaging form, and average storage time; The energy characteristics include consumption data for electric energy, natural gas energy, and diesel energy, the proportion of green electricity self-generation and self-consumption, and energy storage charging and discharging data; The expression for the storage feature vector of the target warehouse is: ； Wherein, V represents the warehouse feature vector; n is a label index, representing the nth feature term; This represents the quantized score corresponding to the nth feature item.

9. The method for dynamic calibration and adaptive allocation of carbon emission factors in warehousing according to claim 1, characterized in that, The carbon emission basic factor library adopts a multi-layer architecture; the first layer is the industry general factor layer, which stores the average warehousing emission factors of different industries; the second layer is the warehouse type factor layer, which stores the basic factors of different types of warehouses; the third layer is the sub-scenario factor layer, which stores the scenario factors under different equipment configurations and different operating modes.

10. A dynamic calibration and adaptive allocation system for carbon emission factors in warehousing, characterized in that, include: The warehouse multi-dimensional feature acquisition module is used to collect and quantify multi-dimensional data affecting warehouse carbon emissions in real time, and build a dynamic warehouse feature vector library for the target warehouse. A carbon emission basic factor library is used to store basic carbon emission factors for warehousing at different levels and of different types. The initial carbon emission factor matching and generation module is used to match and generate initial carbon emission factors from the carbon emission basic factor library based on the warehouse feature vector of the target warehouse. The calibration module is used to dynamically calibrate the initial carbon emission factor based on the energy data of the target warehouse and the corresponding data quality evaluation results, so as to generate a scenario-based accurate carbon emission factor. The adaptive allocation decision module is used to select the optimal carbon emission allocation method from multiple allocation methods based on the data availability of the target warehouse and the business scenario, and generate cargo-level carbon emission allocation coefficients. The factor iteration update and sharing module is used to update the precise carbon emission factor to the carbon emission basic factor library based on the satisfaction of preset update trigger conditions; the factor iteration update and sharing module is also used to provide a sharing interface to support upstream and downstream enterprises in the supply chain to share the precise carbon emission factor. In addition, a visualization interface module is provided to offer visualization functionality; the visualization interface module is also used to provide an interface for interfacing with the carbon accounting platform.

Images