A dual-carbon comprehensive energy consumption monitoring platform for smart parks

Abstract

The present invention discloses a dual-carbon comprehensive energy consumption monitoring platform applied to a smart park, including a park carbon emission sensing unit, an emission reduction potential prediction unit, a maximum emission reduction potential sensing unit, a carbon emission output value ratio sensing unit, and a sorting unit for obtaining the emission reduction potential release of all enterprises in the park, and sorting them by size to obtain the release ranking after sorting; a platform recommendation unit for determining the top M enterprises in the release ranking after sorting as demonstration emission reduction enterprises; the present invention optimizes the potential prediction model so that the data of the current year can reflect the carbon emission prediction emission reduction potential in real time, solving the problem of being unable to obtain historical data for the next year. The carbon emission prediction emission reduction potential of the enterprise is equated with the actual emission reduction potential, and the emission reduction potential release can be calculated and optimized in real time, so that the park management can make scientific and reasonable adjustments to the emission reduction measures in the current year, greatly improving the real-time nature of the emission reduction effect.

Description

Technical Field

[0001] The present invention relates to an energy consumption monitoring platform, in particular to a dual-carbon comprehensive energy consumption monitoring platform applied to a smart park. Background Art

[0002] In modern society, with the acceleration of industrialization and urbanization, greenhouse gas emissions have become one of the main factors affecting climate change. Especially in smart parks, the energy consumption and carbon emissions of enterprises are the core content of carbon emission monitoring in the park. In order to achieve the "dual carbon" goals, namely carbon peak and carbon neutrality, enterprises in the park need to effectively manage and optimize energy consumption and carbon emissions. Therefore, how to scientifically and accurately monitor and predict the carbon emissions of enterprises and evaluate their emission reduction potential has become the key to achieving the "dual carbon" goals of the park. With the increasing severity of global climate change issues, promoting the development of a low-carbon economy has become a global consensus.

[0003] Patent No. CN202310996479.8 discloses a method and system for monitoring and calculating enterprise energy consumption and carbon emissions based on big data, which is based on the consistency and reaction lag of the changes in the carbon emissions data and the coal usage data.

[0004] Then, in the prior art, the carbon emission monitoring method mainly relies on the calculation of historical data and actual data of the next year. However, since the data of the next year cannot be obtained in the current year, the calculation and optimization of emission reduction potential face lags. This lag makes it impossible to evaluate the release of emission reduction potential in real time, which affects the real-time adjustment and decision-making of emission reduction measures. For example, without the carbon emission data of the next year, the actual emission reduction potential of the enterprise cannot be accurately predicted, which leads to large errors in the calculation and optimization decision of the release of emission reduction potential. Such defects limit the accuracy and real-time performance of carbon emission energy efficiency optimization. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention provides a dual-carbon comprehensive energy consumption monitoring platform applied to smart parks, which solves the technical problems raised in the background technology through carbon emission prediction and emission reduction potential.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0007] A dual-carbon comprehensive energy consumption monitoring platform applied to smart parks, including:

[0008] The park carbon emission sensing unit is used to obtain the park carbon emission data dictionary;

[0009] The emission reduction potential prediction unit is used to extract the enterprise area carbon emission data dictionary from the park carbon emission data dictionary, input it into the pre-trained optimization potential prediction model, and output the enterprise's carbon emission prediction reduction potential;

[0010] The maximum emission reduction potential sensing unit is used to preset the maximum carbon emission reduction potential of enterprises in the park;

[0011] The expression of the maximum carbon emission reduction potential is: DCRP max,j =γ·PO j ;

[0012] DCRP max,j Indicates the maximum carbon emission reduction potential of an enterprise, PO j represents the production output value of the enterprise during the carbon emission period, γ represents the proportional coefficient of the production output value and the maximum carbon emission reduction potential;

[0013] A carbon emission output value ratio sensing unit, used for extracting the carbon emission output value ratio of an enterprise from the park carbon emission data dictionary;

[0014] The emission reduction potential release degree calculation unit is used to determine the emission reduction potential release degree according to the maximum carbon emission actual emission reduction potential, the carbon emission predicted emission reduction potential and the carbon emission output value ratio;

[0015] The expression of the emission reduction potential release degree is:

[0016]

[0017] r j Indicates the degree of emission reduction potential released by the enterprise, DCRP mj Represents the carbon emission reduction potential output by the model, CER j represents the carbon emission output value ratio, α is the influence coefficient of the carbon emission output value ratio on the release degree of emission reduction potential;

[0018] The sorting unit is used to obtain the emission reduction potential release degree of all enterprises in the park, sort them by size, and obtain the sorted release degree ranking;

[0019] The platform recommendation unit is used to determine the top M enterprises in the release degree ranking after sorting as demonstration emission reduction enterprises.

[0020] In some embodiments, the step of the park carbon emission sensing unit acquiring the park carbon emission data dictionary includes:

[0021] S1-1. Define the enterprise identifier and collect several carbon emission categories and their carbon emissions for the enterprise in the current year based on the enterprise identifier;

[0022] S1-2, define each carbon emission category as a key, and define the associated carbon emission amount as a value, to construct an enterprise key-value pair;

[0023] S1-3, integrating several of the enterprise key-value pairs to generate an enterprise carbon emission data dictionary;

[0024] The expression of enterprise carbon emission data dictionary is:

[0025] d(E i ) represents enterprise E i The enterprise carbon emission data dictionary is used to store the enterprise E i Several key-value pairs of E i represents the enterprise identifier of the i-th enterprise in the park, C i,j Represents Enterprise E i The jth carbon emission category, V i,j Represents Enterprise E i The carbon emissions associated with the jth carbon emission category, k i Represents Enterprise E i Total number of carbon emission categories, Represents Enterprise E i A key-value pair of the j-th carbon emission category and the associated carbon emission amount.

[0026] In some embodiments, the step of the park carbon emission sensing unit acquiring the park carbon emission data dictionary includes:

[0027] S1-4, define each enterprise identifier as a key, and define the associated enterprise carbon emission data dictionary as a value, to form a park key-value pair;

[0028] S1-5, integrating several park key-value pairs to generate a park carbon emission data dictionary;

[0029] The expression of the park carbon emission data dictionary is:

[0030] D represents the carbon emission data dictionary of enterprises in the park. i ) stores the obtained park carbon emission data dictionary, and H represents the total number of enterprise carbon emission data dictionaries.

[0031] In some of the embodiments, an enterprise identifier is defined, and based on the enterprise identifier, several carbon emission categories and carbon emission amounts of the enterprise in the current year are collected, including:

[0032] S1-1-1. Define several carbon emission categories; carbon emission categories include: direct production energy, indirect production energy and measured emission categories;

[0033] S1-1-2. Obtain the direct production energy, indirect production energy and carbon emissions of measured emission categories respectively;

[0034] S1-1-3. The expression for obtaining carbon emissions from direct production energy is:

[0035]

[0036] CE pd Represents the carbon emissions of direct production energy, P i represents the consumption of the enterprise under the i-th category of direct production energy, F d,i represents the carbon emission factor of the i-th type of direct production energy, N d Indicates the type of energy used for direct production in the production process of the enterprise;

[0037] S1-1-4. The expression for obtaining carbon emissions from indirect production energy is:

[0038]

[0039] CE nd Indicates the carbon emissions from indirect production energy use, N i represents the consumption of the enterprise under the i-th category of indirect production energy, F nd,i represents the carbon emission factor of the i-th type of indirect production energy, N nd It refers to the type of indirect energy used by an enterprise in the production process.

[0040] In some embodiments, the modeling step of the optimization potential prediction model includes:

[0041] S2-1. Calculate the total carbon emissions CE of the enterprise in each historical year based on the direct production energy consumption, indirect production energy consumption and carbon emissions of measured emission categories in the historical year. j ;

[0042]

[0043] CE m represents the carbon emissions of the measured emission categories;

[0044] S2-2. Obtain the production value of the enterprise in the historical year;

[0045] S2-3. Calculate the ratio of total carbon emissions to production output value in each historical year to determine the carbon emission output value ratio;

[0046] The expression of the carbon emission output value ratio is:

[0047] CER jrepresents the carbon emission output value ratio of enterprise j, PO j represents the production value of enterprise j during the carbon emission period;

[0048] S2-4. Calculate the actual carbon emission reduction potential of the current year compared with the previous year in adjacent historical years;

[0049] S2-5. Convert the enterprise carbon emission data dictionary into an input vector, define the corresponding enterprise's real carbon emission reduction potential as a target label, and construct an optimization potential sample;

[0050] S2-6. Obtain optimization potential samples of several historical years and construct an optimization potential training set;

[0051] S2-7. Use the neural network model to perform self-supervised pre-training on the optimization potential training set until the optimization potential prediction model is obtained through iteration.

[0052] In some of the embodiments, calculating the actual carbon emission reduction potential of the current year compared to the previous year in adjacent historical years includes:

[0053] S2-4-1. Obtain the total carbon emissions, total energy consumption and total production value of the enterprise in the previous year;

[0054] S2-4-2. Obtain the total carbon emissions, total energy consumption and total production value of the enterprise in the current year;

[0055] S2-4-3. Calculate the actual carbon emission reduction potential based on the total carbon emissions, total energy consumption and total production value of the enterprise in the previous year, and the total carbon emissions, total energy consumption and total production value of the enterprise in the current year;

[0056] The expression for calculating the actual carbon emission reduction potential is:

[0057]

[0058] DCRP j Indicates the actual reduction potential of carbon emissions in the previous year, CE j (t-1), E j (t-1), PO j (t-1) represents the total carbon emissions, total energy consumption and total production value of the enterprise in the previous year; CE j (t), E j (t), PO j (t) represent the total carbon emissions, total energy consumption and total production value of the enterprise in the current year respectively.

[0059] In some of the embodiments, self-supervised pre-training of a neural network model on an optimization potential training set includes:

[0060] S2-7-1. Extract a number of optimization potential samples from the optimization potential training set in batches and input them into the neural network model;

[0061] S2-7-2, treating the input vector in the optimization potential sample as a characteristic variable, and outputting the carbon emission prediction reduction potential through forward propagation;

[0062] S2-7-3. Define a potential loss function, and calculate the potential loss value of the predicted carbon emission reduction potential and the actual carbon emission reduction potential through the potential loss function;

[0063] The expression of the potential loss function is:

[0064] L represents the potential loss value, DCRP pj DCRP represents the predicted carbon emission reduction potential. j represents the actual carbon emission reduction potential of the previous year, and N is the number of optimization potential samples participating in the training;

[0065] S2-7-4, perform back propagation to update the model parameters of the neural network model and output the carbon emission reduction potential for the next round;

[0066] S2-7-5. Repeat S2-7-2 to S2-7-4 until the potential loss function is minimized.

[0067] In some embodiments, the step of determining the emission reduction potential release degree by the emission reduction potential release degree calculation unit includes:

[0068] S5-1. Calculate the potential change value of the maximum carbon emission actual reduction potential and the carbon emission predicted reduction potential;

[0069] S5-2. Calculate the unit potential change value of the potential change value compared to the actual emission reduction potential of the maximum carbon emission;

[0070] S5-3, calculating the negative correlation product of the unit potential change value and the carbon emission output value ratio;

[0071] S5-4. Define the negative correlation product as the emission reduction potential release degree.

[0072] The present invention provides a dual-carbon comprehensive energy consumption monitoring platform applied to smart parks, which has the following beneficial effects:

[0073] By using historical data to train and optimize the potential prediction model, the data of the current year can reflect the carbon emission reduction potential in real time, solving the problem of not being able to obtain historical data for the next year. The carbon emission reduction potential of the enterprise is equated with the actual reduction potential, and the release degree of reduction potential can be calculated and optimized in real time, so that the park management can make scientific and reasonable adjustments to the emission reduction measures in the current year, greatly improving the real-time nature of the emission reduction effect.

[0074] By optimizing the potential prediction model, the system can accurately capture the relationship between an enterprise's carbon emission categories and carbon emissions, and predict the enterprise's emission reduction potential; the optimization adjustment of the enterprise's emission reduction potential not only depends on the enterprise's carbon emissions, but also combines the deep trends in historical data, providing more precise guidance for the park's carbon emission energy efficiency optimization.

[0075] By incorporating the relationship between the carbon emission output value ratio and the emission reduction potential release degree into the prediction model, the park's emission reduction measures are further optimized. Through the real-time calculation of the emission reduction potential release degree, the park can quickly identify enterprises with high emission reduction potential release degrees. In addition, the emission reduction potential release degree, as an important indicator for evaluating the carbon emission optimization effect of enterprises, can help the park management formulate more scientific emission reduction strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 This is a structural block diagram of a dual-carbon comprehensive energy consumption monitoring platform applied to a smart park in the present invention;

[0077] Figure 2 This is a monitoring flow chart of a dual-carbon comprehensive energy consumption monitoring platform applied to a smart park in the present invention;

[0078] Figure 3 A modeling flow chart of the optimization potential prediction model of the present invention;

[0079] Figure 4 This is a flow chart for obtaining the degree of release of emission reduction potential described in the present invention. DETAILED DESCRIPTION

[0080] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0081] Example 1: Please refer to Figure 1-Figure 2 The present invention provides a dual-carbon comprehensive energy consumption monitoring platform applied to a smart park, comprising:

[0082] The park carbon emission sensing unit is used to obtain the park carbon emission data dictionary;

[0083] In this embodiment, carbon emission data are collected and sorted from all enterprises in the park to build a unified data dictionary, which contains the carbon emissions of each enterprise under different carbon emission categories (such as production energy consumption, measured emissions, etc.).

[0084] The emission reduction potential prediction unit is used to extract the enterprise area carbon emission data dictionary from the park carbon emission data dictionary, input it into the pre-trained optimization potential prediction model, and output the enterprise's carbon emission prediction reduction potential;

[0085] The maximum emission reduction potential sensing unit is used to preset the maximum carbon emission reduction potential of enterprises in the park;

[0086] The expression of the maximum carbon emission reduction potential is: DCRP max,j =γ·PO j ;

[0087] DCRP max,j Indicates the maximum carbon emission reduction potential of an enterprise, PO j represents the production value of the enterprise during the carbon emission period, and γ represents the ratio coefficient between the production value and the maximum carbon emission reduction potential. Generally speaking, it can be estimated based on industry data or historical experience;

[0088] A carbon emission output value ratio sensing unit, used for extracting the carbon emission output value ratio of an enterprise from the park carbon emission data dictionary;

[0089] The emission reduction potential release degree calculation unit is used to determine the emission reduction potential release degree according to the maximum carbon emission actual emission reduction potential, the carbon emission predicted emission reduction potential and the carbon emission output value ratio;

[0090] The expression of the emission reduction potential release degree is:

[0091]

[0092] R j Indicates the degree of emission reduction potential released by the enterprise, DCRP mj Represents the carbon emission reduction potential output by the model, CER j represents the carbon emission output value ratio, α is the influence coefficient of the carbon emission output value ratio on the release degree of emission reduction potential;

[0093] Specifically, γ·PO j -DCRP mj is the emission reduction potential, that is, the theoretical maximum emission reduction potential γ·PO jThe gap between the predicted carbon emission reduction potential of the enterprise and the output of the model. The larger the gap, the greater the room for the enterprise to optimize carbon emissions; and 1-α·CER j It reflects the impact of the carbon emission output value ratio on the release of emission reduction potential, and controls the impact through the size of the impact coefficient α. The higher the carbon emission output value ratio, the greater the carbon emission intensity of the enterprise, and the smaller the release of emission reduction potential (therefore, the impact of emission reduction potential release is negatively correlated).

[0094] The release of emission reduction potential is a ratio that indicates the proportion of carbon emission reduction that can be achieved by the enterprise under the current emission reduction optimization strategy. It represents the gap between the maximum emission reduction potential and the predicted carbon emission reduction potential, which is equivalent to evaluating the actual effect and potential of the enterprise's current emission reduction measures. The carbon emission output value ratio as an influencing factor shows the relationship between the enterprise's carbon emissions and output value. Enterprises with high carbon emission output value ratios may need more emission reduction optimization to achieve the same emission reduction potential.

[0095] For example, if the emission reduction potential release degree of an enterprise is close to 1, it means that the enterprise's carbon emission optimization potential has been fully released and the emission reduction effect is good. If the emission reduction potential release degree of an enterprise is close to 0, it means that the enterprise's carbon emission optimization potential has not been effectively released and more emission reduction measures are needed to tap the potential.

[0096] Through the maximum carbon emission optimization potential, the predicted carbon emission optimization potential and the carbon emission output value ratio, we can calculate the emission reduction potential release degree. This indicator reflects the emission reduction effect and potential of the enterprise under the existing emission reduction strategy.

[0097] The sorting unit is used to obtain the emission reduction potential release degree of all enterprises in the park, sort them by size, and obtain the sorted release degree ranking;

[0098] The platform recommendation unit is used to determine the top M enterprises in the release degree ranking after sorting as demonstration emission reduction enterprises.

[0099] The steps for the park carbon emission sensing unit to obtain the park carbon emission data dictionary include:

[0100] S1-1. Define the enterprise identifier and collect several carbon emission categories and their carbon emissions for the enterprise in the current year based on the enterprise identifier;

[0101] S1-2, define each carbon emission category as a key, and define the associated carbon emission amount as a value, to construct an enterprise key-value pair;

[0102] S1-3. Integrate several of the enterprise key-value pairs to generate an enterprise carbon emission data dictionary.

[0103] The expression of enterprise carbon emission data dictionary is:

[0104] D(E i ) represents enterprise E i The enterprise carbon emission data dictionary is used to store the enterprise E i Several key-value pairs of E i represents the enterprise identifier of the i-th enterprise in the park, C i,j Represents Enterprise E i The jth carbon emission category, V i,j Represents Enterprise E i The carbon emissions associated with the jth carbon emission category, k i Show Enterprise E i Total number of carbon emission categories, Represents Enterprise E i A key-value pair of the j-th carbon emission category and the associated carbon emission amount.

[0105] Specifically, this step first defines an enterprise identifier and collects the enterprise's carbon emission data for the current year based on the identifier, that is, obtains several carbon emission categories and their carbon emissions from the enterprise. These carbon emission categories include but are not limited to production energy, indirect production energy, and measured emission data. Each carbon emission category contains the carbon emissions associated with it. Next, each carbon emission category is defined as a key, and the carbon emissions associated with it are used as the value to construct an enterprise key-value pair.

[0106] Furthermore, the step of the park carbon emission sensing unit obtaining the park carbon emission data dictionary includes:

[0107] S1-4, define each enterprise identifier as a key, and define the associated enterprise carbon emission data dictionary as a value, to form a park key-value pair;

[0108] S1-5. Integrate several of the park key-value pairs to generate a park carbon emission data dictionary.

[0109] The expression of the park carbon emission data dictionary is:

[0110] D represents the carbon emission data dictionary of enterprises in the park. i ) stores the obtained park carbon emission data dictionary, and H represents the total number of enterprise carbon emission data dictionaries (i.e., the number of enterprises in the park).

[0111] Specifically, in steps S1-4 and S1-5, the process of constructing the park carbon emission data dictionary is further refined. S1-4 defines each enterprise identifier as a key and the associated enterprise carbon emission data dictionary as a value, thereby forming a park key-value pair. In the park carbon emission data dictionary, the key of each key-value pair is the enterprise identifier (for example: enterprise A, enterprise B, enterprise C, etc.), and the value is the enterprise's carbon emission data dictionary, which contains all the carbon emission categories of the enterprise and their corresponding carbon emissions. Specifically, the carbon emission data dictionary of enterprise A contains categories such as "electricity consumption", "natural gas consumption", "steam consumption" and their carbon emissions, while enterprises B and C have similar carbon emission data dictionaries. These key-value pairs are distinguished by enterprise identifiers, and the carbon emission data of each enterprise can be managed independently.

[0112] Next, in step S1-5, several park key-value pairs are integrated to generate a park carbon emission data dictionary. The purpose of this step is to aggregate the carbon emission data dictionaries of all enterprises in the park to form a global park carbon emission data dictionary. The data dictionary of each enterprise uses its identifier as the key and the enterprise's carbon emission data as the value, ultimately forming a collection containing the carbon emission information of all enterprises.

[0113] Furthermore, the step S1-1 specifically includes:

[0114] S1-1-1. Define several carbon emission categories; carbon emission categories include: direct production energy, indirect production energy and measured emission categories;

[0115] S1-1-2. Obtain the direct production energy, indirect production energy and carbon emissions of measured emission categories respectively;

[0116] S1-1-3. The expression for obtaining carbon emissions from direct production energy is:

[0117]

[0118] CE pd Represents the carbon emissions of direct production energy, P i represents the consumption of the enterprise under the i-th category of direct production energy, F d,i represents the carbon emission factor of the i-th type of direct production energy, N d Indicates the type of energy used for direct production in the production process of the enterprise;

[0119] Specifically, the calculation steps for direct production energy carbon emissions are:

[0120] First, determine the type of energy (such as electricity, natural gas, steam, etc.) and the amount of energy consumed in production activities.e (Unit: kWh) and natural gas consumption P g (Unit: m 3 );

[0121] Then, the carbon emission per unit of direct production energy is defined as the carbon emission factor. For example, the carbon emission factor F for electricity e (Unit: kg CO2 / kWh), natural gas carbon emission factor F g (Unit: kg CO2 / m 3 );

[0122] Then, based on the determined direct production energy consumption and the corresponding carbon emission factor, calculate the carbon emission of each direct production energy;

[0123] Finally, the carbon emissions from all direct production energy consumption are summed up, and the total amount obtained represents the carbon emissions from direct production energy consumption.

[0124] S1-1-4. The expression for obtaining carbon emissions from indirect production energy is:

[0125]

[0126] CE nd Indicates the carbon emissions from indirect production energy use, N i represents the consumption of the enterprise under the i-th category of indirect production energy, F nd,i represents the carbon emission factor of the i-th type of indirect production energy, N nd Indicates the type of indirect energy used by the enterprise in the production process;

[0127] Specifically, the indirect production energy refers to energy consumption that is not directly used in the production process, such as office electricity, lighting, air conditioning, etc. The energy consumption of the indirect production energy is usually related to the daily operation activities of the enterprise and does not involve the production process, but it still has an impact on carbon emissions. i It refers to each type of energy consumption used by the enterprise in the non-production process. Such energy consumption is collected through automated metering equipment. The above indirect production energy data can be obtained by setting up energy meters on relevant non-production equipment, including water and electricity meters, gas meters (boilers, etc.) of non-production equipment. As mentioned above, the carbon emission factor of indirect production energy refers to the carbon emissions per unit of indirect production energy.

[0128] The measured emission category carbon emissions CE m The calculation can be accurately measured through automated monitoring equipment. The automated monitoring equipment can collect emission samples from the enterprise's emission sources through a sampling pipeline and send the samples to an analysis cabinet or infrared analyzer for analysis, thereby accurately obtaining the enterprise's actual carbon emission data.

[0129] Through the above embodiments, various carbon emission data of enterprises and parks are formed into a dictionary data structure that is easy to store, identify and extract, so that the platform can easily call and operate the energy consumption monitoring data of the park during subsequent data references.

[0130] Example 2: See Figures 1 to 4 The technical solution of this embodiment 2 is different from that of embodiment 1 in that a modeling step of optimizing the potential prediction model is disclosed, and the modeling step includes:

[0131] S2-1. Calculate the total carbon emissions CE of the enterprise in each historical year based on the direct production energy consumption, indirect production energy consumption and carbon emissions of measured emission categories in the historical year. j ;

[0132]

[0133] CE m represents the carbon emissions of the measured emission categories;

[0134] S2-2. Obtain the production value of the enterprise in the historical year;

[0135] S2-3. Calculate the ratio of total carbon emissions to production output value in each historical year to determine the carbon emission output value ratio;

[0136] The expression of the carbon emission output value ratio is:

[0137] CER j represents the carbon emission output value ratio of enterprise j, PO j represents the production value of enterprise j during the carbon emission period;

[0138] S2-4. Calculate the actual carbon emission reduction potential of the current year compared with the previous year in adjacent historical years;

[0139] S2-5. Convert the enterprise carbon emission data dictionary into an input vector, define the corresponding enterprise's real carbon emission reduction potential as a target label, and construct an optimization potential sample;

[0140] S2-6. Obtain optimization potential samples of several historical years and construct an optimization potential training set;

[0141] S2-7. Use the neural network model to perform self-supervised pre-training on the optimization potential training set until the optimization potential prediction model is obtained through iteration.

[0142] Specifically, the enterprise carbon emission data dictionary contains multiple key-value pairs, each of which represents a carbon emission category (such as "direct production energy", "indirect production energy", "measured emissions") and its corresponding carbon emissions. When these key-value pairs are converted into input features, the carbon emissions (value part) of each key-value pair will be used as a dimension in the input vector. Therefore, the enterprise carbon emission data dictionary can be regarded as a feature vector containing multiple carbon emission categories and their carbon emissions, and each dimension corresponds to a specific carbon emission category.

[0143] For example, if the company's carbon emissions data dictionary includes three categories: "direct production energy", "indirect production energy" and "measured emissions", then the input vector will include the carbon emissions values ​​of these three categories, forming a multi-dimensional feature vector. The dimensions of this vector correspond to different carbon emission categories, and the value of each dimension is the carbon emissions corresponding to that category. When constructing the training set, each company's carbon emissions data dictionary will be converted into an input vector containing multi-dimensional features, with the actual carbon emissions reduction potential (that is, the carbon emissions that the company can reduce through optimization) as the target label. Of course, during the construction process, the carbon emissions must be standardized or normalized to avoid the magnitude of certain features being too large or too small.

[0144] Furthermore, the step S2-4 specifically includes:

[0145] S2-4-1. Obtain the total carbon emissions, total energy consumption and total production value of the enterprise in the previous year;

[0146] S2-4-2. Obtain the total carbon emissions, total energy consumption and total production value of the enterprise in the current year;

[0147] S2-4-3. Calculate the actual carbon emission reduction potential based on the total carbon emissions, total energy consumption and total production value of the enterprise in the previous year, and the total carbon emissions, total energy consumption and total production value of the enterprise in the current year;

[0148] The expression for calculating the actual carbon emission reduction potential is:

[0149]

[0150] DCRP j Indicates the actual reduction potential of carbon emissions in the previous year, CE j (t-1), E j (t-1), PO j (t-1) represents the total carbon emissions, total energy consumption and total production value of the enterprise in the previous year; CE j (t), E j(t), PO j (t) represent the total carbon emissions, total energy consumption and total production value of the enterprise in the current year respectively.

[0151] Specifically, if the carbon emissions and energy consumption of the current year have increased compared to the previous year, it may indicate that the company's emission reduction work has not been fully implemented and there is room for further optimization. If the carbon emissions and energy consumption of the current year have decreased, but the production output value has continued to grow, it indicates that the company's carbon emission intensity has decreased and the optimization potential may be large. Therefore, the real carbon emission reduction potential is actually an assessment of the room for companies to reduce future carbon emissions. Based on the company's historical data, it predicts how much carbon emissions the company can reduce under the optimized best practices.

[0152] Furthermore, the step S2-7 specifically includes:

[0153] S2-7-1. Extract a number of optimization potential samples from the optimization potential training set in batches and input them into the neural network model;

[0154] S2-7-2, treating the input vector in the optimization potential sample as a characteristic variable, and outputting the carbon emission prediction reduction potential through forward propagation;

[0155] S2-7-3. Define a potential loss function, and calculate the potential loss value of the predicted carbon emission reduction potential and the actual carbon emission reduction potential through the potential loss function;

[0156] The expression of the potential loss function is:

[0157] L represents the potential loss value, DCRP pj DCRP represents the predicted carbon emission reduction potential. j represents the actual carbon emission reduction potential of the previous year, and N is the number of optimization potential samples participating in the training;

[0158] The potential loss function is used to measure the deviation between the predicted carbon emission reduction potential and the actual carbon emission reduction potential. The expression of this loss function is the mean square error (MSE).

[0159] S2-7-4, perform back propagation to update the model parameters of the neural network model and output the carbon emission reduction potential for the next round;

[0160] Calculate the gradient of the loss function relative to each parameter of the model, and use the gradient descent algorithm to update the parameters of the neural network. This process is to adjust the network weights through back propagation to reduce the value of the potential loss function so that the predicted value gradually approaches the actual emission reduction potential. This process is carried out in each round of training to optimize the performance of the model.

[0161] S2-7-5. Repeat S2-7-2 to S2-7-4 until the potential loss function is minimized.

[0162] Specifically, steps S2-7-2 to S2-7-4 are executed repeatedly until the loss function reaches a minimum value, that is, the model converges. Through multiple rounds of training and parameter adjustment, the neural network continuously learns and improves the prediction results, making the carbon emission reduction potential more accurate.

[0163] In this embodiment, the step of determining the emission reduction potential release degree by the emission reduction potential release degree calculation unit includes:

[0164] S5-1. Calculate the potential change value between the maximum actual carbon emission reduction potential and the predicted carbon emission reduction potential; the potential change value represents the gap between the enterprise's current emission reduction strategy and the maximum possible emission reduction potential.

[0165] S5-2. Calculate the unit potential change value of the potential change value compared to the actual emission reduction potential of the maximum carbon emission;

[0166] The unit potential change value measures the actual progress of the optimization of emission reduction potential, that is, the ratio between the emission reduction potential currently achieved by the enterprise and the maximum possible emission reduction potential.

[0167] S5-3, calculating the negative correlation product of the unit potential change value and the carbon emission output value ratio;

[0168] According to the previously calculated unit potential change value and the negative correlation with the carbon emission output value ratio, a further product calculation is performed. The higher the carbon emission output value ratio, the greater the carbon emission intensity of the enterprise, which may mean that the enterprise needs more efforts to achieve the same emission reduction potential. Therefore, the product of the unit potential change value and the carbon emission output value ratio is used as an important regulating factor for the release of emission reduction potential. The negative correlation product reflects the inhibitory effect of the carbon emission output value ratio on the release of emission reduction potential, that is, the higher the carbon emission output value ratio, the lower the release of emission reduction potential.

[0169] S5-4. Define the negative correlation product as the emission reduction potential release degree.

[0170] Finally, the negative correlation product is defined as the emission reduction potential release degree, which indicates the emission reduction potential that has been released by the enterprise under the existing emission reduction measures. The emission reduction potential release degree reflects the gap between the maximum emission reduction potential and the actual emission reduction potential of the enterprise, and is the core indicator for evaluating the effectiveness of the enterprise's emission reduction measures. The emission reduction potential release degree close to 1 indicates that the enterprise has fully released its emission reduction potential and the emission reduction effect is good; while close to 0 indicates that the emission reduction potential has not been effectively released, and the enterprise may need to further strengthen its emission reduction measures.

[0171] In this embodiment, the emission reduction potential release degree is constructed through historical data, and the potential law between the carbon emission category and carbon emission amount of each enterprise and the emission reduction potential release degree of the current year is learned through the neural network model, so that the carbon emission predicted emission reduction potential is predicted through the annual carbon emission category and carbon emission amount of the current year, without waiting until the next year to calculate the real carbon emission reduction potential formed by the enterprise after the relevant emission reduction measures. And the carbon emission predicted emission reduction potential is equal to the real carbon emission reduction potential of the current year, which can facilitate the release degree of the enterprise's emission reduction potential, thereby obtaining the strength of the enterprises in the park to implement emission reduction measures, and finally realizing more scientific carbon emission energy consumption monitoring.

[0172] The above embodiments can be implemented in whole or in part by software, hardware, firmware or other arbitrary combinations. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions can be transmitted from a website site, computer, server or data center by wired (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center.

[0173] The computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that includes one or more available media. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium may be a solid-state hard disk.

[0174] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division of a waterway underwater terrain change analysis system and method. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0175] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A dual-carbon comprehensive energy consumption monitoring platform applied to smart parks, characterized in that: include: The park carbon emission sensing unit is used to obtain the park carbon emission data dictionary; The emission reduction potential prediction unit is used to extract the enterprise area carbon emission data dictionary from the park carbon emission data dictionary, input it into the pre-trained optimization potential prediction model, and output the enterprise's carbon emission prediction reduction potential; The maximum emission reduction potential sensing unit is used to preset the maximum carbon emission reduction potential of enterprises in the park; The expression of the maximum carbon emission reduction potential is: DCRP max,j =γ·PO j ; DCRP max,j Indicates the maximum carbon emission reduction potential of an enterprise, PO j represents the production output value of the enterprise during the carbon emission period, γ represents the proportional coefficient of the production output value and the maximum carbon emission reduction potential; A carbon emission output value ratio sensing unit, used for extracting the carbon emission output value ratio of an enterprise from the park carbon emission data dictionary; The emission reduction potential release degree calculation unit is used to determine the emission reduction potential release degree according to the maximum carbon emission actual emission reduction potential, the carbon emission predicted emission reduction potential and the carbon emission output value ratio; The expression of the emission reduction potential release degree is: R j Indicates the degree of emission reduction potential released by the enterprise, DCRP mj Represents the carbon emission reduction potential output by the model, CER j represents the carbon emission output value ratio, α is the influence coefficient of the carbon emission output value ratio on the release degree of emission reduction potential; The sorting unit is used to obtain the emission reduction potential release degree of all enterprises in the park, sort them by size, and obtain the sorted release degree ranking; The platform recommendation unit is used to determine the top M enterprises in the release degree ranking after sorting as demonstration emission reduction enterprises.

2. According to claim 1, a dual-carbon comprehensive energy consumption monitoring platform applied to a smart park is characterized in that: The steps for the park carbon emission sensing unit to obtain the park carbon emission data dictionary include: S1-1. Define the enterprise identifier and collect several carbon emission categories and their carbon emissions for the enterprise in the current year based on the enterprise identifier; S1-2, define each carbon emission category as a key, and define the associated carbon emission amount as a value, to construct an enterprise key-value pair; S1-3, integrating several of the enterprise key-value pairs to generate an enterprise carbon emission data dictionary; The expression of enterprise carbon emission data dictionary is: D(E i ) represents enterprise E i The enterprise carbon emission data dictionary is used to store the enterprise E i Several key-value pairs of E i represents the enterprise identifier of the i-th enterprise in the park, C i,j Represents Enterprise E i The jth carbon emission category, V i,j Represents Enterprise E i The carbon emissions associated with the jth carbon emission category, k i Show Enterprise E i Total number of carbon emission categories, Represents Enterprise E i A key-value pair of the j-th carbon emission category and the associated carbon emission amount.

3. According to claim 2, a dual-carbon comprehensive energy consumption monitoring platform applied to a smart park is characterized in that: The steps of the park carbon emission sensing unit obtaining the park carbon emission data dictionary include: S1-4, define each enterprise identifier as a key, and define the associated enterprise carbon emission data dictionary as a value, to form a park key-value pair; S1-5, integrating several park key-value pairs to generate a park carbon emission data dictionary; The expression of the park carbon emission data dictionary is: D represents the carbon emission data dictionary of enterprises in the park. i ) stores the obtained park carbon emission data dictionary, and H represents the total number of enterprise carbon emission data dictionaries.

4. According to claim 2, a dual-carbon comprehensive energy consumption monitoring platform applied to a smart park is characterized in that: Define corporate identifiers and collect several carbon emission categories and their carbon emissions for the current year based on the corporate identifiers, including: S1-1-1. Define several carbon emission categories; carbon emission categories include: direct production energy, indirect production energy and measured emission categories; S1-1-2. Obtain the direct production energy, indirect production energy and carbon emissions of measured emission categories respectively; S1-1-3. The expression for obtaining carbon emissions from direct production energy is: CE pd Represents the carbon emissions of direct production energy, P i represents the consumption of the enterprise under the i-th category of direct production energy, F d,i represents the carbon emission factor of the i-th type of direct production energy, N d Indicates the type of energy used for direct production in the production process of the enterprise; S1-1-4. The expression for obtaining carbon emissions from indirect production energy is: CE nd Indicates the carbon emissions from indirect production energy use, N i represents the consumption of the enterprise under the i-th category of indirect production energy, F nd,i represents the carbon emission factor of the i-th type of indirect production energy, N nd It refers to the type of indirect energy used by an enterprise in the production process.

5. According to claim 1, a dual-carbon comprehensive energy consumption monitoring platform applied to a smart park is characterized in that: The modeling step of the optimization potential prediction model comprises: S2-1. Calculate the total carbon emissions CE of the enterprise in each historical year based on the direct production energy consumption, indirect production energy consumption and carbon emissions of measured emission categories in the historical year. j ; CE m represents the carbon emissions of the measured emission categories; S2-2. Obtain the production value of the enterprise in the historical year; S2-3. Calculate the ratio of total carbon emissions to production output value in each historical year to determine the carbon emission output value ratio; The expression of the carbon emission output value ratio is: CER j represents the carbon emission output value ratio of enterprise j, PO j represents the production value of enterprise j during the carbon emission period; S2-4. Calculate the actual carbon emission reduction potential of the current year compared with the previous year in adjacent historical years; S2-5. Convert the enterprise carbon emission data dictionary into an input vector, define the corresponding enterprise's real carbon emission reduction potential as a target label, and construct an optimization potential sample; S2-6. Obtain optimization potential samples of several historical years and construct an optimization potential training set; S2-7. Use the neural network model to perform self-supervised pre-training on the optimization potential training set until the optimization potential prediction model is obtained through iteration.

6. According to claim 5, a dual-carbon comprehensive energy consumption monitoring platform applied to a smart park is characterized in that: Calculate the actual carbon emission reduction potential of the current year compared with the previous year in adjacent historical years, including: S2-4-1. Obtain the total carbon emissions, total energy consumption and total production value of the enterprise in the previous year; S2-4-2. Obtain the total carbon emissions, total energy consumption and total production value of the enterprise in the current year; S2-4-3. Calculate the actual carbon emission reduction potential based on the total carbon emissions, total energy consumption and total production value of the enterprise in the previous year, and the total carbon emissions, total energy consumption and total production value of the enterprise in the current year; The expression for calculating the actual carbon emission reduction potential is: DCRP j Indicates the actual reduction potential of carbon emissions in the previous year, CE j (t-1), E j (t-1), PO j (t-1) represents the total carbon emissions, total energy consumption and total production value of the enterprise in the previous year; CE j (t), E j (t), PO j (t) represent the total carbon emissions, total energy consumption and total production value of the enterprise in the current year respectively.

7. According to claim 5, a dual-carbon comprehensive energy consumption monitoring platform applied to a smart park is characterized in that: Self-supervised pre-training of a neural network model on an optimization potential training set, including: S2-7-1. Extract a number of optimization potential samples from the optimization potential training set in batches and input them into the neural network model; S2-7-2, treating the input vector in the optimization potential sample as a characteristic variable, and outputting the carbon emission prediction reduction potential through forward propagation; S2-7-3. Define a potential loss function, and calculate the potential loss value of the predicted carbon emission reduction potential and the actual carbon emission reduction potential through the potential loss function; The expression of the potential loss function is: L represents the potential loss value, DCRP pj DCRP represents the predicted carbon emission reduction potential. j represents the actual carbon emission reduction potential of the previous year, and N is the number of optimization potential samples participating in the training; S2-7-4, perform back propagation to update the model parameters of the neural network model and output the carbon emission reduction potential for the next round; S2-7-5. Repeat S2-7-2 to S2-7-4 until the potential loss function is minimized.

8. According to claim 1, a dual-carbon comprehensive energy consumption monitoring platform applied to a smart park is characterized in that: The step of determining the emission reduction potential release degree by the emission reduction potential release degree calculation unit includes: S5-1. Calculate the potential change value of the maximum carbon emission actual reduction potential and the carbon emission predicted reduction potential; S5-2. Calculate the unit potential change value of the potential change value compared to the actual emission reduction potential of the maximum carbon emission; S5-3, calculating the negative correlation product of the unit potential change value and the carbon emission output value ratio; S5-4. Define the negative correlation product as the emission reduction potential release degree.

Images