Enterprise power consumption optimization method and system considering electric power carbon emission responsibility allocation

Abstract

The invention provides an enterprise power consumption optimization method and system considering electric power carbon emission responsibility allocation, and the method comprises the steps: dividing the power consumption source of an enterprise into traceable power consumption and non-traceable power consumption based on the power consumption data of the enterprise; the method comprises the following steps: firstly, calculating the power carbon emission of an enterprise according to the principle of accurate calculation of traceable electric quantity and average allocation of non-traceable electric quantity, then comprehensively considering the power purchase cost and the power carbon emission of the enterprise, constructing an enterprise low-carbon power consumption optimization model, and finally solving the enterprise low-carbon power consumption optimization model to obtain an enterprise low-carbon power consumption optimization scheme. According to the invention, accurate accounting of power carbon emission of an enterprise and refined low-carbon management of power are facilitated, and repeated calculation of green power environment rights and interests is avoided.

Description

Technical Field

[0001] The present invention belongs to the field of electric power technology, and specifically relates to a method and system for optimizing enterprise electricity consumption taking into account the allocation of electricity carbon emission responsibilities. Background Art

[0002] In the context of the global low-carbon transition, the power industry, as a core sector of carbon emissions, faces a critical bottleneck in ensuring the scientific and fair allocation of carbon responsibility. The current mainstream, one-way responsibility allocation model, characterized by "full responsibility on the power generation side" or "full responsibility on the power consumption side," is ill-suited to the complex carbon flows inherent in the deep interplay between source, grid, and load in the new power system. This leads to systemic biases in supply chain carbon footprint accounting, green electricity trading decisions, and production layout optimization.

[0003] At present, the method of allocating carbon emission responsibilities for electricity still has the following technical bottlenecks. First, it does not fully consider the dynamic impact of load-side electricity consumption behavior on grid carbon emissions. Especially in cross-regional power transmission scenarios, the division of responsibilities often ignores the implicit carbon emissions generated by auxiliary services such as grid losses, peak shaving and frequency regulation, resulting in difficulty in the coordinated release of emission reduction potential on both the source and load sides; second, the annual static emission factor adopted by the current standard is difficult to capture the minute-level carbon intensity changes caused by fluctuations in new energy output, resulting in enterprises being unable to identify high-carbon emission electricity consumption periods and implement demand-side response; third, the current accounting system has not established a dynamic correlation model between carbon emission responsibilities and environmental rights products such as green electricity and green certificates, resulting in a lack of quantitative low-carbon decision-making basis for enterprises when implementing strategies such as power purchase agreements. Summary of the Invention

[0004] The purpose of the present invention is to address the above-mentioned problems existing in the prior art and to provide an enterprise electricity optimization method and system that can achieve accurate accounting of enterprise electricity carbon emissions and refined low-carbon management of electricity, taking into account the allocation of electricity carbon emissions responsibilities.

[0005] To achieve the above objectives, the technical solutions of the present invention are as follows:

[0006] In a first aspect, the present invention proposes a method for optimizing enterprise electricity consumption taking into account the allocation of electricity carbon emission responsibilities, comprising:

[0007] S1. Based on the enterprise's electricity consumption data, the enterprise's electricity sources are divided into traceable electricity and non-traceable electricity;

[0008] S2. Calculate the company's electricity carbon emissions based on the principle of accurate accounting of traceable electricity and equal allocation of non-traceable electricity;

[0009] S3. Comprehensively consider the enterprise's electricity purchase cost and electricity carbon emissions, and build an enterprise low-carbon electricity optimization model;

[0010] S4. Solve the enterprise's low-carbon electricity consumption optimization model and obtain the enterprise's low-carbon electricity consumption optimization plan.

[0011] S2 uses the following formula to calculate the company's electricity carbon emissions:

[0012]

[0013] Q T,g =(G self,g -G net,g )+G con,g +G green

[0014]

[0015] In the above formula, C ele is the carbon emissions of corporate electricity, Q all is the total electricity consumption of the enterprise, Q T,g , Q NT,g are the enterprise's traceable g-th non-fossil energy electricity consumption and the non-traceable g-th non-fossil energy electricity consumption, κ i 、W i 、V i are the carbon dioxide emission factor, the amount used for thermal power generation, and the average lower calorific value of fossil fuel i consumed for power generation in the region where the enterprise is located, respectively. fire , Q bio They are the thermal power generation and biomass power generation in the area where the enterprise is located, G self,g , G net,g are the self-generated electricity and grid-connected electricity of the enterprise’s g-th non-fossil energy source, G con,g , G green They are the non-fossil energy electricity of the gth type in the medium- and long-term bilateral negotiated transactions of enterprises and the electricity of the green electricity transaction contract, Q mar,g , Q mar,g,T are the total non-fossil energy electricity in the electricity market, the total traceable non-fossil energy electricity, and Q mar , Q mar,T They are the total traded electricity and the total traceable electricity in the electricity market respectively.

[0016] S1 uses dynamic labeling technology to divide the company's electricity sources into traceable electricity and non-traceable electricity, including:

[0017] S11. Collect the enterprise's electricity consumption data and classify it to obtain the enterprise's electricity consumption data dynamic labels in each time period. The enterprise's electricity consumption data dynamic labels D(T) in time period T are:

[0018] D(T)={Q(T),T,u}

[0019] In the above formula, Q(T) is the total electricity consumption in the T period, T is the timestamp, and u is the type of electricity source, including self-generated and self-consumed electricity, medium- and long-term bilateral negotiated transactions, green power transactions, priority electricity purchase, grid agency purchase, retail market transactions, and wholesale market transactions. Among them, self-generated and self-consumed electricity, medium- and long-term bilateral negotiated transactions, and green power transactions are traceable electricity, while priority electricity purchase, grid agency purchase, retail market transactions, and wholesale market transactions are not traceable electricity.

[0020] S12. Use the following verification function to verify the electricity consumption data whose source type is traceable electricity:

[0021]

[0022]

[0023] In the above formula, V self (T), V con (T), V green (T) are the verification functions of self-generated and self-used electricity, medium- and long-term bilateral negotiated transaction electricity, and green electricity transaction electricity, respectively. self,T , G net,T are the self-generated electricity and grid-connected electricity of the enterprise in period T, G con is the contracted electricity, [t a ,t b ] is the effective time window of the contract, G green The amount of electricity in green electricity trading contracts.

[0024] In S3, the objective function of the enterprise low-carbon electricity optimization model includes:

[0025]

[0026] In the above formula, α(T) is the dynamic weight coefficient of the T period, Q j,T , G net,T are the electricity consumption of the jth type of electricity source in period T and the enterprise's online electricity consumption, λ j ,λ net are the electricity price of the jth type of electricity source and the enterprise response electricity price, N green,T is the number of green certificates generated by the enterprise's renewable energy power generation during period T, λ green is the price corresponding to the green certificate, X max 、X min are the highest and lowest electricity purchase costs in the history of the enterprise, C ele,T is the carbon emissions of electricity generated by the enterprise during period T, C max 、C minare the highest and lowest electricity carbon emissions in the history of the enterprise, c0 and c(T) are the annual average electricity carbon emission factor of the region where the enterprise is located and the average electricity carbon emission factor during period T, respectively; k is the adjustment factor;

[0027] Constraints include:

[0028]

[0029] In the above formula, G self,T is the self-generated electricity of the enterprise during period T, Q need,T is the amount of electricity required to maintain normal operation of the enterprise during period T, B is the enterprise's electricity purchase budget, and N green,T is the number of green certificates generated by the enterprise’s renewable energy power generation during period T, G self,g,T is the self-generated electricity of the enterprise using the g-th non-fossil energy in period T.

[0030] The S4 uses the improved NSGA-III algorithm to solve the enterprise electricity optimization model, including:

[0031] S41, randomly generate a power consumption plan X=(Q1, Q2, ..., Q j ,N green ) to initialize the population, where Q j is the power consumption of the jth power source type, N green The number of green certificates generated for the company’s new energy power generation;

[0032] S42. Determine corresponding dynamic weight coefficients according to different needs of the enterprise, calculate the objective function value, generate non-uniform dynamic reference point sets under different needs in the normalized target space, and match the distribution of dynamic reference points with the demand priority;

[0033] S43, performing non-dominated sorting on the population, calculating the distance between each individual and the dynamic reference point, and using the distance from the individual to the nearest dynamic reference point as a normalized value representing the superiority of the individual in the normalized target space;

[0034] S44, using crossover and mutation operations to generate new individuals and form a progeny population;

[0035] S45. Merge the parent population and the child population, perform local normalization on the newly added individuals, re-perform non-dominated sorting, and select the corresponding optimal solution based on the dynamic reference points of different requirements, and update the population. The local normalization includes updating the ideal point through the sliding window and locally screening the extreme points.

[0036] S46. Adjust the position of the dynamic reference point according to the distribution density of the new population;

[0037] S47. Repeat S43-S46 until the algorithm converges, extract the non-dominated solutions in the Pareto front, and obtain the low-carbon electricity optimization plan under different enterprise needs.

[0038] In said S45, the updated ideal point is the minimum value of the historical and newly added individuals;

[0039] Local screening of extreme points means: calculating the ASF value of the newly added individuals and comparing the ASF values ​​to update only the extreme points affected by the newly added individuals;

[0040] The S46 includes:

[0041] S461. Calculate the individual distribution density of the new population in each dimension in the normalized target space:

[0042]

[0043] In the above formula, α i is the individual distribution density of the i-th dimension, N i is the number of individuals in the i-th dimension in the normalized target space, and N is the total number of individuals in the normalized target space;

[0044] S462. Determine the adjusted dynamic reference point position according to the following formula:

[0045]

[0046] In the above formula, They are the dynamic reference point positions before and after adjustment respectively.

[0047] In a second aspect, the present invention proposes an enterprise electricity optimization system that takes into account the allocation of electricity carbon emission responsibilities, including an electricity source division module, an electricity carbon emission accounting module, a model construction module, and a model solution module;

[0048] The power source classification module is used to classify the power sources of the enterprise into traceable power and non-traceable power based on the enterprise's power consumption data;

[0049] The electricity carbon emission accounting module is used to calculate the enterprise's electricity carbon emissions based on the principle of accurate accounting of traceable electricity and average allocation of non-traceable electricity;

[0050] The model building module is used to comprehensively consider the enterprise's electricity purchase cost and electricity carbon emissions to build an enterprise low-carbon electricity optimization model;

[0051] The model solving module is used to solve the enterprise low-carbon electricity consumption optimization model and obtain the enterprise low-carbon electricity consumption optimization plan.

[0052] The electricity carbon emissions calculation module calculates the enterprise's electricity carbon emissions based on the following formula:

[0053]

[0054] Q T,g =(G self,g -G net,g )+G con,g +G green

[0055]

[0056] In the above formula, C ele is the carbon emissions of corporate electricity, Q all is the total electricity consumption of the enterprise, Q T,g , Q NT,g are the enterprise's traceable g-th non-fossil energy electricity consumption and non-traceable g-th non-fossil energy electricity consumption, κ i 、W i 、V i are the carbon dioxide emission factor, the amount used for thermal power generation, and the average lower calorific value of fossil fuel i consumed for power generation in the region where the enterprise is located, respectively. fire , Q bio They are the thermal power generation and biomass power generation in the area where the enterprise is located, G self,g , G net,g are the self-generated electricity and grid-connected electricity of the enterprise’s g-th non-fossil energy source, G con,g , G green They are the non-fossil energy electricity of the gth type in the medium- and long-term bilateral negotiated transactions of enterprises and the electricity of the green electricity transaction contract, Q mar,g , Q mar,g,T are the total non-fossil energy electricity in the electricity market, the total traceable non-fossil energy electricity, and Q mar , Q mar,T They are the total traded electricity and the total traceable electricity in the electricity market respectively;

[0057] The objective function of the enterprise low-carbon electricity optimization model includes:

[0058]

[0059] In the above formula, α(T) is the dynamic weight coefficient of the T period, Q j,T , G net,T are the electricity consumption of the jth type of electricity source in period T and the enterprise's online electricity consumption, λ j ,λ net are the electricity price of the jth type of electricity source and the enterprise response price, N green,T is the number of green certificates generated by the enterprise's renewable energy power generation during period T, λ green is the price corresponding to the green certificate, X max、X min are the highest and lowest electricity purchase costs in the history of the enterprise, C ele,T is the carbon emissions of electricity generated by the enterprise during period T, C max 、C min are the highest and lowest electricity carbon emissions in the history of the enterprise, c0 and c(T) are the annual average electricity carbon emission factor of the region where the enterprise is located and the average electricity carbon emission factor during period T, respectively; k is the adjustment factor;

[0060] Constraints include:

[0061]

[0062] In the above formula, G self,T is the self-generated electricity of the enterprise during period T, Q need,T is the amount of electricity required to maintain normal operation of the enterprise during period T, B is the enterprise's electricity purchase budget, and N green,T is the number of green certificates generated by the enterprise’s renewable energy power generation during period T, G self,g,T is the self-generated electricity of the enterprise using the g-th non-fossil energy in period T.

[0063] The electricity source classification module uses dynamic labeling technology to classify the enterprise's electricity sources into traceable electricity and non-traceable electricity, including a classification unit and a verification unit;

[0064] The classification unit is used to collect the enterprise's electricity consumption data for classification processing to obtain the enterprise's electricity consumption data dynamic labels in each time period. The enterprise's electricity consumption data dynamic labels D(T) in time period T are:

[0065] D(T)={Q(T),T,u}

[0066] In the above formula, Q(T) is the total electricity consumption in the T period, T is the timestamp, and u is the type of electricity source, including self-generated and self-consumed electricity, medium- and long-term bilateral negotiated transactions, green power transactions, priority electricity purchase, grid agency purchase, retail market transactions, and wholesale market transactions. Among them, self-generated and self-consumed electricity, medium- and long-term bilateral negotiated transactions, and green power transactions are traceable electricity, while priority electricity purchase, grid agency purchase, retail market transactions, and wholesale market transactions are not traceable electricity.

[0067] The verification unit verifies the electricity consumption data whose source type is traceable electricity using the following verification function:

[0068]

[0069] In the above formula, V self (T), V con (T), V green(T) are the verification functions of self-generated and self-used electricity, medium- and long-term bilateral negotiated transaction electricity, and green electricity transaction electricity, respectively. self,T , G net,T are the self-generated electricity and grid-connected electricity of the enterprise in period T, G con is the contracted electricity, [t a ,t b ] is the effective time window of the contract, G green The amount of electricity in green electricity trading contracts.

[0070] The model solving module uses the improved NSGA-III algorithm to solve the enterprise electricity optimization model. The specific process of the algorithm includes:

[0071] A. Randomly generate a power consumption plan X=(Q1, Q2, ..., Q j ,N green ) to initialize the population, where Q j is the power consumption of the jth power source type, N green The number of green certificates generated for the company’s new energy power generation;

[0072] B. Determine the corresponding dynamic weight coefficient based on the different needs of the enterprise, calculate the objective function value, and generate a non-uniform set of dynamic reference points under different needs in the normalized target space, so that the distribution of dynamic reference points matches the demand priority;

[0073] C. Perform non-dominated sorting on the population, calculate the distance between each individual and the dynamic reference point, and use the distance from the individual to the nearest dynamic reference point as the normalized value representing the superiority of the individual in the normalized target space;

[0074] D. Use crossover and mutation operations to generate new individuals and form a progeny population;

[0075] E. Merge the parent population and the child population, perform local normalization on the newly added individuals, re-perform non-dominated sorting, and select the corresponding optimal solution based on the dynamic reference points of different requirements to update the population. The local normalization process includes:

[0076] Update the ideal point through the sliding window so that the updated ideal point is the minimum value of the history and the newly added individual;

[0077] Locally screen extreme points, including calculating the ASF value of the newly added individuals and comparing the ASF values ​​to update only the extreme points affected by the newly added individuals;

[0078] F. Adjust the position of the dynamic reference point according to the distribution density of the new population, including:

[0079] F1, statistics the individual distribution density of the new population in each dimension in the normalized target space:

[0080]

[0081] In the above formula, α i is the individual distribution density of the i-th dimension, N i is the number of individuals in the i-th dimension in the normalized target space, and N is the total number of individuals in the normalized target space;

[0082] F2. Determine the adjusted dynamic reference point position according to the following formula:

[0083]

[0084] In the above formula, They are the dynamic reference point positions before and after adjustment respectively;

[0085] G. Repeat CF until the algorithm converges, extract the non-dominated solutions in the Pareto front, and obtain the low-carbon electricity optimization plan under different enterprise needs.

[0086] Compared with the prior art, the present invention has the following beneficial effects:

[0087] 1. The present invention provides an enterprise electricity optimization method that takes into account the sharing of electricity carbon emissions responsibilities. First, based on the enterprise electricity consumption data, the enterprise's electricity sources are divided into traceable electricity and non-traceable electricity. Then, the enterprise's electricity carbon emissions are calculated according to the principle of accurate accounting of traceable electricity and average sharing of non-traceable electricity. Then, the enterprise's electricity purchase cost and electricity carbon emissions are comprehensively considered to construct an enterprise low-carbon electricity optimization model. Finally, the enterprise low-carbon electricity optimization model is solved to obtain the enterprise low-carbon electricity optimization plan. This method divides the enterprise's electricity sources into traceable electricity and non-traceable electricity, and fully considers the sharing of responsibilities for non-traceable electricity. It not only helps the enterprise to accurately calculate electricity carbon emissions and conduct refined low-carbon management of electricity, but also avoids the repeated calculation of green electricity environmental rights and interests, and enhances the enthusiasm of enterprises for green electricity consumption.

[0088] 2. The present invention provides an enterprise electricity optimization method that takes into account the sharing of electricity carbon emission responsibilities. It introduces a dynamic weight coefficient into the enterprise low-carbon electricity optimization model, and adjusts the dynamic weight coefficient according to the different needs of the enterprise, and finally obtains a low-carbon electricity optimization plan under different needs. The enterprise can select the corresponding low-carbon electricity optimization plan based on actual needs, which is suitable for scenarios where multi-objective decision-making and rapid matching of needs are required.

[0089] 3. The present invention provides an enterprise electricity optimization method that takes into account the sharing of electricity carbon emission responsibilities. The improved NSGA-III algorithm is used to solve the enterprise low-carbon electricity optimization model. In terms of the reference point mechanism, the improved NSGA-III algorithm can, on the one hand, realize dynamic reference point generation, and adaptively update the reference point through the weight adjustment mechanism of the dynamic weight coefficient, so that the reference point is more in line with actual needs; on the other hand, the dynamic reference point position is adjusted according to the population distribution during the evolution process, so that the adjusted dynamic reference point position can adapt to the current solution set characteristics. In terms of convergence, the algorithm determines the reference point based on the demand preferences of the enterprise, and dynamically adjusts the reference point during the evolution process, which can improve the convergence speed. In terms of computational efficiency, the algorithm adopts an incremental normalization method. After merging the parent population and the child population, only the newly added individuals are locally normalized, which can reduce the complexity of normalization from the exponential level to the linear level, effectively reducing repeated calculations, speeding up the processing speed of high-dimensional problems, and improving computational efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0090] Figure 1 This is a flowchart of the method described in Example 1.

[0091] Figure 2 This is a structural diagram of the system described in Example 2. DETAILED DESCRIPTION

[0092] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0093] Example 1:

[0094] This embodiment takes a typical enterprise as the research object and implements the enterprise electricity optimization method of the present invention taking into account the allocation of electricity carbon emission responsibilities. Figure 1 The specific steps are as follows:

[0095] 1. Using dynamic labeling technology, the electricity consumption data of enterprises is divided into traceable electricity and non-traceable electricity according to the source of the electricity consumption of the enterprises. Specifically including:

[0096] 1.1. Collect enterprise electricity consumption data and carry out classification processing. The dynamic labels D(T) of the enterprise's electricity consumption data in period T are:

[0097] D(T)={Q(T),T,u}

[0098] In the above formula, Q(T) is the total electricity consumption in time period T, T is the timestamp, and u is the type of electricity source, including self-generated and self-consumed electricity, medium- and long-term bilateral negotiated transaction electricity, green electricity transaction electricity, priority electricity purchase, grid agency purchase of electricity, retail market-based transaction, and wholesale market-based transaction. Among them, self-generated and self-consumed electricity, medium- and long-term bilateral negotiated transaction electricity, and green electricity transaction electricity are traceable electricity, while priority electricity purchase, grid agency purchase of electricity, retail market-based transaction, and wholesale market-based transaction are non-traceable electricity.

[0099] The power status set S = {T, NT, NY}. When the power source type is traceable power, the power status is T; when the power source type is non-traceable power, the power status is NT; when the traceable power fails to pass the verification, the power status is NY.

[0100] 1.2. Use the following verification function to verify the electricity consumption data whose source type is traceable electricity:

[0101]

[0102] In the above formula, V self (T), V con (T), V green (T) are the verification functions of self-generated and self-used electricity, medium- and long-term bilateral negotiated transaction electricity, and green electricity transaction electricity, respectively. self,T , G net,T are the self-generated electricity and grid-connected electricity of the enterprise in period T, G con is the contracted electricity, [t a ,t b ] is the effective time window of the contract, G green The amount of electricity in green electricity trading contracts.

[0103] 1.3. When the verification function value of the electricity usage data is 1, it means that the verification has passed, and the power status label remains unchanged. When the verification function value of the electricity usage data is 0, it means that the verification has failed, and the power status label is changed from T to NY. Subsequently, manual review is performed to determine whether the electricity usage data is traceable or non-traceable.

[0104] 2. Calculate the company's electricity carbon emissions based on the principle of accurate accounting of traceable electricity and equal allocation of non-traceable electricity, specifically including:

[0105] 2.1. Traceable electricity is divided into traceable non-fossil energy electricity and traceable fossil energy electricity. Traceable non-fossil energy electricity is the direct accumulation of each part of electricity, which is:

[0106] Q T,g =(G self,g -G net,g )+G con,g +G green

[0107] In the above formula, Q T,g is the traceable non-fossil energy electricity consumption of the enterprise, G self,g , G net,g are the self-generated electricity and grid-connected electricity of the enterprise’s g-th non-fossil energy source, G con,g , G green They are respectively the g-th non-fossil energy electricity portion of the enterprise's medium- and long-term bilateral negotiated transactions and the green electricity transaction contract electricity.

[0108] The non-fossil energy electricity consumption that cannot be traced is calculated based on the scale of the enterprise's non-traceable electricity and the proportion of non-fossil energy electricity in the power market. The non-fossil energy electricity consumption that cannot be traced by the enterprise is:

[0109]

[0110] In the above formula, Q NT,g Q is the enterprise's non-fossil energy electricity consumption of type g that cannot be traced. all is the total electricity consumption of the enterprise, Q mar,g , Q mar,g,T are the total non-fossil energy electricity in the electricity market, the total traceable non-fossil energy electricity, and Q mar , Q mar,T They are the total traded electricity and the total traceable electricity in the electricity market respectively.

[0111] 2.2. Based on the total non-fossil energy electricity consumption of the enterprise, the enterprise electricity carbon emissions accounting is carried out, including:

[0112]

[0113] In the above formula, C ele is the carbon emissions of the enterprise's electricity, κ i 、W i 、V i are the carbon dioxide emission factor, the amount used for thermal power generation, and the average lower calorific value of fossil fuel i consumed for power generation in the region where the enterprise is located, respectively. fire , Q bio They are respectively the thermal power generation and biomass power generation in the area where the enterprise is located.

[0114] This typical enterprise consumed 137.51 million kWh of electricity in June. The electricity consumption structure is shown in Table 1. The carbon dioxide emission factor of fossil energy electricity in the local area is 0.84 kgCO2 / kWh. It can be calculated that the enterprise's electricity carbon emissions in June were 87,600 tons.

[0115] Table 1 Electricity consumption structure of a typical enterprise in June

[0116]

[0117] 3. Considering the enterprise's electricity purchase cost and electricity carbon emissions, a low-carbon electricity optimization model is constructed. The objective function of this model is:

[0118]

[0119] In the above formula, α(T) is the dynamic weight coefficient of the T period, and 0.3≤α(T)≤0.7, to avoid the dynamic weight coefficient from being too extreme. j,T , G net,T are the electricity consumption of the jth type of electricity source in period T and the enterprise's online electricity consumption, λ j ,λ net are the electricity price of the jth type of electricity source and the enterprise response electricity price, N green,T is the number of green certificates generated by the enterprise's renewable energy power generation during period T, λ green is the price corresponding to the green certificate, X max 、X min are the highest and lowest electricity purchase costs in the history of the enterprise, C ele,T is the carbon emissions of electricity generated by the enterprise during period T, C max 、C min are the highest and lowest electricity carbon emissions in the history of the enterprise, c0 and c(T), respectively, are the annual average electricity carbon emission factor of the region where the enterprise is located and the average electricity carbon emission factor during period T. When c(T) increases, the carbon emission pressure increases, and α(T) decreases, focusing on emission reduction, while on the contrary, focusing on cost. k is the adjustment factor;

[0120] The constraints are:

[0121] Electricity demand balancing constraints:

[0122]

[0123] Power purchase budget constraints:

[0124]

[0125] Green Certificate Constraints:

[0126]

[0127] In the above formula, G self,T is the self-generated electricity of the enterprise during period T, Q need,T is the amount of electricity required to maintain normal operation of the enterprise during period T, B is the enterprise's electricity purchase budget, and N green,T is the number of green certificates generated by the enterprise’s renewable energy power generation during period T, G self,g,T is the self-generated electricity of the enterprise using the g-th non-fossil energy in period T.

[0128] 4. The improved NSGA-III algorithm is used to solve the enterprise low-carbon electricity consumption optimization model and obtain the enterprise low-carbon electricity consumption optimization plan.

[0129] Specifically include:

[0130] 4.1. Adjust the enterprise's preference for electricity purchase cost and carbon emissions through the dynamic weight coefficient α(T), and adjust the priority of electricity purchase cost and carbon emissions in real time based on the fluctuation of the electricity carbon emission factor.

[0131] 4.2. Randomly generate a power consumption plan X=(Q1, Q2, ..., Q j ,N green ) to initialize the population, where Q j is the power consumption of the jth power source type, N green The number of green certificates generated for the company’s new energy power generation;

[0132] 4.3. Determine the corresponding dynamic weight coefficients based on the different needs of enterprises (including low-carbon priority, cost optimization, and low-carbon and cost balance), calculate the objective function value, and generate a non-uniform set of dynamic reference points under different needs in the normalized target space, so that the distribution of dynamic reference points matches the demand priority;

[0133] 4.4. Perform non-dominated sorting on the population, calculate the distance between each individual and the dynamic reference point, and use the distance from the individual to the nearest dynamic reference point as the normalized value representing the superiority of the individual in the normalized target space;

[0134] 4.5. Use crossover and mutation operations to generate new individuals and form a progeny population. have:

[0135]

[0136] In the above formula, β is the cross-distribution parameter, is the parent individual;

[0137] 4.6. Merge the parent and child populations, perform local normalization on the newly added individuals, re-perform non-dominated sorting, and select the optimal solution corresponding to the dynamic reference point based on different requirements to update the population. Local normalization includes updating the ideal point through a sliding window (i.e., taking the minimum value of the historical and newly added individuals) and local screening of extreme points (calculating the ASF value of the newly added individuals and comparing the ASF values ​​to update only the extreme points affected by the newly added individuals). The ideal point and extreme point represent the minimum and maximum values ​​of each target, respectively. For details, see the adaptive normalization process of population individuals.

[0138] 4.7. Adjust the position of the dynamic reference point according to the new population distribution density, including:

[0139] 4.7.1. Calculate the individual distribution density of the new population in each dimension in the normalized target space:

[0140]

[0141] In the above formula, α i is the individual distribution density of the i-th dimension, N i is the number of individuals in the i-th dimension in the normalized target space, and N is the total number of individuals in the normalized target space.

[0142] 4.7.2. Determine the adjusted dynamic reference point position according to the following formula:

[0143]

[0144] In the above formula, They are the dynamic reference point positions before and after adjustment respectively.

[0145] 4.8. Repeat 4.4-4.7 until the algorithm converges, extract the non-dominated solutions in the Pareto front, and obtain the low-carbon electricity optimization plan under different enterprise needs.

[0146] Example 2:

[0147] An enterprise electricity optimization system that takes into account the allocation of electricity carbon emission responsibilities, such as Figure 2 As shown, it includes an electricity source classification module, an electricity carbon emission accounting module, a model building module, and a model solving module.

[0148] The electricity source classification module is used to classify the enterprise's electricity sources into traceable electricity and non-traceable electricity based on the enterprise's electricity consumption data, and includes a classification unit and a verification unit.

[0149] The classification unit is used to collect the enterprise's electricity consumption data for classification processing to obtain the enterprise's electricity consumption data dynamic labels in each time period. The enterprise's electricity consumption data dynamic labels D(T) in time period T are:

[0150] D(T)={Q(T),T,u}

[0151] In the above formula, Q(T) is the total electricity consumption in time period T, T is the timestamp, and u is the type of electricity source, including self-generated and self-consumed electricity, medium- and long-term bilateral negotiated transaction electricity, green electricity transaction electricity, priority electricity purchase, grid agency purchase of electricity, retail market-based transaction, and wholesale market-based transaction. Among them, self-generated and self-consumed electricity, medium- and long-term bilateral negotiated transaction electricity, and green electricity transaction electricity are traceable electricity, while priority electricity purchase, grid agency purchase of electricity, retail market-based transaction, and wholesale market-based transaction are non-traceable electricity.

[0152] The verification unit verifies the electricity consumption data whose source type is traceable electricity using the following verification function:

[0153]

[0154] In the above formula, V self (T), V con (T), V green (T) are the verification functions of self-generated and self-used electricity, medium- and long-term bilateral negotiated transaction electricity, and green electricity transaction electricity, respectively. self,T , G net,T are the self-generated electricity and grid-connected electricity of the enterprise in period T, G con is the contracted electricity, [t a ,t b ] is the effective time window of the contract, G green The amount of electricity in green electricity trading contracts.

[0155] The electricity carbon emissions accounting module is used to calculate the enterprise's electricity carbon emissions based on the principle of accurate accounting of traceable electricity and average allocation of non-traceable electricity. The specific calculation formula is as follows:

[0156]

[0157] Q T,g =(G self,g -G net,g )+G con,g +G green

[0158]

[0159] In the above formula, C ele is the carbon emissions of corporate electricity, Q all is the total electricity consumption of the enterprise, Q T,g , Q NT,g are the enterprise's traceable g-th non-fossil energy electricity consumption and the non-traceable g-th non-fossil energy electricity consumption, κ i 、W i 、V i are the carbon dioxide emission factor, the amount used for thermal power generation, and the average lower calorific value of fossil fuel i consumed for power generation in the region where the enterprise is located, respectively. fire , Q bio They are the thermal power generation and biomass power generation in the area where the enterprise is located, G self,g , G net,g are the self-generated electricity and grid-connected electricity of the enterprise’s g-th non-fossil energy source, G con,g , G green They are the non-fossil energy electricity of the gth type in the medium- and long-term bilateral negotiated transactions of enterprises and the electricity of the green electricity transaction contract, Qmar,g , Q mar,g,T are the total non-fossil energy electricity in the electricity market, the total traceable non-fossil energy electricity, and Q mar , Q mar,T They are the total traded electricity and the total traceable electricity in the electricity market respectively.

[0160] The model building module is used to comprehensively consider the enterprise's electricity purchase cost and electricity carbon emissions to build an enterprise low-carbon electricity optimization model. The objective function of the model includes:

[0161]

[0162] In the above formula, α(T) is the dynamic weight coefficient of the T period, Q j,T 、Gnet, T are the electricity consumption of the jth type of electricity source in period T and the enterprise's online electricity consumption, λ j ,λ net are the electricity price of the jth type of electricity source and the enterprise response price, N green,T is the number of green certificates generated by the enterprise's renewable energy power generation during period T, λ green is the price corresponding to the green certificate, X max 、X min are the highest and lowest electricity purchase costs in the history of the enterprise, C ele,T is the carbon emissions of electricity generated by the enterprise during period T, C max 、C min are the highest and lowest electricity carbon emissions in the company’s history, c0 and c(T), respectively, are the annual average electricity carbon emission factor of the company’s region and the average electricity carbon emission factor during period T, and k is the adjustment factor.

[0163] Constraints include:

[0164]

[0165] In the above formula, G self,T is the self-generated electricity of the enterprise during period T, Q need,T is the amount of electricity required to maintain normal operation of the enterprise during period T, B is the enterprise's electricity purchase budget, and N green,T is the number of green certificates generated by the enterprise’s renewable energy power generation during period T, G self,g,T is the self-generated electricity of the enterprise using the g-th non-fossil energy in period T.

[0166] The model solving module is used to solve the enterprise low-carbon electricity optimization model using the improved NSGA-III algorithm to obtain the enterprise low-carbon electricity optimization plan. The specific process of the improved NSGA-III algorithm includes:

[0167] A. Randomly generate a power consumption plan X=(Q1, Q2, ..., Qj ,N green ) to initialize the population, where Q j is the power consumption of the jth power source type, N green The number of green certificates generated for the company’s new energy power generation;

[0168] B. Determine the corresponding dynamic weight coefficient based on the different needs of the enterprise, calculate the objective function value, and generate a non-uniform set of dynamic reference points under different needs in the normalized target space, so that the distribution of dynamic reference points matches the demand priority;

[0169] C. Perform non-dominated sorting on the population, calculate the distance between each individual and the dynamic reference point, and use the distance from the individual to the nearest dynamic reference point as the normalized value representing the superiority of the individual in the normalized target space;

[0170] D. Use crossover and mutation operations to generate new individuals and form a progeny population;

[0171] E. Merge the parent population and the child population, perform local normalization on the newly added individuals, re-perform non-dominated sorting, and select the corresponding optimal solution based on the dynamic reference points of different requirements to update the population. The local normalization process includes:

[0172] Update the ideal point through the sliding window so that the updated ideal point is the minimum value of the history and the newly added individual;

[0173] Locally screen extreme points, including calculating the ASF value of the newly added individuals and comparing the ASF values ​​to update only the extreme points affected by the newly added individuals;

[0174] F. Adjust the position of the dynamic reference point according to the distribution density of the new population, including:

[0175] F1. Calculate the individual distribution density of the new population in each dimension in the normalized target space:

[0176]

[0177] In the above formula, α i is the individual distribution density of the i-th dimension, N i is the number of individuals in the i-th dimension in the normalized target space, and N is the total number of individuals in the normalized target space;

[0178] F2. Determine the adjusted dynamic reference point position according to the following formula:

[0179]

[0180] In the above formula, They are the dynamic reference point positions before and after adjustment respectively;

[0181] G. Repeat CF until the algorithm converges, extract the non-dominated solutions in the Pareto front, and obtain the low-carbon electricity optimization plan under different enterprise needs.

Claims

1. A method for optimizing enterprise electricity consumption taking into account the allocation of electricity carbon emission responsibilities, characterized in that: The method comprises: S1. Based on the enterprise's electricity consumption data, the enterprise's electricity sources are divided into traceable electricity and non-traceable electricity; S2. Calculate the company's electricity carbon emissions based on the principle of accurate accounting of traceable electricity and equal allocation of non-traceable electricity; S3. Comprehensively consider the enterprise's electricity purchase cost and electricity carbon emissions, and build an enterprise low-carbon electricity optimization model; S4. Solve the enterprise's low-carbon electricity consumption optimization model and obtain the enterprise's low-carbon electricity consumption optimization plan.

2. The method for optimizing enterprise electricity consumption taking into account the allocation of electricity carbon emission responsibilities according to claim 1, characterized in that: S2 uses the following formula to calculate the company's electricity carbon emissions: Q T ,g=(Gself,g-Gnet,g)+Gcon,g+Ggreen In the above formula, Cele is the carbon emission of the enterprise electricity, Qall is the total electricity consumption of the enterprise, and Q T ,g,Q NT , g are the enterprise's traceable g-th non-fossil energy power consumption and non-traceable g-th non-fossil energy power consumption, respectively, i 、W i , Vi are the carbon dioxide emission factor, the amount used for thermal power generation, and the average lower calorific value of fossil fuel i consumed for power generation in the enterprise's region, respectively. Qfire and Qbio are the thermal power generation and biomass power generation in the enterprise's region, respectively. Gself,g and Gnet,g are the self-generated electricity and grid-connected electricity of the g-th non-fossil energy of the enterprise, respectively. Gcon,g and Ggreen are the g-th non-fossil energy electricity volume of the enterprise's medium- and long-term bilateral negotiated transactions and the green electricity transaction contract electricity volume, respectively. Qmar,g, Qmar,g, T are the total non-fossil energy electricity in the electricity market, the total traceable non-fossil energy electricity, and Q mar 、Qmar, T They are the total traded electricity and the total traceable electricity in the electricity market respectively.

3. A method for optimizing enterprise electricity consumption taking into account the allocation of electricity carbon emission responsibilities according to claim 1 or 2, characterized in that: S1 uses dynamic labeling technology to divide the company's electricity sources into traceable electricity and non-traceable electricity, including: S11. Collect the enterprise's electricity consumption data and classify it to obtain the enterprise's electricity consumption data dynamic labels in each time period. The enterprise's electricity consumption data dynamic labels D(T) in time period T are: D(T)={Q(T),T,u} In the above formula, Q(T) is the total electricity consumption in the T period, T is the timestamp, and u is the type of electricity source, including self-generated and self-consumed electricity, medium- and long-term bilateral negotiated transactions, green power transactions, priority electricity purchase, grid agency purchase, retail market transactions, and wholesale market transactions. Among them, self-generated and self-consumed electricity, medium- and long-term bilateral negotiated transactions, and green power transactions are traceable electricity, while priority electricity purchase, grid agency purchase, retail market transactions, and wholesale market transactions are not traceable electricity. S12. Use the following verification function to verify the electricity consumption data whose source type is traceable electricity: In the above formula, V self (T), V con (T), V green (T) are the verification functions of self-generated and self-used electricity, medium- and long-term bilateral negotiated transaction electricity, and green electricity transaction electricity, respectively. self,T 、Gnet, T They are the self-generated electricity and grid-connected electricity of the enterprise in period T, Gcon is the contracted electricity, [ta, tb] is the effective time window of the contract, and Ggreen is the green electricity transaction contracted electricity.

4. A method for optimizing enterprise electricity consumption taking into account the allocation of electricity carbon emission responsibilities according to claim 1 or 2, characterized in that: In S3, the objective function of the enterprise low-carbon electricity optimization model includes: In the above formula, α(T) is the dynamic weight coefficient of the T period, Q j,T 、Gnet, T are the electricity consumption of the jth type of electricity source in period T and the enterprise's online electricity consumption, λ j ,λ net are the electricity price of the jth type of electricity source and the enterprise response price, N green,T is the number of green certificates generated by the enterprise's renewable energy power generation during period T, λ green is the price corresponding to the green certificate, X max 、X min These are the highest and lowest electricity purchase costs in the company's history, respectively. T is the carbon emissions of electricity generated by the enterprise during period T, C max 、C min are the highest and lowest electricity carbon emissions in the history of the enterprise, c0 and c(T) are the annual average electricity carbon emission factor of the region where the enterprise is located and the average electricity carbon emission factor during period T, respectively; k is the adjustment factor; Constraints include: In the above formula, G self,T is the self-generated electricity of the enterprise during period T, Qneed, T is the amount of electricity required to maintain normal business operations during period T, B is the electricity purchase budget of the enterprise, Ngreen, T is the number of green certificates generated by the enterprise's renewable energy power generation during period T, Gself,g, T is the self-generated electricity of the enterprise using the g-th non-fossil energy in period T.

5. A method for optimizing enterprise electricity consumption taking into account the allocation of electricity carbon emission responsibilities according to claim 1 or 2, characterized in that: The S4 uses the improved NSGA-III algorithm to solve the enterprise electricity optimization model, including: S41. Randomly generate an electricity consumption plan X = (Q1, Q2, ..., Qj, Ngreen) that meets the constraints to initialize the population, where Qj is the electricity consumption of the jth electricity source type, and Ngreen is the number of green certificates generated by the enterprise's renewable energy power generation; S42. Determine corresponding dynamic weight coefficients according to different needs of the enterprise, calculate the objective function value, generate non-uniform dynamic reference point sets under different needs in the normalized target space, and match the distribution of dynamic reference points with the demand priority; S43, performing non-dominated sorting on the population, calculating the distance between each individual and the dynamic reference point, and using the distance from the individual to the nearest dynamic reference point as a normalized value representing the superiority of the individual in the normalized target space; S44, using crossover and mutation operations to generate new individuals and form a progeny population; S45. Merge the parent population and the child population, perform local normalization on the newly added individuals, re-perform non-dominated sorting, and select the corresponding optimal solution based on the dynamic reference points of different requirements, and update the population. The local normalization includes updating the ideal point through the sliding window and locally screening the extreme points. S46. Adjust the position of the dynamic reference point according to the distribution density of the new population; S47. Repeat S43-S46 until the algorithm converges, extract the non-dominated solutions in the Pareto front, and obtain the low-carbon electricity optimization plan under different enterprise needs.

6. The method for optimizing enterprise electricity consumption taking into account the allocation of electricity carbon emission responsibilities according to claim 5, characterized in that: In said S45, the updated ideal point is the minimum value of the historical and newly added individuals; Local screening of extreme points means: calculating the ASF value of the newly added individuals and comparing the ASF values ​​to update only the extreme points affected by the newly added individuals; The S46 includes: S461. Calculate the individual distribution density of the new population in each dimension in the normalized target space: In the above formula, α i is the individual distribution density of the i-th dimension, N i is the number of individuals in the i-th dimension in the normalized target space, and N is the total number of individuals in the normalized target space; S462. Determine the adjusted dynamic reference point position according to the following formula: In the above formula, They are the dynamic reference point positions before and after adjustment respectively.

7. An enterprise electricity optimization system taking into account the allocation of electricity carbon emission responsibilities, characterized in that: The system includes an electricity source classification module, an electricity carbon emission accounting module, a model construction module, and a model solving module; The power source classification module is used to classify the power sources of the enterprise into traceable power and non-traceable power based on the enterprise's power consumption data; The electricity carbon emission accounting module is used to calculate the enterprise's electricity carbon emissions based on the principle of accurate accounting of traceable electricity and average allocation of non-traceable electricity; The model building module is used to comprehensively consider the enterprise's electricity purchase cost and electricity carbon emissions to build an enterprise low-carbon electricity optimization model; The model solving module is used to solve the enterprise low-carbon electricity consumption optimization model and obtain the enterprise low-carbon electricity consumption optimization plan.

8. The enterprise electricity optimization system taking into account the allocation of electricity carbon emission responsibilities according to claim 7 is characterized in that: The electricity carbon emissions calculation module calculates the enterprise's electricity carbon emissions based on the following formula: Q T ,g=(Gself,g-Gnet,g)+Gcon,g+Ggreen In the above formula, Cele is the carbon emission of the enterprise electricity, Qall is the total electricity consumption of the enterprise, and Q T ,g,Q NT , g are the enterprise's traceable g-th non-fossil energy power consumption and non-traceable g-th non-fossil energy power consumption, respectively, i 、W i , Vi are the carbon dioxide emission factor, the amount used for thermal power generation, and the average lower calorific value of fossil fuel i consumed for power generation in the enterprise's region, respectively. Qfire and Qbio are the thermal power generation and biomass power generation in the enterprise's region, respectively. Gself,g and Gnet,g are the self-generated electricity and grid-connected electricity of the g-th non-fossil energy of the enterprise, respectively. Gcon,g and Ggreen are the g-th non-fossil energy electricity volume of the enterprise's medium- and long-term bilateral negotiated transactions and the green electricity transaction contract electricity volume, respectively. Qmar,g, Qmar,g, T are the total non-fossil energy electricity in the electricity market, the total traceable non-fossil energy electricity, and Q mar 、Qmar, T They are the total traded electricity and the total traceable electricity in the electricity market; The objective function of the enterprise low-carbon electricity optimization model includes: In the above formula, α(T) is the dynamic weight coefficient of the T period, Q j,T 、Gnet, T are the electricity consumption of the jth type of electricity source in period T and the enterprise's online electricity consumption, λ j ,λ net are the electricity price of the jth type of electricity source and the enterprise response electricity price, N green,T is the number of green certificates generated by the enterprise's renewable energy power generation during period T, λ green is the price corresponding to the green certificate, X max 、X min These are the highest and lowest electricity purchase costs in the company's history, respectively. T is the carbon emissions of electricity generated by the enterprise during period T, C max 、C min are the highest and lowest electricity carbon emissions in the history of the enterprise, c0 and c(T) are the annual average electricity carbon emission factor of the region where the enterprise is located and the average electricity carbon emission factor during period T, respectively; k is the adjustment factor; Constraints include: In the above formula, G self,T is the self-generated electricity of the enterprise in period T, Qneed, T is the amount of electricity required to maintain normal business operations during period T, B is the electricity purchase budget of the enterprise, Ngreen, T is the number of green certificates generated by the enterprise's renewable energy power generation during period T, Gself,g, T is the self-generated electricity of the enterprise using the g-th non-fossil energy in period T.

9. The enterprise electricity optimization system taking into account the allocation of electricity carbon emission responsibilities according to claim 7 or 8, characterized in that: The electricity source classification module uses dynamic labeling technology to classify the enterprise's electricity sources into traceable electricity and non-traceable electricity, including a classification unit and a verification unit; The classification unit is used to collect the enterprise's electricity consumption data for classification processing to obtain the enterprise's electricity consumption data dynamic labels in each time period. The enterprise's electricity consumption data dynamic labels D(T) in time period T are: D(T)={Q(T),T,u} In the above formula, Q(T) is the total electricity consumption in the T period, T is the timestamp, and u is the type of electricity source, including self-generated and self-consumed electricity, medium- and long-term bilateral negotiated transactions, green power transactions, priority electricity purchase, grid agency purchase, retail market transactions, and wholesale market transactions. Among them, self-generated and self-consumed electricity, medium- and long-term bilateral negotiated transactions, and green power transactions are traceable electricity, while priority electricity purchase, grid agency purchase, retail market transactions, and wholesale market transactions are not traceable electricity. The verification unit verifies the electricity consumption data whose source type is traceable electricity using the following verification function: In the above formula, V self (T), V con (T), V green (T) are the verification functions of self-generated and self-used electricity, medium- and long-term bilateral negotiated transaction electricity, and green electricity transaction electricity, respectively. self,T 、Gnet, T They are the self-generated electricity and grid-connected electricity of the enterprise in period T, Gcon is the contracted electricity, [ta, tb] is the effective time window of the contract, and Ggreen is the green electricity transaction contracted electricity.

10. The enterprise electricity optimization system taking into account the allocation of electricity carbon emission responsibilities according to claim 7 or 8, characterized in that: The model solving module uses the improved NSGA-III algorithm to solve the enterprise electricity optimization model. The specific process of the algorithm includes: A. Randomly generate a power consumption plan X = (Q1, Q2, ..., Qj, Ngreen) that meets the constraints to initialize the population, where Qj is the power consumption of the jth power source type and Ngreen is the number of green certificates generated by the enterprise's renewable energy power generation; B. Determine the corresponding dynamic weight coefficient based on the different needs of the enterprise, calculate the objective function value, and generate a non-uniform set of dynamic reference points under different needs in the normalized target space, so that the distribution of dynamic reference points matches the demand priority; C. Perform non-dominated sorting on the population, calculate the distance between each individual and the dynamic reference point, and use the distance from the individual to the nearest dynamic reference point as the normalized value representing the superiority of the individual in the normalized target space; D. Use crossover and mutation operations to generate new individuals and form a progeny population; E. Merge the parent population and the child population, perform local normalization on the newly added individuals, re-perform non-dominated sorting, and select the corresponding optimal solution based on the dynamic reference points of different requirements to update the population. The local normalization process includes: Update the ideal point through the sliding window so that the updated ideal point is the minimum value of the history and the newly added individual; Locally screen extreme points, including calculating the ASF value of the newly added individuals and comparing the ASF values ​​to update only the extreme points affected by the newly added individuals; F. Adjust the position of the dynamic reference point according to the distribution density of the new population, including: F1, statistics the individual distribution density of the new population in each dimension in the normalized target space: In the above formula, α i is the individual distribution density of the i-th dimension, N i is the number of individuals in the i-th dimension in the normalized target space, and N is the total number of individuals in the normalized target space; F2. Determine the adjusted dynamic reference point position according to the following formula: In the above formula, They are the dynamic reference point positions before and after adjustment respectively; G. Repeat CF until the algorithm converges, extract the non-dominated solutions in the Pareto front, and obtain the low-carbon electricity optimization plan under different enterprise needs.

Images