Carbon emissions accounting system and method for urban water supply plants and water conservation and carbon reduction for large water users

Abstract

The present invention relates to the field of ecological and environmental protection technology, and discloses a system and method for calculating carbon emissions of urban water supply plants and water-saving and carbon reduction of large water users. The method comprises real-time collection of data of water supply plants and large water users, obtaining and fusing artificial design features and machine learning features through preprocessing and feature extraction; calculating the carbon footprint of the water supply plant and the carbon emissions of large water users according to the fused features to obtain a total carbon emission accounting value; obtaining a total carbon emission forecast value based on short-term and long-term prediction models; setting a calculation threshold and a target value, implementing water-saving transformation and evaluating the carbon reduction effect when the threshold is exceeded; the present invention can accurately calculate and predict the carbon emissions of a water supply system by fusing artificial design features and machine learning features, combined with short-term and long-term prediction models, and provides a scientific basis and technical support for water-saving and carbon reduction of the water supply system.

Description

Technical Field

[0001] The present invention relates to the field of ecological environmental protection technology, and more specifically, to a system and method for calculating carbon emissions from urban water supply plants and water conservation and carbon reduction for large water users. Background Art

[0002] Carbon emissions in the water industry mainly come from urban water supply systems, including water supply plants and large water users. How to accurately calculate carbon emissions from water supply systems and formulate effective water-saving and carbon-reduction measures accordingly has become a key issue that needs to be addressed urgently.

[0003] The patent application with publication number CN116932998A discloses a method for calculating carbon emissions in regional water environment governance. This method identifies each target area within the carbon emission accounting boundary, obtains carbon emission information and carbon emission reduction information in each area, and calculates carbon emissions and carbon emission reductions based on emission source activity levels, carbon emission factors, emission reduction activity levels, and carbon reduction factors, thereby achieving quantitative accounting of carbon emissions in regional water environment governance. However, this method is mainly aimed at carbon accounting at the macro level of regional water environment governance, and does not involve carbon emission accounting for the main bodies within the water supply system. In addition, this method does not consider the formulation and evaluation of water-saving and carbon reduction measures.

[0004] The patent application with publication number CN115641026A discloses a method for calculating carbon emissions from sewage treatment plants. This method uses the life cycle assessment method, takes into account the carbon emissions of the entire operation process of the sewage treatment plant, and evaluates the emission reduction effects of three carbon emission reduction technologies, which can provide a basis for improving the carbon emissions of sewage treatment plants. However, this method mainly focuses on the carbon accounting of a single entity and does not consider the carbon emission accounting of multiple entities in the water supply system. In addition, this method does not consider the prediction of carbon emissions and the dynamic optimization of water-saving and carbon reduction measures.

[0005] In summary, existing technologies lack a refined accounting method for carbon emissions from multiple entities within water supply systems; carbon emission accounting is disconnected from water-saving and carbon reduction measures, failing to form a closed decision-making loop; and carbon emission forecasts are insufficiently considered, making it difficult to judge future trends. Summary of the invention

[0006] In order to overcome the above-mentioned defects of the prior art, the present invention provides a system and method for calculating carbon emissions of urban water supply plants and water-saving and carbon reduction for large water users. The method can collect data of water supply plants and large water users in real time, obtain artificial design features and machine learning features through preprocessing and feature extraction, and merge them, so as to finely calculate the carbon emissions of water supply plants and large water users. On the basis of obtaining the carbon emission accounting value, the future carbon emission trend is predicted by combining short-term and long-term prediction models. When the set threshold is exceeded, the water-saving transformation is automatically triggered, and the transformation effect is regularly evaluated. This method is expected to break through the bottleneck of existing technologies and realize the accurate calculation of carbon emissions of water supply systems, future predictions, and closed-loop optimization of water saving and carbon reduction.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] Carbon emissions accounting methods for urban water supply plants and water conservation and carbon reduction for large water users, including:

[0009] Collect the first data of the water supply plant, the first data of the large water user and the electricity consumption data by item in real time to form the second data; pre-process the second data to obtain the third data; extract features from the third data to obtain artificial design features; obtain machine learning features based on the artificial design features; integrate the artificial design features with the machine learning features to construct the fourth data;

[0010] According to the fourth data, the carbon footprint of the water supply plant is calculated to obtain the first carbon emission data; according to the fourth data, the carbon emissions of large water users are calculated to obtain the second carbon emission data; the first carbon emission data and the second carbon emission data are input into a preset carbon emission accounting model to obtain a total carbon emission accounting value; according to the total carbon emission accounting value, as well as the pre-constructed short-term carbon emission prediction model and the long-term carbon emission prediction model, a total carbon emission prediction value is obtained;

[0011] Set accounting thresholds and target values ​​for total carbon emissions. When the calculated value of total carbon emissions exceeds the accounting threshold or the predicted value of total carbon emissions exceeds the target value, implement water-saving transformation of the water supply system. Regularly evaluate changes in carbon emissions before and after the water-saving transformation of the water supply system, and quantify the carbon reduction effect of the water-saving transformation.

[0012] Furthermore, the artificial design features include energy consumption intensity characteristic indicators of water supply plants and water use pattern characteristic indicators of large water users; the energy consumption intensity characteristic indicators of water supply plants include unit water supply power consumption, net water consumption and pipe network leakage rate; the water use pattern characteristic indicators of large water users include peak water consumption, water balance coefficient and minimum water consumption at night;

[0013] The method of obtaining machine learning features based on artificially designed features includes:

[0014] The artificially designed features are centralized, and the mean of each feature is shifted to 0 to obtain the original feature matrix X; the covariance matrix of the centralized features is calculated; the covariance matrix is ​​eigenvalue decomposed to obtain eigenvalues ​​and eigenvectors; the eigenvalues ​​are sorted from large to small, and the eigenvectors corresponding to the first k largest eigenvalues ​​are selected to form the transformation matrix P;

[0015] The transformation matrix P is used to multiply the original feature matrix X after centralization on the left to obtain the principal component feature matrix Z=PX after dimensionality reduction; the variance contribution rate of each principal component feature is calculated according to the eigenvalue; feature selection is performed on the principal component feature matrix Z, and the principal component features whose variance contribution rate is lower than the preset variance contribution rate threshold are eliminated to obtain machine learning features.

[0016] Further, the calculating the carbon footprint of the water supply plant according to the fourth data to obtain the first carbon emission data includes:

[0017] The carbon emissions of the electricity use of the water supply plant are calculated based on the unit water supply power consumption of the water supply plant in the fourth data and the real-time carbon emission factor of the power grid; the carbon emissions of the electricity use of the water supply plant are equal to the product of the unit water supply power consumption and the water supply and the real-time carbon emission factor of the power grid;

[0018] According to the water consumption of the water supply plant in the fourth data, calculate the carbon emissions of the water supply plant in the water intake link;

[0019] According to the pipe network leakage rate of the water supply plant in the fourth data, calculate the carbon emissions of the water supply plant's transmission and distribution links;

[0020] The carbon emissions from the electricity usage, water intake and transmission and distribution links of the water supply plant are summed up to obtain the carbon footprint of the water supply plant, forming the first carbon emission data.

[0021] Furthermore, the calculation of the carbon emissions of the water supply plant's transmission and distribution link according to the pipe network leakage rate of the water supply plant in the fourth data includes:

[0022] The process life cycle assessment method is used to calculate the total carbon emissions of the pipeline network throughout its life cycle. The total carbon emissions of the pipeline network throughout its life cycle are divided by the water supply to obtain the carbon intensity per unit water supply, which is then multiplied by the pipeline leakage rate to obtain the carbon emissions of the water supply plant's transmission and distribution links.

[0023] Furthermore, the carbon emissions of large water users are calculated based on the fourth data to obtain the second carbon emissions data, including:

[0024] Based on the peak water consumption of large water users in the fourth data, calculate the carbon emissions of transmission and distribution during peak hours;

[0025] According to the water balance coefficient of large water users in the fourth data, calculate the balanced carbon emissions of transmission and distribution energy consumption;

[0026] Based on the minimum nighttime water consumption of large water users in the fourth data, calculate the background leakage carbon emissions of the pipe network;

[0027] By adding together the carbon emissions from transmission and distribution during peak hours, the balanced carbon emissions from transmission and distribution energy consumption, and the background leakage carbon emissions of the pipeline network, we can obtain the carbon emission footprint of large water users and form the second carbon emission data.

[0028] Furthermore, the calculation of the carbon emissions of transmission and distribution during the peak period according to the peak water consumption of large water users in the fourth data includes:

[0029] Determine the peak load regulation operation mode of the water supply network; the peak load regulation operation mode of the water supply network includes two types: variable frequency speed regulation and pump start and stop;

[0030] According to the peak load regulation operation mode of the water supply network and the peak water consumption of large water users, the network energy consumption during peak hours is calculated; under the variable frequency speed regulation operation mode, the network energy consumption during peak hours is proportional to the square of the peak water consumption of large water users; under the start-stop pump operation mode, the network energy consumption during peak hours is proportional to the peak water consumption of large water users;

[0031] According to the real-time carbon emission factor of the power grid during peak hours, the energy consumption of the pipeline network during peak hours is converted into the transmission and distribution carbon emissions during peak hours.

[0032] Furthermore, obtaining the predicted value of total carbon emissions includes:

[0033] Obtain weather data and holiday data, build a short-term carbon emission prediction model, and predict the daily carbon emissions in the future short-term period based on the total carbon emission accounting value, weather data, holiday data and the short-term carbon emission prediction model;

[0034] Obtaining long-term forecast-related data, building a long-term carbon emission forecast model, and forecasting monthly carbon emissions in the future long-term period based on the total carbon emission accounting value, long-term forecast-related data and the long-term carbon emission forecast model; the long-term forecast-related data includes seasonal regularity data, statutory holiday data and facility maintenance plans;

[0035] The daily carbon emissions and monthly carbon emissions are superimposed and combined to form a forecast value of total carbon emissions.

[0036] The carbon emission accounting system for urban water supply plants and water saving and carbon reduction for large water users is used to implement the above-mentioned carbon emission accounting method for urban water supply plants and water saving and carbon reduction for large water users. The system includes:

[0037] Data acquisition module: used for real-time collection of first data of water supply plants, first data of large water users and itemized electricity consumption data to form second data; pre-processing the second data to obtain third data;

[0038] Feature extraction module: used to extract features from the third data to obtain artificially designed features; obtain machine learning features based on the artificially designed features; and fuse the artificially designed features with the machine learning features to construct fourth data;

[0039] Carbon emission accounting module: according to the fourth data, the carbon footprint of the water supply plant is calculated to obtain the first carbon emission data; according to the fourth data, the carbon emission of large water users is calculated to obtain the second carbon emission data; the first carbon emission data and the second carbon emission data are input into the preset carbon emission accounting model to obtain the total carbon emission accounting value;

[0040] Carbon emission prediction module: obtains the total carbon emission prediction value based on the total carbon emission accounting value, as well as the pre-built short-term carbon emission prediction model and long-term carbon emission prediction model;

[0041] Water-saving and carbon reduction assessment module: used to set the accounting threshold and target value of total carbon emissions. When the calculated value of total carbon emissions exceeds the accounting threshold or the predicted value of total carbon emissions exceeds the target value, water-saving transformation of the water supply system is implemented; regular assessment of carbon emission changes before and after the water-saving transformation of the water supply system is conducted to quantify the carbon reduction effect of the water-saving transformation.

[0042] An electronic device comprises a memory, a central processing unit and a computer program stored in the memory and executable on the central processing unit, wherein the central processing unit implements the above-mentioned carbon emission accounting method for urban water supply plants and water saving and carbon reduction accounting method for large water users when executing the computer program.

[0043] A computer-readable storage medium having a computer program stored thereon, which, when executed, implements the above-mentioned method for calculating carbon emissions of urban water supply plants and water-saving and carbon reduction for large water users.

[0044] Compared with the prior art, the present invention has the following beneficial effects:

[0045] The present invention realizes the accurate accounting and prediction of carbon emissions of water supply plants and large water users by integrating artificial design features with machine learning features. Combined with short-term and long-term prediction models, the system can accurately predict future carbon emission trends and help managers formulate countermeasures in advance. Through the real-time collection of key data by intelligent sensors, the system can detect anomalies in time and take water-saving transformation measures, thereby effectively reducing carbon emissions. The multi-level evaluation method not only analyzes the changes in carbon emissions before and after water-saving transformation, but also quantifies the carbon reduction effect of water-saving transformation, providing a basis for further optimization. The system integrates multi-source data, provides a reliable data foundation through preprocessing and feature extraction, and uses machine learning technology for feature selection and dimensionality reduction, which improves the efficiency and accuracy of feature analysis and supports more scientific decision-making. By analyzing energy consumption intensity and water use pattern indicators, the system can identify the weak links of the water supply system and put forward targeted energy-saving and management optimization suggestions. At the same time, by monitoring the leakage rate of the pipeline network, it helps managers to discover and repair pipeline network problems in time. The system effectively reduces carbon emissions during the water supply process, provides important support for achieving energy conservation and emission reduction goals, has significant environmental benefits, and reduces operating costs by optimizing water supply and water use efficiency, bringing economic benefits to related companies and users. In summary, the present invention significantly improves the water-saving and carbon-reduction capabilities of urban water supply plants and large water users, and makes important contributions to ecological and environmental protection. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0047] Figure 1 It is a principle flow chart of the carbon emission accounting method of urban water supply plants and water saving and carbon reduction accounting method of large water users in the present invention;

[0048] Figure 2 A flow chart of a method for obtaining machine learning features based on artificially designed features in the carbon emission accounting method for urban water supply plants and water saving and carbon reduction for large water users of the present invention;

[0049] Figure 3 A flow chart of a method for calculating the carbon footprint of a water supply plant and obtaining first carbon emission data in the method for calculating carbon emissions of a town water supply plant and water saving and carbon reduction for large water users of the present invention;

[0050] Figure 4 A flow chart of a method for calculating carbon emissions of large water users and obtaining second carbon emission data in the method for calculating carbon emissions of urban water supply plants and water saving and carbon reduction of large water users of the present invention;

[0051] Figure 5 A flow chart of a method for calculating the carbon emissions of urban water supply plants and water-saving and carbon-reduction accounting methods for large water users during peak hours;

[0052] Figure 6 A flow chart of a method for quantifying the carbon reduction effect of water-saving transformation in the carbon emission accounting method for urban water supply plants and water-saving and carbon reduction for large water users of the present invention;

[0053] Figure 7 This is a functional module diagram of the carbon emission accounting system for urban water supply plants and water saving and carbon reduction for large water users in the present invention. DETAILED DESCRIPTION

[0054] 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.

[0055] Example 1

[0056] See also Figure 1 As shown, this embodiment provides a method for calculating carbon emissions of urban water supply plants and water saving and carbon reduction of large water users, including:

[0057] Step S1000, collecting first data of the water supply plant, first data of large water users and item-by-item electricity consumption data in real time to form second data; preprocessing the second data to obtain third data; extracting features from the third data to obtain artificial design features; obtaining machine learning features based on the artificial design features; fusing the artificial design features with the machine learning features to construct fourth data;

[0058] Furthermore, step S1000 includes:

[0059] Step S1100, installing intelligent sensors at key nodes of water supply plants and large water users, collecting first data in real time, and collecting itemized electricity consumption data of water supply plants and large water users at the same time; using the first data and itemized electricity consumption data as second data; the first data includes first data of water supply plants and first data of large water users; the first data of water supply plants includes water flow data, pipe network pressure data and equipment power consumption data; the first data of large water users includes total water consumption data, itemized water consumption data and water distribution pump power consumption data;

[0060] Specifically, smart water meters are installed at key nodes such as water intakes, water purification workshops, and pump stations of water supply plants to collect water flow data in real time; pressure sensors are installed at key nodes such as trunk pipes and branch pipes of the water supply network to collect network pressure data in real time; smart electric meters are installed on high-energy-consuming equipment such as water supply pumps, fans, and dosing equipment to collect equipment power consumption data in real time; water flow data, network pressure data, and equipment power consumption data are used as the first data of the water supply plant; smart water meters are installed at the main water meter of large water users to collect total water consumption data in real time; sub-meters are installed in the main water use places of large water users such as bathing, kitchens, air conditioners, boilers, etc., to collect sub-item water consumption data in real time; smart electric meters are installed in the water distribution pump room of large water users to collect water distribution pump power consumption data in real time; total water consumption data, sub-item water consumption data, and water distribution pump power consumption data are used as the first data of large water users; through the smart electricity gateway of the water supply plant and large water users, the sub-item power consumption data of various pumps, fans, motors and other electrical equipment are collected, and the sub-item power consumption data are integrated with the first data to form the second data;

[0061] Both the first data and the second data are time series data.

[0062] The first data example of a water supply plant:

[0063] Timestamp, water flow (m³ / h), pipe network pressure (MPa), equipment power consumption (kWh);

[0064] 2023-01-01 00:00:00, 100.5, 0.3, 50.2;

[0065] 2023-01-01 01:00:00, 95.2, 0.28, 48.6;

[0066] 2023-01-01 02:00:00, 90.1, 0.32, 52.4;

[0067] Example of the first data of large water users:

[0068] Timestamp, total water consumption (m³), itemized water consumption (m³), power consumption of water distribution pump (kWh);

[0069] 2023-01-01 00:00:00, 20.6, 10.2|5.8|4.6, 5.5;

[0070] 2023-01-01 01:00:00, 18.4, 9.5|4.9|4.0, 5.1;

[0071] 2023-01-01 02:00:00, 22.5, 11.8|6.2|4.5, 5.9;

[0072] Examples of electricity consumption data for water supply plants and large water users:

[0073] Timestamp, power consumption of water supply pump (kWh), power consumption of fan (kWh), power consumption of dosing pump (kWh);

[0074] 2023-01-01 00:00:00, 25.5, 15.4, 5.2;

[0075] 2023-01-01 01:00:00, 27.2, 16.1, 5.8;

[0076] 2023-01-01 02:00:00, 23.9, 14.8, 4.9;

[0077] Step S1100 realizes the real-time collection of multiple parameters such as water volume, pressure, and power consumption by deploying smart sensors at key nodes of the water supply system. On the water supply plant side, the smart water meter accurately measures the inflow and outflow of water in each process link to evaluate the purification and distribution efficiency; the pressure sensor monitors the pressure change of the pipe network and promptly discovers the abnormality of the pipe network; the smart meter grasps the energy consumption level of the equipment and finds the energy saving space. On the large water user side, the smart water meter not only collects the total meter reading, but also installs sub-meters to monitor key water use places and depict the refined water use mode; at the same time, the power consumption of the water distribution pump is monitored to reflect the energy efficiency of the secondary water supply. In addition, the power consumption data of each item is collected through the power consumption smart gateway and refined to a single device, so as to realize the energy consumption monitoring of the whole process and all elements of the water supply, and provide data support for the subsequent carbon accounting. Multi-sensor collaborative collection makes the data granularity finer and more timely, which can fully perceive the status of the water supply system, timely optimize the operation plan, and tap the potential for energy saving and emission reduction. The second data gathers multi-source heterogeneous data of the water supply system, but still needs to be pre-processed and converted into analyzable structured data.

[0078] Step S1200, preprocessing the second data to obtain third data;

[0079] Further, step S1200 includes:

[0080] Step S1210, performing data cleaning on the second data to identify and remove abnormal values, missing values, and noise data therein;

[0081] Step S1220, normalizing the cleaned data, and uniformly mapping data of different dimensions to the range of 0-1;

[0082] Step S1230, performing data standardization on the normalized data, converting the data into a standard normal distribution with a mean of 0 and a variance of 1, so that data with different distributions are comparable.

[0083] Specifically, step S1200 uses a series of data preprocessing techniques to convert massive heterogeneous data into regular analytical data. The first is data cleaning, which removes outliers to eliminate measurement errors and outlier interference. Data cleaning methods include threshold filtering, 3σ criterion, distance clustering, etc.; threshold filtering sets upper and lower limits based on experience, and data beyond the range is considered as outliers and removed; the 3σ criterion assumes that the data follows a normal distribution, and data exceeding 3 times the standard deviation is an outlier; distance clustering finds outliers that are far from the cluster center through clustering. Data normalization methods include maximum and minimum value normalization, zero mean normalization, etc.; data normalization solves the comparability problem of data of different dimensions. Maximum and minimum value normalization linearly maps the data to [0,1] according to its maximum and minimum values; zero mean normalization shifts the mean of the data to 0. The normalized data is distributed on the same scale, which is convenient for comprehensive evaluation. Data standardization methods include Z-score, decimal scaling, etc.; data standardization further converts normalized data into a standard normal distribution. Z-score uses the difference between the data and the mean divided by the standard deviation to standardize, and decimal scaling standardizes by moving the decimal point of the data. Standardization makes the mean and variance of data with different distributions consistent and obey the same normal distribution, which is conducive to horizontal comparison of different objects. After preprocessing, multi-source data of water supply systems from different sources, different dimensions, and different distributions are integrated into a structured, analyzable, high-quality data set, which not only expands the sample size and improves statistical significance, but also lays the foundation for subsequent data mining. At the same time, regular structured data is easy to store, manage, and call, and a unified data warehouse can be established to achieve data standardization and centralized management.

[0084] Step S1300, performing feature engineering on the third data to construct fourth data;

[0085] Furthermore, step S1300 includes:

[0086] Step S1310, extracting features from the third data to obtain artificial design features; the artificial design features include energy consumption intensity characteristic indicators of water supply plants and water use pattern characteristic indicators of large water users; the energy consumption intensity characteristic indicators of water supply plants include unit water supply power consumption, net water consumption and pipe network leakage rate; the water use pattern characteristic indicators of large water users include peak water consumption, water balance coefficient and minimum water consumption at night;

[0087] Specifically, step S1310 makes full use of experts' prior knowledge of the physical mechanisms of the water supply system and manually designs a series of characteristic indicators with strong explanatory power and clear engineering significance. These characteristics are closely related to the two core issues of energy consumption intensity of water supply plants and water use patterns of large water users, and characterize the energy efficiency and water-saving and carbon reduction potential of water supply systems from the supply side and demand side respectively.

[0088] On the water supply plant side, based on the physical mechanism of the water supply system, characteristic indicators reflecting the energy consumption intensity of the water supply plant are designed, including:

[0089] a. Unit water supply power consumption, that is, the amount of electricity consumed by the water supply plant to produce 1 cubic meter of water. The calculation formula is the total power consumption divided by the total water supply;

[0090] b. Net water consumption, that is, the amount of raw water consumed by the water supply plant to produce 1 cubic meter of factory water. The calculation formula is the total water intake minus the total water supply, divided by the total water supply;

[0091] c. Pipeline network leakage rate, that is, the proportion of water lost in the water supply network due to leakage and other reasons to the total water supply. The calculation formula is the pipeline network leakage amount divided by the total water supply.

[0092] The unit water supply power consumption and water purification consumption reflect the power consumption and water consumption efficiency of water supply. The lower the unit water supply power consumption, the higher the energy efficiency of the water supply equipment; the lower the water purification consumption, the higher the raw water utilization rate and the more advanced the water purification process. The pipeline leakage rate reflects the airtightness and integrity of the water supply pipeline network. The lower the leakage rate, the fewer leakage points in the pipeline network and the higher the transmission and distribution efficiency. By analyzing these three indicators, we can comprehensively evaluate the energy efficiency and resource utilization level of the water supply plant, identify the weak links in the water supply system, formulate targeted energy-saving transformation and pipeline network repair plans, and tap the potential for water saving and carbon reduction. The fourth data is time series data.

[0093] On the large water users side, based on the physical mechanism of the water supply system, characteristic indicators reflecting the water use patterns of large water users are designed, including:

[0094] a. Peak water consumption, i.e. the average water consumption of large water users during the period of the day when water consumption is the highest, reflects the water supply pressure during peak water consumption;

[0095] b. Water balance coefficient, which is the standard deviation of water consumption of large water users in a day divided by the average value, reflects the volatility of users' water demand;

[0096] c. The minimum water consumption at night, that is, the average water consumption of large water users between 0:00 and 5:00 at night, reflects the background leakage level of the pipeline network.

[0097] Peak water consumption, water balance coefficient, and minimum water consumption at night characterize the characteristics of users' water use behavior. The greater the peak water consumption, the more concentrated the water demand and the greater the peak-shaving pressure of the water supply network; the higher the water balance coefficient, the greater the volatility of water demand and the greater the difficulty of regulating the water supply network; the greater the minimum water consumption at night, the higher the background leakage level of the network and the more prominent the leakage problem of the network. By analyzing these three indicators, we can gain insight into the water use patterns of large water users, predict peak water use periods, and optimize water supply scheduling plans. At the same time, we can guide large water users to use water at off-peak times, reduce peak water consumption, and balance water supply pressure; find hidden leakage points in the network, reduce background leakage levels, and improve network efficiency. Therefore, these characteristics are of great significance for formulating precise water-saving measures, guiding user behavior changes, and achieving water-saving and carbon reduction in the water supply system.

[0098] Then, statistical methods are used to screen the energy consumption intensity characteristic indicators of water supply plants and the water use pattern characteristic indicators of large water users; correlation analysis can be used to calculate the correlation coefficient matrix between characteristics. If the correlation coefficient of two characteristics is close to 1 or -1, it means that they are highly correlated and carry similar information. One of them can be eliminated to eliminate redundancy. Using the t test and F test in hypothesis testing, it can be determined whether the characteristics are significantly correlated with the target variable. The smaller the P value, the more significant the correlation. The irrelevant characteristics with large P values ​​can be eliminated.

[0099] In general, step S1310 uses a combination of expert knowledge and statistical tools to ensure the interpretability of feature engineering while taking into account the statistical significance of features, so that the selected feature set not only conforms to the physical laws of the water supply system, but also improves the learning and prediction effects of subsequent models, thereby better serving the application goals of water conservation and carbon reduction. These artificially designed features are intuitive and clear, and are convenient for engineering practice, but they are limited by expert experience and may not be able to discover the inherent laws of the data. Therefore, it is necessary to further use machine learning methods on the basis of artificial features to automatically extract high-order features implicit in the data, explore information not covered by artificial features, and build a more comprehensive and abstract feature representation.

[0100] Artificially designed features, examples are as follows:

[0101] Time period, unit water supply power consumption (kWh / m³), clean water consumption (m³ / m³), pipe network leakage rate (%), peak water consumption (m³), water balance coefficient, minimum water consumption at night (m³);

[0102] 2023-01-01, 0.521, 0.212, 0.102, 26.5, 0.214, 8.2;

[0103] 2023-01-02, 0.495, 0.203, 0.098, 28.2, 0.198, 8.9;

[0104] Step S1320, obtaining machine learning features based on the manually designed features;

[0105] Furthermore, if Figure 2 As shown, step S1320 includes:

[0106] Step S1321, centralize the artificially designed features, shift the mean of each feature to 0, and obtain the original feature matrix X;

[0107] Step S1322, calculating the covariance matrix of the centered features;

[0108] Step S1323, performing eigenvalue decomposition on the covariance matrix to obtain eigenvalues ​​and eigenvectors;

[0109] Step S1324, sort the eigenvalues ​​from large to small, and select the eigenvectors corresponding to the first k largest eigenvalues ​​to form the transformation matrix P;

[0110] Step S1325, multiplying the original feature matrix X after centralization by the transformation matrix P on the left to obtain the principal component feature matrix Z=PX after dimensionality reduction;

[0111] Step S1326, calculating the variance contribution rate of each principal component feature according to the eigenvalue;

[0112] Step S1327, perform feature selection on the principal component feature matrix Z, remove the principal component features whose variance contribution rate is lower than a preset variance contribution rate threshold, and obtain machine learning features.

[0113] Specifically, step S1320 uses the unsupervised feature extraction method of principal component analysis (PCA), which automatically learns the intrinsic structure of the data based on the artificially designed features and extracts a few principal component features that best represent the essence of the data. Principal component analysis converts linearly related variables in the original feature space into linearly independent principal component space through orthogonal transformation, thereby achieving feature compression and dimensionality reduction. The generated principal component features are mutually orthogonal and have the largest variance, which eliminates the multicollinearity between the original features and simplifies the model structure. At the same time, by setting the variance contribution rate threshold, the number of principal components can be adaptively selected to minimize information loss while reducing the dimensionality.

[0114] The core of PCA is the eigenvalue decomposition of the covariance matrix. First, the original feature matrix is ​​centered, and the mean is translated from the origin to the center of the coordinate axis to eliminate the dimensional influence of the feature. Then, the covariance matrix between features is calculated based on the centered matrix to characterize the correlation between features. Then, the covariance matrix is ​​eigenvalue decomposed to decompose the eigenvector matrix P and the eigenvalue matrix Λ. The eigenvalue represents the variance of each principal component. The larger the variance, the more information the principal component carries. The eigenvector gives the linear transformation matrix from the original feature space to the principal component space, mapping the original feature matrix X to the principal component feature matrix Z.

[0115] After the transformation, the principal component feature matrix Z removes the linear correlation of the original features, but still retains the main information of X. The variance contribution rate of the principal component is obtained by dividing the corresponding eigenvalue by the sum of the eigenvalues, which reflects the importance of the principal component. By setting the variance contribution rate threshold, a few of the most important principal components are selected as machine learning features, and components with low variance contribution rates and less information are eliminated, achieving dimensionality reduction compression while minimizing information loss.

[0116] For example, if the original features include 10 indicators such as unit water supply power consumption X1, net water consumption X2, and pipe network leakage rate X3, these indicators may have a certain linear correlation. The principal component features Z1, Z2, and Z3 obtained by principal component analysis may be only 3, but they concentrate 90% of the information of the original features. Z1 may reflect the overall energy consumption intensity of the water supply system, Z2 reflects the transmission and distribution efficiency of the pipe network, and Z3 reflects the user's water balance. These three principal component features cover the main characteristics of the energy efficiency and water use pattern of the water supply system, but the dimension is greatly reduced, which is conducive to subsequent modeling and prediction.

[0117] Examples of machine learning features include:

[0118] Time period, principal component 1, principal component 2, principal component 3;

[0119] 2023-01-01, -1.25, 0.89, -0.31;

[0120] 2023-01-02, -1.07, 1.12, -0.25;

[0121] Although the machine learning features extracted by principal component analysis are not as interpretable as manually designed features, they can retain the original information of the data to the greatest extent, eliminate feature redundancy, and reveal the inherent laws and hidden structures of the data. The principal component features obtained by PCA can often characterize high-order statistical information that is not covered by artificial features, and deepen the understanding of the target problem. Machine learning features and artificial features complement each other. When used in combination, they can comprehensively and accurately characterize the state of complex water supply systems, laying a solid data foundation for building a water-saving and carbon-reduction prediction model.

[0122] In summary, step S1320 uses principal component analysis to achieve feature dimensionality reduction with minimal information loss, eliminating multicollinearity between original features and retaining the inherent variation of data to the maximum extent. The combination of machine learning feature extraction and artificial feature design overcomes the limitations of a single feature engineering method, characterizes the energy consumption and water use patterns of the water supply system from multiple perspectives such as linear and nonlinear, low-order and high-order, and can better support data mining and knowledge discovery for water conservation and carbon reduction.

[0123] Step S1330, integrating the manually designed features with the machine learning features to construct fourth data.

[0124] The fourth data integrates the key characteristic indicators of water supply plants and large water users in the form of a structured matrix, covering multi-dimensional information such as water volume, electricity consumption, pressure, leakage, peaks and valleys, comprehensively describing the energy consumption intensity and water use patterns of the water supply system, and revealing the intrinsic connection between various parameters. It provides data support for the subsequent carbon emission accounting and identification of water-saving and carbon reduction opportunities. The fourth data is time series data.

[0125] The fourth data is as follows:

[0126] Time period, unit water supply power consumption, net water consumption, pipe network leakage rate, peak water consumption, water balance coefficient, minimum water consumption at night, principal component 1, principal component 2, principal component 3;

[0127] 2023-01-01, 0.521, 0.212, 0.102, 26.5, 0.214, 8.2, -1.25, 0.89, -0.31;

[0128] 2023-01-02, 0.495, 0.203, 0.098, 28.2, 0.198, 8.9, -1.07, 1.12, -0.25;

[0129] Step S1000 builds the data foundation for carbon accounting of water supply system from data perception, data preprocessing to feature engineering. Real-time collection of multi-source data is achieved through online monitoring, and heterogeneous data is integrated into a unified and regular data set by means of data cleaning, normalization and standardization to ensure data quality; on this basis, a compact and comprehensive feature set is extracted from the bottom up by combining manual design and machine learning, so as to prepare for subsequent tasks such as carbon emission accounting and water saving identification. The rigorous data processing process ensures the representativeness and accuracy of the data, while feature engineering removes noise interference while retaining the essence of the data, making the data more refined and conducive to fast and efficient analysis. At the same time, the well-interpreted features give the data a clear physical meaning, making the data analysis results interpretable and feasible. High-quality data and features are the foundation of digital applications and the basis for refined management and intelligent decision-making of water supply systems. Establishing a full-process data quality management system and strengthening data governance capabilities are important links in the digital transformation of the water supply industry.

[0130] Step S2000, calculate the carbon footprint of the water supply plant according to the energy consumption intensity characteristic index of the water supply plant in the fourth data to obtain the first carbon emission data; calculate the carbon emissions of large water users according to the water use pattern characteristic index of large water users in the fourth data to obtain the second carbon emission data; input the first carbon emission data and the second carbon emission data into a preset carbon emission accounting model to obtain a total carbon emission accounting value; obtain a total carbon emission forecast value based on the total carbon emission accounting value, as well as a pre-constructed short-term carbon emission forecast model and a long-term carbon emission forecast model.

[0131] Furthermore, step S2000 includes:

[0132] Step S2100, calculating the carbon footprint of the water supply plant according to the energy consumption intensity characteristic index of the water supply plant in the fourth data to obtain the first carbon emission data;

[0133] Furthermore, if Figure 3 As shown, step S2100 includes:

[0134] Step S2110, calculating the carbon emissions of the electricity use of the water supply plant according to the unit water supply power consumption of the water supply plant in the fourth data and the real-time carbon emission factor of the power grid; the carbon emissions of the electricity use of the water supply plant are equal to the product of the unit water supply power consumption and the water supply and the real-time carbon emission factor of the power grid;

[0135] Specifically, the real-time carbon emission factor of the power grid refers to the indicator of the carbon dioxide equivalent emitted by the power system to produce unit electricity (such as 1 kilowatt-hour) at a specific point in time. The unit is usually grams of CO2 / kWh or kgCO2 / MWh, which reflects the carbon emission intensity of the power grid.

[0136] The calculation principle of the real-time carbon emission factor of the power grid is as follows:

[0137] Determine the composition of the power grid: Count the installed capacity, output and carbon emission levels of various types of generators in the power grid. Generators usually include thermal power (coal power, gas power), hydropower, nuclear power, wind power, photovoltaic power, etc.;

[0138] Calculate the proportion of power generation of each type of power source: According to the real-time output of each type of unit, calculate the proportion of its power generation in the total power generation;

[0139] Calculate the CO2 emissions of various power sources: multiply the carbon emission factor of each unit of power generation by the corresponding power generation to obtain the respective CO2 emissions; fossil energy such as coal-fired power and gas-fired power has a higher carbon emission factor, while renewable energy such as hydropower, nuclear power, wind and solar power has a lower or even zero carbon emission factor;

[0140] Calculate the weighted average: add up the CO2 emissions of various power sources and divide it by the total power generation to get the average carbon emission factor of the power grid, which is the real-time carbon emission factor.

[0141] The higher the proportion of renewable energy generation, the lower the carbon emission factor; coal-fired units are often required to generate electricity at full load during peak electricity consumption periods, which increases the carbon emission factor; different units have different parameters such as energy efficiency levels and emission control measures, which affect carbon emission performance.

[0142] For the electricity use of water supply plants, carbon emissions mainly come from the production of electricity. Due to the dynamic nature of the regional power grid power generation structure, the actual power consumption of water supply plants also changes at any time. When the proportion of renewable energy in the power grid increases, the proportion of thermal power decreases, the carbon emission intensity on the power generation side decreases, and the carbon footprint per unit of electricity consumption of the water supply plant also decreases. Conversely, when the power grid is dominated by thermal power generation, the carbon intensity of electricity production increases, and the carbon emissions of electricity consumption in the water supply plant will also increase. Therefore, it is necessary to introduce a real-time carbon emission factor for the power grid to dynamically reflect the impact of changes in the power structure on the carbon footprint of the water supply system. Through this factor, the comprehensive carbon emission intensity on the power grid side is obtained. The real-time carbon emission factor avoids the limitations of the static electricity carbon emission factor and can objectively evaluate the dynamic carbon emission level of electricity use in the water supply plant. Multiplying the real-time carbon emission factor, the water supply volume and the unit water supply power consumption in the fourth data can obtain the carbon emissions in the electricity use link of the water supply plant.

[0143] Step S2120, calculating the carbon emissions of the water intake link of the water supply plant according to the water consumption of the water supply plant in the fourth data;

[0144] Specifically, for the water intake link of the water supply plant, carbon emissions not only include the direct energy consumption of facilities such as water intake pumps, but also the environmental load carried by water resources themselves. The acquisition and use of water resources will interfere with the ecosystem, leading to problems such as loss of biodiversity, soil degradation, and decline in groundwater levels, which in turn affect the carbon sequestration function of terrestrial ecosystems. Therefore, it is necessary to calculate the comprehensive carbon emission intensity of unit water resource loss. Here, the input-output life cycle assessment method is used to measure the direct and indirect impacts of water resource acquisition, utilization, and emission on the environment from a macro scale. First, based on the input-output table, the complete consumption level and pollutant emission level of each department are calculated to obtain the carbon intensity of water resource loss at the department level. Secondly, considering the differences in the scarcity of different water resources, corresponding weights are assigned to surface water, groundwater, and external water, and the comprehensive water resource carbon emission intensity at the regional scale is calculated. Finally, the carbon intensity of water resources, water supply, and net water consumption in the fourth data are multiplied to obtain the carbon emissions in the water intake link. This process not only quantifies the direct carbon emissions from water intake, but also internalizes the impact of water scarcity on the ecological environment, making the carbon accounting results more comprehensive and objective. The greater the water consumption of the water supply plant, the greater the water intake, and the greater the carbon emissions caused by water resource consumption.

[0145] The calculation of carbon emissions from water intake at a water supply plant includes:

[0146] ;

[0147] in:

[0148] : Carbon emissions from water extraction.

[0149] : Water supply of water supply plant (Unit: ), which indicates the total amount of clean water delivered by the water supply plant. This parameter can be obtained through flow meter or user water meter data.

[0150] :The fourth data, water consumption, indicates the volume of raw water consumed to produce a unit volume of clean water. This parameter is calculated from the water inlet and outlet data of the water supply plant.

[0151] : No. The weight coefficients of various water sources (such as surface water, groundwater, external water, etc.) are calculated, and the sum of the weights of each water source is 1. These parameters are determined based on the proportion of water supply of each type of water source, reflecting the differences in scarcity of different water sources.

[0152] : No. These parameters are estimated through the input-output life cycle assessment method, reflecting the ecological and environmental impacts of the development and utilization of different water sources.

[0153] : Water Supply Plant No. These parameters are calculated based on the water intake and supply of various water sources and reflect the efficiency of water resource utilization.

[0154] : Power consumption of the water intake pump. This parameter is determined by the rated power and actual load rate of the water intake pump.

[0155] : Carbon emission factor per unit of electricity. This parameter is estimated based on the power generation structure and fuel composition of the regional power grid.

[0156] : The overall efficiency of the water intake pump. This parameter is determined by the equipment parameters and operating conditions of the water intake pump.

[0157] : Ecological water use ratio. This parameter is calculated based on the regional ecological water use and total water supply.

[0158] :Carbon sink loss coefficient per unit ecological water use. This parameter is estimated based on the relationship between ecological water use and ecosystem carbon sequestration function.

[0159] : Ecological water use scheduling coefficient. This parameter reflects the impact of the ecological water use scheduling scheme on the carbon sink function.

[0160] The physical meaning of this formula is that the carbon emissions in the water supply plant's water intake link are equal to the ratio of direct water intake carbon emissions (including water resource carbon emissions and water intake pump energy consumption carbon emissions) to ecological water use carbon sink losses. Among them, the numerator of direct water intake carbon emissions represents the comprehensive impact of water supply, net water consumption, water resource endowment and water intake pump energy efficiency on carbon emissions; the denominator of ecological water use carbon sink losses describes the reduction of water ecological carbon sink function caused by ecological water use scheduling, which offsets part of the carbon emissions to a certain extent.

[0161] This formula uses a fractional form to compare direct carbon emissions with ecological carbon sink losses, making carbon emission accounting more comprehensive and balanced; the numerator comprehensively considers multi-dimensional factors such as water volume, water consumption, water sources, and energy consumption, and describes the "water-energy-carbon" coupling relationship of the water supply system; the introduction of water pump energy efficiency parameters can sensitively reflect the emission reduction effects of energy-saving measures; the denominator incorporates ecological water use scheduling factors, which can evaluate the carbon sink benefits of ecological water use security and realize "water-ecology-carbon" coupling.

[0162] Measures to reduce carbon emissions from water withdrawal:

[0163] Demand management: Strengthen demand management, improve water use efficiency, and control total water supply .

[0164] Optimize water purification process: improve water purification rate and reduce water consumption .

[0165] Optimal allocation of water sources: Coordinated allocation of multiple water sources such as surface water, groundwater, and unconventional water, and optimize water source weights .

[0166] Water intake pump management: Strengthen water intake pump operation management and reduce power consumption , improve motor and inverter efficiency .

[0167] Power structure optimization: optimize the power structure, increase the proportion of renewable energy, and reduce the carbon emission factor of electricity .

[0168] Ecological water management: Strengthen ecological water management and improve ecological water dispatch coefficient , maximizing the ecological carbon sink benefits.

[0169] Water resource control: strictly control water resources and improve water resource utilization , control the carbon emission intensity per unit of water resources .

[0170] In summary, the formula is based on the entire process of water cycle, starting from the two dimensions of water supply and ecology, comprehensively considering multiple influencing factors, quantitatively describing the internal mechanism of "water-energy-ecology-carbon" in the water supply system, and forming a closed carbon accounting framework covering water carbon sources and sinks and ecological feedback. By parameterizing various influencing factors, the formula can flexibly evaluate the carbon reduction potential of different water-saving, energy-saving, water transfer, and water protection measures, screen out the optimal water-saving and emission reduction strategy, and conduct scenario analysis and policy evaluation based on this. This has important theoretical and practical value for promoting the low-carbon transformation of the water supply system and supporting the refined management of water resources and ecological environment in the basin.

[0171] Step S2130, calculating the carbon emissions of the water supply plant's transmission and distribution link according to the pipe network leakage rate of the water supply plant in the fourth data;

[0172] The calculation of the carbon emissions of the water supply plant's transmission and distribution link according to the pipe network leakage rate of the water supply plant in the fourth data includes:

[0173] The process life cycle assessment method is used to calculate the total carbon emissions of the pipeline network throughout its life cycle. The total carbon emissions of the pipeline network throughout its life cycle are divided by the water supply to obtain the carbon intensity per unit water supply, which is then multiplied by the pipeline leakage rate to obtain the carbon emissions of the water supply plant's transmission and distribution links.

[0174] Specifically, for the transmission and distribution of water supply plants, carbon emissions not only occur in the treatment of water loss in the pipeline network, but are also implicit in the entire life cycle of the pipeline network assets. Traditional methods focus more on the impact of pipeline network leakage rate on water supply efficiency, but ignore the life cycle carbon footprint of the water supply network itself. During the pipeline network construction phase, the production, transportation, and laying of pipes consume resources and emit pollution; during the operation phase of the pipeline network, maintenance operations such as emergency repairs and renovations, as well as the treatment of leaked water, will also generate carbon emissions; during the scrapping phase of the pipeline network, landfilling and incineration of waste pipes will still have an impact on the environment. Only by centrally accounting for these carbon emissions scattered in various stages of the pipeline network life cycle can the carbon footprint of the water supply system transmission and distribution link be accurately evaluated. Here, the process life cycle assessment method is adopted, based on the ISO 14040 standard, to calculate the total carbon emissions of the pipeline network throughout its life cycle, and to analyze the carbon emission inventory of each process of the water supply network from material acquisition to final disposal. Process LCA requires a lot of on-site investigation and data support. Although the calculation results are accurate, the data quality requirements are very high and the workload is large. The total carbon emissions of the entire life cycle of the pipeline network are divided by the water supply to get the unit water supply carbon intensity, and then multiplied by the pipeline network leakage rate to get the carbon emissions of the transmission and distribution link. It can be seen that the higher the pipeline network leakage rate, the higher the frequency of leakage water treatment and pipeline network maintenance, and the greater the carbon emission intensity of the transmission and distribution link.

[0175] Step S2140, summing up the carbon emissions of the electricity use, water intake and transmission and distribution links of the water supply plant to obtain the carbon footprint of the water supply plant and form the first carbon emission data.

[0176] Step S2100 uses the life cycle assessment method to systematically calculate the carbon emissions of each link of the water supply plant. Traditional carbon accounting often only focuses on fossil energy consumption in the operation stage, ignoring carbon emissions in other links of the water supply system throughout its life cycle. The life cycle assessment takes into account the resource consumption and pollution emissions of the water supply plant construction, operation, maintenance, and scrapping from a full perspective, revealing the hidden carbon footprint. The carbon footprint of the water supply plant can be obtained by summing up the carbon emissions of the three parts of electricity use, water intake, and transmission and distribution. The data quality and calculation method of carbon emission measurement in each link directly affect the accuracy of the carbon footprint. Through systematic accounting from a life cycle perspective, the water supply system is regarded as an organic whole, and the direct and indirect carbon emissions caused by the water-energy relationship are fully included. The impact of resource and environmental efficiency on the carbon footprint is quantitatively evaluated, thereby revealing the implicit carbon emission mechanism of the water supply system. Compared with the carbon accounting of a single link, the carbon footprint of the water supply plant can more comprehensively and objectively reflect the resource and environmental costs of the operation of the water supply system, and provide a decision-making basis for the formulation of water-saving and emission reduction policies.

[0177] The first carbon emission data is time series data, exemplarily:

[0178] Time period, carbon emissions in the use stage (tCO2), carbon emissions in the water intake stage (tCO2), and carbon emissions in the transmission and distribution stage (tCO2);

[0179] 2023-01-01, 1.25, 2.62, 3.54;

[0180] 2023-01-02, 1.32, 2.58, 3.67;

[0181] Step S2200, calculating the carbon emissions of large water users according to the water use pattern characteristic indicators of large water users in the fourth data to obtain second carbon emission data;

[0182] Furthermore, if Figure 4 As shown, step S2200 includes:

[0183] Step S2210, calculating the carbon emissions of transmission and distribution during the peak period according to the peak water consumption of large water users in the fourth data;

[0184] Furthermore, if Figure 5 As shown, step S2210 includes:

[0185] Step S2211, determining the peak load regulation operation mode of the water supply network; the peak load regulation operation mode of the water supply network includes two types: variable frequency speed regulation and pump start and stop;

[0186] Step S2212, according to the peak load regulation operation mode of the water supply network and the peak water consumption of large water users, the network energy consumption during the peak period is calculated; under the variable frequency speed regulation operation mode, the network energy consumption during the peak period is proportional to the square of the peak water consumption of large water users; under the start-stop pump operation mode, the network energy consumption during the peak period is proportional to the peak water consumption of large water users;

[0187] Step S2213, based on the real-time carbon emission factor of the power grid during the peak period, convert the energy consumption of the pipeline network during the peak period into the transmission and distribution carbon emissions during the peak period.

[0188] Specifically, step S2210 accurately calculates the carbon emissions of distribution during peak hours according to the peak-shaving operation mode of the water supply network. The water supply network faces greater water supply pressure during peak hours, and needs to meet user needs through peak-shaving means. Common peak-shaving operation modes include variable frequency speed regulation and start-stop pumps. Variable frequency speed regulation refers to controlling the speed of the water pump through a frequency converter, dynamically adjusting the water supply of the water pump according to the feedback of the network pressure and flow, and achieving real-time matching with water demand. This method can smoothly adjust the flow, reduce water supply fluctuations, and reduce network pressure shocks, and has the advantages of energy saving, stability, and reduced leakage. Start-stop pumps are traditional peak-shaving methods, which adjust the water supply by a combination of start-stop pumps. When water consumption increases, more water pumps are turned on; when water consumption decreases, some water pumps are turned off. Although the start-stop pump is simple to control and the equipment investment is small, frequent start-stopping will reduce the service life of the water pump, large water supply fluctuations, high energy consumption, and unstable network pressure. In variable frequency speed regulation mode, the energy consumption of the pipeline network is proportional to the square of the peak water consumption, and in pump start-stop mode, the energy consumption of the pipeline network is proportional to the peak water consumption.

[0189] Finally, the real-time carbon emission factor of the power grid is used to convert the electricity consumption of the pipeline network into carbon emission equivalents. The real-time carbon emission factor during peak hours needs to be matched during calculation to obtain more accurate carbon emission accounting results. The calculation of peak transmission and distribution carbon emissions reflects the peak energy consumption and carbon emission characteristics of the water supply network. The greater the peak water consumption, the higher the energy consumption of the pipeline network during peak hours, and the greater the carbon emissions of transmission and distribution. Reasonable peak-shaving operation mode can minimize the energy consumption of the pipeline network while ensuring the water use of residents, and achieve energy conservation and carbon reduction in a coordinated manner. This provides an important basis for the low-carbon optimization operation of the water supply network, guides the pipeline network to adopt advanced variable frequency speed regulation technology, pre-peak shaving before the peak water use arrives, and reduce energy consumption waste during the ramp process. At the same time, based on the carbon emission factor matching the real-time emission characteristics of the power grid, priority is given to peak shaving and water storage when low-carbon electricity is abundant, and water supply is reduced during high-carbon periods. Active participation in power demand side response helps the low-carbon transformation of the energy system. Step S2210 quantifies the impact of large water users’ water use behavior on the carbon emissions of the pipeline network, reflects the interactive relationship between supply and demand, deepens the understanding of the carbon emission mechanism of the water supply system, and lays the foundation for formulating precise system water-saving measures.

[0190] Step S2220, calculating the balanced carbon emissions of transmission and distribution energy consumption according to the water balance coefficient of the large water users in the fourth data;

[0191] Specifically, for balanced carbon emissions, it mainly depends on the fluctuation characteristics of the load curve of large water users. The dynamic changes in water demand will cause frequent adjustments to the pressure and flow of the pipe network, affecting the energy consumption and efficiency of the system. If the load curve fluctuates violently, the peak regulation of the water supply network will be more difficult, energy consumption will increase, and carbon emissions will rise. On the contrary, if the water demand is relatively stable, the operation of the pipe network will also tend to be balanced, and the carbon intensity will decrease. Balanced carbon emissions are calculated using the energy integration method. The product of the standard deviation of the water consumption time series and the unit energy consumption of the water supply network is integrated over time to obtain balanced carbon emissions. The energy integration method is used to calculate balanced carbon emissions. The volatility of the load curve is evaluated by integrating the energy consumption of the water supply network over time. Energy consumption is closely related to flow fluctuations. The standard deviation of the dynamically changing flow (i.e., the water consumption time series) is calculated, multiplied by the unit water network energy consumption, and integrated over time. The result represents the response of the energy consumption of the water supply network to demand fluctuations. The larger the integral value, the more violent the fluctuation of water supply energy consumption with demand, and the more balanced carbon emissions. Introducing the water balance coefficient feature in the fourth data into the energy integration formula, we can get a quantitative balanced carbon emission value. It can be seen that the smaller the water balance coefficient, the more stable the user's water demand, the more balanced the operation of the water supply network, and the lower the carbon emission intensity.

[0192] Step S2230, calculating the background leakage carbon emissions of the pipe network according to the minimum nighttime water consumption of the large water users in the fourth data;

[0193] Specifically, the background leakage carbon emissions are mainly related to the nighttime water use characteristics of large water users. Since the pressure of the pipeline network where large water users are located and the water demand at night are usually very low, the water consumption during the night period is mainly caused by the leakage of the pipeline network itself. The larger the minimum water consumption at night, the higher the background leakage level of the pipeline network where the large water users are located. The leakage water will generate additional carbon emissions in the transmission and distribution link, and its carbon footprint cannot be ignored. Background leakage carbon emissions can be estimated by multiplying the carbon intensity of pipeline energy consumption by the minimum water consumption at night. The carbon intensity of pipeline energy consumption is calculated based on the unit water distribution power consumption of the water supply pipeline network and the real-time carbon emission factor of the power grid. The minimum water consumption at night reflects the volume of leakage in the pipeline network where large water users are located, and directly determines the scale of carbon emissions from background leakage. Although the nighttime water use behavior of large water users has no direct impact on the occurrence of pipeline network leakage, its data characteristics can characterize the leakage status of the pipeline network where they are located. Incorporating the minimum water consumption at night in the fourth data into carbon accounting can reveal the intrinsic connection between the water demand characteristics of large water users and the efficiency of their pipeline networks.

[0194] The background leakage carbon emissions of the pipe network are calculated as follows:

[0195] ;

[0196] in:

[0197] : Background leakage carbon emissions from pipeline networks.

[0198] :The carbon intensity of the unit water distribution energy consumption of the pipeline network, that is, the carbon emissions of distributing 1 cubic meter of water. It is calculated based on the unit water distribution power consumption of the pipeline network and the real-time carbon emission factor of the power grid.

[0199] : Minimum water consumption at night by large water users.

[0200] : Leakage correction factor, depends on the pipe material characteristics and leakage type, obtained through experiments or empirical estimation.

[0201] : Leakage index, depends on crack propagation characteristics, soil properties, etc., and is obtained through experiments or empirical estimates.

[0202] :The daily average pressure of the pipeline network is collected in real time by installing pressure sensors at key nodes of the pipeline network.

[0203] : The minimum pressure of the pipeline network at night, which corresponds to the pipeline network pressure when the water consumption is minimum at night, is also collected through the pressure sensor.

[0204] and Proportional: When When the carbon emission of background leakage increases It also increases accordingly, indicating that the higher the carbon intensity of water distribution energy consumption in the pipeline network, the greater the carbon emissions per unit of water distribution.

[0205] and Proportional: When When it increases, the background leakage volume increases, resulting in an increase in carbon emissions corresponding to the leakage water.

[0206] and Exponentially positively correlated: When it increases, the leakage correction coefficient increases and the background leakage carbon emission increases significantly.

[0207] and Exponentially negatively correlated: When it increases, the leakage correction factor decreases and the increase in background leakage carbon emissions is mitigated.

[0208] when and When the background leakage increases, the carbon emission Pressure fluctuations The stronger the response, the more sensitive the pipeline network is to pressure.

[0209] The introduction of leakage correction items improves the level of refinement of carbon emissions from background leakage. Under the ideal condition of constant pipe network pressure, the background leakage is only proportional to the minimum water consumption at night. However, the pressure of the actual water supply network often fluctuates greatly, especially during peak hours when large water users use water in large quantities. The sudden increase in pressure will significantly aggravate the leakage of the pipe network. The leakage correction item takes into account the impact of pressure fluctuations and dynamically describes the nonlinear response of background leakage to pressure, making the carbon emission accounting more in line with the actual operating status of the pipe network. At the same time, the leakage coefficient λ and the leakage index α reflect the differences in leakage characteristics of different pipe networks. The larger the λ value, the more fragile the pipe network is, and the greater the leakage at the same pressure; the larger the α value, the more sensitive the pipe network is to pressure fluctuations, and the faster the leakage increases with the increase in pressure. For weak pipe networks with high λ and α values, we should strengthen transformation and maintenance, balance the pipe network pressure, and curb background leakage at the source.

[0210] In general, this formula starts from the operation mechanism of the pipeline network. On the basis of characterizing the background leakage volume by the minimum water consumption at night, it further considers the amplification effect of pressure fluctuations and pipeline network characteristics on leakage, forming a carbon accounting model that is both explanatory and applicable. Based on the calculation results of this model, the background leakage carbon emission level of the pipeline network where large water users are located can be quantitatively evaluated, key governance targets can be identified, and the effect of water-saving transformation can be monitored. At the same time, the explicit introduction of pressure and leakage parameters also provides an entry point and quantitative basis for the implementation of refined management measures such as pipeline network zoning metering, pressure regulation, and leakage control, which will help the water supply network to save energy, increase efficiency, improve quality and reduce carbon emissions.

[0211] Step S2240, adding the carbon emissions of transmission and distribution during peak hours, the balanced carbon emissions of transmission and distribution energy consumption, and the background leakage carbon emissions of the pipeline network to obtain the carbon emission footprint of large water users and form the second carbon emission data.

[0212] Step S2200 analyzes the impact of water use behavior of large water users on the carbon footprint of the water supply system from the perspective of water demand. The water use patterns of different users vary greatly, which is reflected in water use intensity, water use time period and fluctuation pattern. The water use behavior on the demand side is transmitted to the supply side through the dynamic balance mechanism of supply and demand, thereby affecting the carbon emission level of the water supply system. As the main body of water demand, the peak and valley water use characteristics and load curve shape of large water users directly determine the distribution mode and energy consumption intensity of the water supply network. Therefore, it is necessary to calculate the carbon emissions of the water use patterns of large water users and explore the changes in the carbon footprint of the water supply system caused by demand-side behavior.

[0213] The second carbon emission data is time series data, exemplarily:

[0214] Time period, peak period carbon emissions (tCO2), balanced carbon emissions of transmission and distribution energy consumption (tCO2), and background leakage carbon emissions of the pipeline network (tCO2);

[0215] 2023-01-01, 1.85, 0.92, 0.54;

[0216] 2023-01-02, 1.98, 0.88, 0.61;

[0217] The carbon emissions of water supply networks are affected by the peak and valley characteristics of the demand side, load curves, nighttime flow and other factors, and cannot be simply attributed to the design and operation of the network itself. Analyzing the carbon footprint of the network based on the characteristics of water demand is helpful to reveal the carbon emission mechanism of the water supply system from the perspective of supply and demand. The total carbon emissions of large water users are obtained by adding up the carbon emissions of the three dimensions of peak period transmission and distribution, balance, and background leakage. Compared with a single indicator, multi-dimensional synthesis can fully reflect the carbon effect of water use behavior and provide a more detailed and reliable carbon accounting basis for demand-side management and water-saving policy formulation.

[0218] Step S2300, input the first carbon emission data and the second carbon emission data into a preset carbon emission accounting model to obtain a total carbon emission accounting value; obtain a total carbon emission prediction value based on the total carbon emission accounting value, as well as a pre-built short-term carbon emission prediction model and a long-term carbon emission prediction model.

[0219] Furthermore, step S2300 includes:

[0220] Step S2310, inputting the first carbon emission data and the second carbon emission data into a preset carbon emission accounting model to obtain a total carbon emission accounting value;

[0221] Specifically, step S2310 superimposes the carbon footprint of the water supply plant and the carbon footprint of large water users to obtain the total carbon emissions of the water supply system from water intake, water production to distribution and water use. The carbon emission accounting model is based on the carbon emission inventory method, combined with the water supply process route and water use scenario, to construct a carbon accounting methodology suitable for urban water supply systems. This methodology draws on internationally accepted standards such as ISO 14064 and IPCC, takes into account the characteristics of China's water supply industry, and adopts a bottom-up and step-by-step accounting approach. The water supply system is divided into subsystems such as raw water, purified water, distribution, and terminal water use, and the direct and indirect carbon emissions of each subsystem are calculated separately, and then the total carbon emission accounting value is obtained by adding them up. Among them, direct emissions mainly come from the use of water purification agents and disinfectants, and indirect emissions mainly come from energy consumption such as electricity and heat. Finally, the carbon accounting model uses the total carbon emission accounting value as a comprehensive evaluation indicator to form a highly comparable and scientific accounting result, which provides an important basis for carrying out carbon management of water supply systems and formulating water-saving and carbon reduction policies. The total carbon emissions calculation value is time series data, for example:

[0222] Time period, Carbon emissions from water supply plants (tCO2), Carbon emissions from large water users (tCO2), Total carbon emissions (tCO2);

[0223] 2023-01-01, 7.41, 3.31, 10.72;

[0224] 2023-01-02, 7.57, 3.47, 11.04;

[0225] The carbon emission accounting model overcomes the technical difficulties of cross-domain and cross-border accounting by constructing a full-process carbon accounting framework. The water supply system involves two major sectors, industrial and civil, and two stages of production and consumption. The types of carbon emission sources are diverse and complex, making it difficult to uniformly account. The model innovatively connects the production and use links closely, opening up the carbon accounting channel for water supply and water use. The integrated accounting perspective of the plant and network not only avoids repeated calculations of carbon emissions, but also tracks carbon losses caused by pipeline leakage, making the carbon accounting results more comprehensive and accurate. At the same time, the model also makes full use of multi-source data to improve the spatiotemporal granularity and dynamics of carbon emission quantification, which can support refined carbon management. As a weather vane for macro-control, the total carbon emission accounting value can guide the formation of a green and low-carbon oriented water supply scheduling and demand-side management mechanism. Therefore, the carbon emission accounting model is a powerful tool for the water supply industry to respond to climate change and promote carbon peak and carbon neutrality. It is of great significance to explore water-carbon coupling, collaborative governance paths, and accelerate the low-carbon transformation of water supply.

[0226] For example, a city's water supply company used the carbon accounting model to calculate the total annual carbon emissions of the water supply system to be 30,000 tons. Among them, the carbon footprint of the water supply plant accounted for 60%, mainly from the power consumption of water purification agents and water pumps; the carbon footprint of large water users accounted for 40%, mainly from the water use of air conditioning cooling towers and boilers. By comparing the sub-item carbon emission data, the company found that the unit consumption of the water purification process was high, which was the focus of carbon reduction; the air conditioning system was a major carbon emitter among water users, and recycling condensed water could significantly reduce carbon. Based on this, the water company formulated a package of carbon reduction measures such as upgrading and transforming water purification processes and promoting the recycling of reclaimed water, striving to reduce the total carbon emissions by 20% in three years. It can be seen that the carbon accounting model provides a carbon emission portrait from a global perspective, which helps to identify key carbon sources, focus on major contradictions, formulate water-saving and carbon reduction paths according to local conditions, and promote high-quality and low-carbon development of the water supply industry.

[0227] Step S2320, obtaining weather data and holiday data, building a short-term carbon emission prediction model, and predicting daily carbon emissions in the future short-term period based on the total carbon emission accounting value, weather data, holiday data and the short-term carbon emission prediction model;

[0228] Specifically, short-term carbon emission forecasting aims to capture the dynamic changes in carbon emissions and achieve early warning and real-time optimization and regulation of carbon emissions. Carbon emissions from water supply systems show obvious short-term fluctuations, which are affected by factors such as weather conditions and holiday effects. A large amount of empirical evidence shows that hot and dry weather will significantly increase water supply and carbon emissions, while rainy and low-temperature weather will suppress water demand and carbon emissions. During holidays, water use peaks frequently occur, and carbon emissions often appear abnormally high. Therefore, incorporating weather data and holiday data into the carbon emission forecasting model can effectively improve the accuracy of short-term forecasts.

[0229] The total carbon emission accounting value, weather data, and holiday data are all time series data; short-term carbon emission forecasts use a strategy that combines time series analysis and machine learning. Time series analysis uses traditional methods such as moving average and exponential smoothing to characterize the trend, periodicity, and randomness of carbon emissions, but it is difficult to integrate exogenous variables. Machine learning models such as support vector machines and random forests can establish a nonlinear relationship between carbon emissions and weather and holidays, capture the dynamic response of carbon emissions to external shocks, and have better forecasting effects. The two methods complement each other. The short-term carbon emission forecasting model combines the grasp of the endogenous laws of carbon emissions by time series analysis and the learning ability of machine learning models for exogenous factors to form a fusion algorithm that overcomes the limitations of a single model.

[0230] In terms of feature selection, the model uses daily carbon emissions as the dependent variable, meteorological factors such as temperature, humidity, rainfall, and holiday dummy variables as independent variables, and historical carbon emissions are included to characterize the inertia of carbon emissions. When training the model, time series analysis is first used to strip off the long-term trend of carbon emissions, and then the machine learning model is used to fit the short-term fluctuation of carbon emissions. Finally, the two items are combined to form daily carbon emissions. The model verification uses methods such as cross-validation to continuously optimize hyperparameters and improve the generalization ability of prediction. The prediction results are output on a daily scale to show the short-term nonlinear fluctuation trajectory of carbon emissions.

[0231] Short-term carbon emission forecasts can provide a decision-making basis for precise peak load regulation. According to carbon emission warnings, water supply companies can proactively adjust their operation plans, increase water supply in advance before the peak water use arrives, guide users to use water in staggered periods, and alleviate water supply and carbon emission pressures. At the same time, short-term forecasts can also optimize carbon quota management. Water supply companies can dynamically adjust daily carbon emission quotas based on the predicted carbon emission values ​​for the next week, minimize carbon emissions and smooth the carbon emission curve while ensuring water supply safety. In addition, carbon emission forecasts for key users can guide them to adjust water use periods and processes, shave peaks and fill valleys, and reduce ineffective carbon emissions. In short, short-term carbon emission forecasts can support the formation of a water demand response mechanism under carbon constraints by proactively perceiving future carbon emission levels, coordinate efforts at the source and the terminal, improve the efficiency of carbon resource allocation in the water supply system, and help achieve the carbon peak target as soon as possible.

[0232] For example, under continuous high temperature weather, the daily carbon emissions of the water supply system in City A climbed to 1,500 tons, approaching the carbon emission warning line. Short-term forecasts show that carbon emissions will increase at a rate of 5% in the next three days, with a risk of exceeding the standard. Based on this warning, the water supply dispatching department raised the water supply pressure in advance and encouraged large users to use water at night; water-saving service personnel conducted household inspections and promptly repaired leaks; industrial parks adopted rotating production suspension and production restrictions to stagger production. Through the coordination of water supply and use, the growth rate of carbon emissions in the next three days will be reduced to less than 2%, effectively alleviating the carbon reduction pressure on the water supply system. It can be seen that short-term carbon emission forecasts can accurately warn of carbon emission anomalies, drive the linkage of carbon reduction in the entire chain, and win more regulatory space for low-carbon water supply operations.

[0233] Step S2330, obtaining long-term prediction related data, building a long-term carbon emission prediction model, and predicting the monthly carbon emissions in the future long-term period based on the total carbon emission accounting value, the long-term prediction related data and the long-term carbon emission prediction model; the long-term prediction related data includes seasonal regularity data, statutory holiday data and facility maintenance plan data;

[0234] Specifically, long-term carbon emission forecasts mainly deal with seasonal changes in carbon emissions and describe the periodic fluctuations of carbon emissions in the water supply system within the year. The seasonality of carbon emissions is mainly due to seasonal changes in the climate. In different seasons, there are significant differences in temperature, precipitation, and water demand, which in turn affect the energy consumption and carbon emission levels of water supply. Usually, carbon emissions in summer are significantly higher than in winter, and carbon emissions in the dry season are higher than in the wet season. The periodic changes in carbon emissions caused by seasonal factors, that is, seasonal regularity data, can be described by seasonal indexes. In addition, major holidays such as the Spring Festival and National Day often cause large fluctuations in water consumption, which is an important factor affecting carbon emissions. Therefore, by analyzing the daily water supply data before and after statutory holidays in previous years, the water use patterns of different types of holidays can be found, and the water supply and carbon emissions in the same period of the future year can be predicted. These data are usually based on the time granularity of days. Planned events such as facility maintenance and process transformation will also affect the carbon emission level by changing the water supply operation mode. Taking into account seasonal laws, holiday effects, facility maintenance and other factors, carbon emissions in the next quarter to a year can be more accurately predicted, providing a reference for carbon budget preparation and carbon management decisions.

[0235] The long-term forecast-related data is time series data. The long-term carbon emission forecast is based on a monthly scale and adopts a structured time series model. The structured time series model decomposes the carbon emission sequence into long-term trend terms, seasonal fluctuation terms, and irregular random terms, and then models and predicts each sub-item. Among them, the long-term trend term describes the development trend of carbon emissions and can be fitted using the differential autoregressive moving average model ARIMA; the input is the time series of carbon emission historical data, and the output is the point forecast value of carbon emissions in the future.

[0236] Seasonal fluctuations describe the cyclical law of carbon emissions, and the seasonal index method is a commonly used modeling tool; it assumes that seasonal fluctuations are stable over several years, and uses the average value of the same period of several years as the seasonal index to indicate the relative position of the period. Irregular random terms describe the response of carbon emissions to incidental events such as holidays and facility maintenance, and can be introduced into the model as dummy variables. By predicting the three components separately and then combining them, the monthly carbon emission forecast can be obtained. The structural time series model combines long-term trend extrapolation with seasonal fluctuation interpolation, making full use of the multi-scale information of the carbon emission sequence to form a more robust forecast of carbon emissions in the future period.

[0237] In the long-term potential modeling, ARIMA is widely used in non-stationary time series forecasting. It first differentiates the sequence, then builds autoregression and moving average models on the difference sequence, and finally restores the difference to obtain the predicted value. ARIMA eliminates the non-stationarity of the carbon emission series through differentiation and autoregression, making the residual term more consistent with the white noise assumption.

[0238] Long-term carbon emission forecasts are an important basis for the top-level design of carbon management. Based on long-term forecasts, water supply companies can estimate the total carbon emission accounting value for the next quarter to a year, coordinate carbon emission quotas in advance, build a carbon emission budget management framework, and achieve forward-looking and refined management of carbon assets. At the same time, water supply authorities can refer to the forecast values ​​to reasonably allocate regional carbon emission indicators, formulate differentiated carbon emission reduction plans, and guide regional water supply to coordinate carbon reduction. In addition, long-term forecasts can also be used to demonstrate the feasibility and benefits of water-saving and carbon-reduction projects. Simulate different water-saving and carbon-reduction scenarios, estimate future carbon emission reduction potential, screen preferred plans with high carbon cost-effectiveness, and guide enterprises to carry out technical transformation and equipment renewal. In short, the long-term forecast of carbon emissions in the water supply system, based on a macro perspective and focusing on long-term development, can guide the formation of a strategically oriented carbon management pattern and promote the precise implementation and scientific implementation of the carbon peak and carbon neutrality goals of the water supply industry on the timetable and roadmap.

[0239] For example, the carbon emissions of the water supply system in City K have obvious seasonality. The average monthly carbon emissions in summer are 5,000 tons, and only 3,000 tons in winter. The structural time series model predicts that due to the maintenance of facilities, carbon emissions in the first quarter of the next year will decrease by 10% year-on-year. Based on this, the Water Supply Group lowered the quarterly carbon emission quotas of each subsidiary. At the same time, each subsidiary also took the opportunity to implement water-saving and carbon-reduction transformation, introduce high-efficiency motors, and optimize water purification processes. It is estimated that after the transformation, 500 tons of carbon can be reduced each quarter, which basically offsets the loss of facility maintenance. It can be seen that long-term carbon emission forecasts can maximize the preservation and appreciation of carbon assets by prospectively judging future carbon emission levels, optimizing the spatial and temporal configuration of carbon emissions, and guiding enterprises to flexibly adjust internal carbon resource configuration. Establishing a carbon budget management system and carrying out carbon asset value management are important means for the water supply industry to implement carbon peak and carbon neutrality goals and accelerate green and low-carbon development.

[0240] Step S2340, the daily carbon emissions and the monthly carbon emissions are superimposed and combined to form a total carbon emissions forecast value.

[0241] Specifically, the shorter the time scale of carbon emission prediction, the stronger the ability to depict the dynamic changes of carbon emissions, but it is greatly affected by random fluctuations and has high uncertainty; the longer the prediction scale, the stronger the ability to reflect the long-term evolution trend of carbon emissions, but it is slow to respond to peaks and troughs and has low sensitivity. Short-term and long-term carbon emission predictions look at the carbon emission operation status of the water supply system from the micro and macro perspectives respectively, each with its own emphasis and complementing each other. The combination of daily and monthly carbon emission forecast values ​​can take into account the timeliness and stability of carbon emission forecasts, forming a more comprehensive and reliable forecast result.

[0242] In specific implementation, carbon emission forecasting can be carried out in two steps. First, using the short-term carbon emission forecasting model, based on weather forecasts and holiday arrangements, a rolling forecast of daily carbon emissions for the next 15 days or so is made. Since the accuracy of weather forecasts decays rapidly over time, short-term forecasts should be controlled within 2 weeks. Secondly, a long-term forecast of carbon emissions in subsequent months is carried out. Taking into account the relatively stable seasonal patterns, the monthly forecast accuracy for the next 3-6 months is acceptable. Then, starting from the half month covered by the short-term forecast, the monthly carbon emissions obtained from the long-term forecast are evenly distributed to each day, seamlessly connected with the short-term forecast, and finally forming a complete sequence of daily carbon emissions for the next 3-6 months. This sequence covers short-term fluctuations and long-term trends, and can be used to guide full-time domain decision-making for carbon emission management in water supply systems.

[0243] Example of carbon emission forecast:

[0244] Time period, daily carbon emission forecast value (tCO2), monthly carbon emission forecast value (tCO2);

[0245] 2023-01-01, 10.85, 332.6;

[0246] 2023-01-02, 11.12, 332.6;

[0247] 2023-01-03, 10.93, 332.6;

[0248] The organic integration of short-term and long-term carbon emission forecast results lays the foundation for the joint management of carbon emissions in the "two dimensions" of the water supply system. From the perspective of time, short-term forecasts focus on carbon emissions, "one picture to guide refined carbon control; long-term forecasts to grasp carbon emissions" as a chess game, and promote systematic carbon governance. From the perspective of space, based on daily scale forecasts, vertical benchmarking such as year-on-year and month-on-month can be carried out to find out the shortcomings of carbon management; with monthly scale forecasts as a reference, regional coordination can be coordinated, horizontal comparisons and carbon trading can be carried out, and the spatial allocation of carbon resources can be optimized. Vertical benchmarking and horizontal comparison form a closed loop, which can accurately diagnose the weak links in carbon emission management, continuously optimize the carbon emission reduction path, and stimulate the endogenous motivation of market players to reduce carbon. In short, short-term forecasts enable refined carbon control, and long-term forecasts lead systematic carbon governance. The two types of forecasts support each other and combine up and down to form a three-dimensional and comprehensive carbon emission forecast and management pattern. As carbon forecasts evolve towards real-time and intelligent directions, the efficiency of carbon resource allocation in the water supply industry will continue to improve, and the carbon peak and carbon neutrality goals can be expected to be achieved as scheduled.

[0249] For example, the S City Water Supply Group has connected short-term and long-term carbon emission prediction modules to the smart water supply platform. Through the carbon emission prediction screen, the operation and maintenance personnel can intuitively grasp the daily carbon emission trend chart for the next 10 days. In view of the predicted sudden increase in carbon emissions in the next three days, the operation and maintenance personnel promptly increased the frequency of pump room inspections, optimized the start and stop of the pump group, and reduced the peak of carbon emissions. At the same time, the sales staff promoted the water reuse plan to high-carbon emission industrial users, which was favored by users and is expected to reduce annual carbon emissions by 5%. In addition, facing the expected increase in carbon emissions next month, S City took the initiative to purchase surplus carbon emission quotas from the neighboring D City, which revitalized the regional carbon asset stock. It can be seen that the carbon emission predictions at the daily and monthly scales are seamlessly connected, forming a three-dimensional carbon emission prediction picture covering all time and space. This new carbon management tool gives water supply companies greater freedom in carbon resource regulation, enabling them to flexibly respond to carbon market fluctuations, find the best balance between carbon reduction and emission reduction and economic benefits, and promote enterprises to achieve synergistic efficiency in pollution reduction and carbon reduction.

[0250] Step S3000, set the accounting threshold and target value of the total carbon emissions. When the calculated value of the total carbon emissions exceeds the accounting threshold or the predicted value of the total carbon emissions exceeds the target value, implement water-saving transformation of the water supply system; regularly evaluate the changes in carbon emissions before and after the water-saving transformation of the water supply system, and quantify the carbon reduction effect of the water-saving transformation.

[0251] Furthermore, step S3000 includes:

[0252] Step S3100, setting a calculation threshold and target value for the total carbon emissions, and implementing water-saving transformation of the water supply system when the total carbon emissions calculation value exceeds the calculation threshold or the total carbon emissions forecast value exceeds the target value;

[0253] Specifically, step S3100 sets the conditions for triggering the water-saving and carbon-reduction feedback mechanism based on the total carbon emissions accounting and forecasting results. First, it is necessary to set two reference benchmarks: the carbon emissions accounting threshold and the target value. The accounting threshold reflects the upper limit of the normal range. Exceeding the accounting threshold means that carbon emissions are abnormally high, and water-saving and carbon-reduction measures are urgently needed. The target value represents the direction of efforts in water-saving and carbon-reduction work, and is a strategic goal that requires long-term persistence to achieve. The accounting threshold mainly refers to historical data for the same period, and years with higher carbon emissions are selected as the benchmark; the target value needs to comprehensively consider water-saving and carbon-reduction plans, carbon peak targets, carbon neutrality visions, etc., and formulate practical and realistic phased indicators.

[0254] After setting the accounting threshold and target value, the actual calculated carbon emissions are compared with the accounting threshold to determine whether the current carbon emission intensity is abnormal; the predicted future carbon emissions are compared with the target value to determine whether the water-saving and carbon-reduction tasks can be completed under the current trend. When the actual value exceeds the threshold or the predicted value deviates from the target value, it means that the carbon emission level of the water supply system is high, the water-saving and carbon-reduction situation is severe, and timely response measures are needed, which will trigger the subsequent water-saving and carbon-reduction feedback mechanism and implement water-saving transformation of the water supply system.

[0255] The core of the feedback mechanism is to push personalized water-saving and carbon-reduction suggestions to water supply plants and large water users. For water supply plants, measures such as optimizing equipment operating parameters, transforming high-energy-consuming equipment, and repairing pipe network leaks can be pushed to guide the water supply system to improve quality and efficiency; for large water users, measures such as optimizing water use time, selecting water-saving appliances, and finding indoor leaks can be pushed to guide users to develop water-saving habits. At the same time, advanced units and individuals who adopt suggestions and achieve carbon reduction targets can also be commended and rewarded, so as to mobilize the enthusiasm of all parties to participate in water-saving and carbon reduction in a positive incentive way.

[0256] In general, step S3100 sets thresholds and targets to warn of abnormal carbon emissions and guide water conservation and carbon reduction work. When the warning is triggered, a closed-loop feedback loop is formed through personalized suggestion push and reward mechanism, which mobilizes the endogenous motivation of water supply plants and large water users to participate in water conservation and carbon reduction, builds a water conservation and carbon reduction pattern jointly promoted by the government, enterprises, and citizens, forms a joint force of carbon reduction at the source and water conservation at the terminal, and jointly promotes the water supply system to achieve carbon peak and carbon neutrality goals from the supply side and the demand side.

[0257] Step S3200, regularly evaluate the changes in carbon emissions before and after the water-saving transformation of the water supply system, and quantify the carbon reduction effect of the water-saving transformation.

[0258] Furthermore, if Figure 6 As shown, step S3200 includes:

[0259] Step S3210, collecting the operation data of water supply plants and large water users before and after water-saving transformation, and calculating the changes in the total carbon emissions before and after the implementation of water-saving measures;

[0260] Step S3220, calculating the difference between the total carbon emissions before the water-saving transformation and the total carbon emissions after the water-saving transformation, and obtaining the carbon reduction amount of the water-saving measures;

[0261] Step S3230, calculating the ratio of carbon reduction of water-saving measures to the total carbon emissions before water-saving transformation, and obtaining the carbon reduction rate of water-saving measures;

[0262] Step S3240, evaluating the carbon reduction effect of the water-saving transformation based on the carbon reduction amount and carbon reduction rate, and forming a water-saving and carbon reduction assessment report.

[0263] Specifically, step S3200 systematically evaluates the effectiveness of water-saving measures and objectively quantifies the contribution of water-saving to carbon reduction. By tracking the changes in operating parameters of each link of the water supply system before and after the water-saving transformation, the direct carbon reduction brought about by water-saving measures can be calculated, and the reduction in carbon emission intensity can be evaluated. The evaluation indicators include two quantitative indicators: carbon reduction amount and carbon reduction rate. The carbon reduction amount measures the carbon reduction scale of water-saving measures from an absolute value perspective, and the carbon reduction rate examines the carbon reduction efficiency of water-saving measures from a relative value perspective. The combination of the two can fully reflect the effectiveness of water-saving and carbon reduction.

[0264] First, step S3210 collects the operating data of water supply plants and large water users before and after water-saving transformation, including water supply, pipe network leakage, power consumption, etc., and uses the carbon accounting method of step S2000 to calculate the total carbon emissions of the water supply system before and after the implementation of water-saving measures. The total carbon emissions count the carbon footprint of each link in the entire life cycle of water supply, reflecting the overall carbon reduction potential of the system.

[0265] Next, step S3220 quantifies the direct carbon reduction of water-saving measures. The carbon reduction is equal to the total carbon emissions before the water-saving transformation minus the total carbon emissions after the water-saving transformation, which represents the absolute carbon reduction contribution of water-saving measures. Taking the transformation of the water supply network as an example, the leakage rate of the network can be significantly reduced by replacing old pipe sections, repairing leakage points and other measures; the reduction in leakage means that the operating time of water supply pumps, water purification equipment, etc. is shortened, thereby reducing equipment energy consumption and reducing carbon emissions in the water supply production process. At the same time, the saved water resources can replace other high-energy-consuming water supply methods and achieve carbon emission reduction at the user end. The difference in the total carbon emissions before and after the transformation of the pipeline network is the direct carbon reduction of the measure, which quantifies the synergistic benefits of water saving and carbon reduction.

[0266] Step S3230 further calculates the carbon reduction rate of the water-saving measures, that is, the percentage of carbon reduction in the total carbon emissions before the water-saving transformation. The carbon reduction rate reflects the relative carbon reduction efficiency of water-saving measures, which facilitates the horizontal comparison of the carbon reduction effects of different measures and different periods. For example, for two water-saving measures with equivalent investments, the measure with a higher carbon reduction rate is more cost-effective; for a certain water-saving measure that is continuously monitored, if the carbon reduction rate decreases year by year, it may be necessary to adjust the intensity of the measure or seek new ways to save water. The carbon reduction rate links water-saving measures with carbon emission intensity, guides efforts towards water-saving with higher carbon reduction efficiency, and optimizes water-saving strategies under carbon constraints.

[0267] Finally, step S3240 comprehensively considers the carbon reduction amount and carbon reduction rate, evaluates the carbon reduction effect of water-saving measures, and writes a water-saving and carbon reduction assessment report. The assessment report examines the environmental benefits of water-saving measures from a carbon perspective, clarifies the inherent connection between water-saving and carbon reduction through quantitative analysis of carbon reduction, and highlights the climate change response value of water-saving measures; through horizontal and vertical comparisons of carbon reduction rates, the efficiency and benefits of water-saving measures are judged, the key paths of water-saving and carbon reduction are identified, and water-saving planning is optimized based on evidence. The assessment report can also include an analysis of the return on investment of water-saving projects, convert carbon reduction into carbon trading income, and broaden the financing channels for water-saving projects. The government can use the assessment results as a basis for adjusting and assessing water-saving policies to encourage the promotion and application of efficient water-saving technologies.

[0268] In short, step S3200 conducts a post-evaluation of water-saving measures from the perspective of carbon accounting, takes carbon reduction and carbon reduction rate as key indicators, establishes a quantitative evaluation system and dynamic monitoring mechanism for water-saving and carbon reduction, realizes the quantitative assessment of water-saving results, the optimization and iteration of water-saving schemes, and promotes the coordination of water-saving practices and carbon peak targets. Through data empowerment, water-saving policies have moved from "empirical judgment" to evidence-based decision-making, and carbon reduction effects have become an important evaluation dimension for water-saving work. The solid data foundation and standardized accounting process of the water-saving and carbon reduction assessment report make the assessment conclusions more scientific and credible, and the carbon reduction effect is more expected. The systematic assessment process can be replicated and promoted, forming a normalized water-saving and carbon reduction assessment mechanism, providing strong support for regional water-saving planning, water resources allocation in the basin, and the improvement of national water-saving policies. Step S3200 allows water-saving and carbon reduction to achieve "same frequency resonance, carbon reduction targets force the optimization of water-saving paths, and water determines carbon" to generate new momentum for water-saving, creating a new situation in which water resources management in the basin and the vision of carbon neutrality work together.

[0269] Example 2

[0270] This embodiment provides a carbon emission accounting system for urban water supply plants and water saving and carbon reduction for large water users based on embodiment 1. Figure 7 As shown, including:

[0271] Data acquisition module: used for real-time collection of first data of water supply plants, first data of large water users and itemized electricity consumption data to form second data; pre-processing the second data to obtain third data;

[0272] Feature extraction module: used to extract features from the third data to obtain artificially designed features; obtain machine learning features based on the artificially designed features; and fuse the artificially designed features with the machine learning features to construct fourth data;

[0273] Carbon emission accounting module: according to the fourth data, the carbon footprint of the water supply plant is calculated to obtain the first carbon emission data; according to the fourth data, the carbon emission of large water users is calculated to obtain the second carbon emission data; the first carbon emission data and the second carbon emission data are input into the preset carbon emission accounting model to obtain the total carbon emission accounting value;

[0274] Carbon emission prediction module: obtains the total carbon emission prediction value based on the total carbon emission accounting value, as well as the pre-built short-term carbon emission prediction model and long-term carbon emission prediction model;

[0275] Water-saving and carbon reduction assessment module: used to set the accounting threshold and target value of total carbon emissions. When the calculated value of total carbon emissions exceeds the accounting threshold or the predicted value of total carbon emissions exceeds the target value, water-saving transformation of the water supply system is implemented; regular assessment of carbon emission changes before and after the water-saving transformation of the water supply system is conducted to quantify the carbon reduction effect of the water-saving transformation.

[0276] In the data acquisition module, the real-time collection of the first data of the water supply plant, the first data of large water users and the itemized electricity consumption data to form the second data includes: installing intelligent sensors at key nodes of the water supply plant and large water users to collect the first data in real time, and at the same time collecting the itemized electricity consumption data of the water supply plant and large water users; using the first data and the itemized electricity consumption data as the second data; the first data include the first data of the water supply plant and the first data of large water users; the first data of the water supply plant include water flow data, pipe network pressure data and equipment power consumption data; the first data of large water users include total water consumption data, itemized water consumption data and water distribution pump power consumption data.

[0277] In the feature extraction module, the artificially designed features include characteristic indicators of energy consumption intensity of water supply plants and characteristic indicators of water use patterns of large water users; the characteristic indicators of energy consumption intensity of water supply plants include unit water supply electricity consumption, net water consumption and pipe network leakage rate; the characteristic indicators of water use patterns of large water users include peak water consumption, water balance coefficient and minimum water consumption at night.

[0278] The method of obtaining machine learning features based on artificially designed features includes:

[0279] Step S1321, centralize the artificially designed features, shift the mean of each feature to 0, and obtain the original feature matrix X;

[0280] Step S1322, calculating the covariance matrix of the centered features;

[0281] Step S1323, performing eigenvalue decomposition on the covariance matrix to obtain eigenvalues ​​and eigenvectors;

[0282] Step S1324, sort the eigenvalues ​​from large to small, and select the eigenvectors corresponding to the first k largest eigenvalues ​​to form the transformation matrix P;

[0283] Step S1325, multiplying the original feature matrix X after centralization by the transformation matrix P on the left to obtain the principal component feature matrix Z=PX after dimensionality reduction;

[0284] Step S1326, calculating the variance contribution rate of each principal component feature according to the eigenvalue;

[0285] Step S1327, perform feature selection on the principal component feature matrix Z, remove the principal component features whose variance contribution rate is lower than a preset variance contribution rate threshold, and obtain machine learning features.

[0286] In the carbon emission accounting module, the carbon footprint of the water supply plant is calculated according to the fourth data to obtain the first carbon emission data, including:

[0287] Step S2110, calculating the carbon emissions of the electricity use of the water supply plant according to the unit water supply power consumption of the water supply plant in the fourth data and the real-time carbon emission factor of the power grid; the carbon emissions of the electricity use of the water supply plant are equal to the product of the unit water supply power consumption and the water supply and the real-time carbon emission factor of the power grid;

[0288] Step S2120, calculating the carbon emissions of the water intake link of the water supply plant according to the water consumption of the water supply plant in the fourth data;

[0289] Step S2130, calculating the carbon emissions of the water supply plant's transmission and distribution link according to the pipe network leakage rate of the water supply plant in the fourth data;

[0290] Step S2140, summing up the carbon emissions of the electricity use, water intake and transmission and distribution links of the water supply plant to obtain the carbon footprint of the water supply plant and form the first carbon emission data.

[0291] In the carbon emission accounting module, the carbon emission of large water users is calculated according to the fourth data to obtain the second carbon emission data, including:

[0292] Step S2210, calculating the carbon emissions of transmission and distribution during the peak period according to the peak water consumption of large water users in the fourth data;

[0293] Step S2220, calculating the balanced carbon emissions of transmission and distribution energy consumption according to the water balance coefficient of the large water users in the fourth data;

[0294] Step S2230, calculating the background leakage carbon emissions of the pipe network according to the minimum nighttime water consumption of the large water users in the fourth data;

[0295] Step S2240, adding the carbon emissions of transmission and distribution during peak hours, the balanced carbon emissions of transmission and distribution energy consumption, and the background leakage carbon emissions of the pipeline network to obtain the carbon emission footprint of large water users and form the second carbon emission data.

[0296] In the carbon emission prediction module, the total carbon emission prediction value is obtained based on the total carbon emission accounting value, and the pre-built short-term carbon emission prediction model and long-term carbon emission prediction model, including:

[0297] Step S2310, inputting the first carbon emission data and the second carbon emission data into a preset carbon emission accounting model to obtain a total carbon emission accounting value;

[0298] Step S2320, obtaining weather data and holiday data, building a short-term carbon emission prediction model, and predicting daily carbon emissions in the future short-term period based on the total carbon emission accounting value, weather data, holiday data and the short-term carbon emission prediction model;

[0299] Step S2330, obtaining long-term prediction related data, building a long-term carbon emission prediction model, and predicting the monthly carbon emissions in the future long-term period based on the total carbon emission accounting value, the long-term prediction related data and the long-term carbon emission prediction model; the long-term prediction related data includes seasonal regularity data, statutory holiday data and facility maintenance plan data;

[0300] Step S2340, the daily carbon emissions and the monthly carbon emissions are superimposed and combined to form a total carbon emissions forecast value.

[0301] In the water-saving and carbon reduction assessment module, the regular assessment of the carbon emission changes before and after the water-saving transformation of the water supply system and the quantification of the carbon reduction effect of the water-saving transformation include:

[0302] Step S3210, collecting the operation data of water supply plants and large water users before and after water-saving transformation, and calculating the changes in the total carbon emissions before and after the implementation of water-saving measures;

[0303] Step S3220, calculating the difference between the total carbon emissions before the water-saving transformation and the total carbon emissions after the water-saving transformation, and obtaining the carbon reduction amount of the water-saving measures;

[0304] Step S3230, calculating the ratio of carbon reduction of water-saving measures to the total carbon emissions before water-saving transformation, and obtaining the carbon reduction rate of water-saving measures;

[0305] Step S3240, evaluating the carbon reduction effect of the water-saving transformation based on the carbon reduction amount and carbon reduction rate, and forming a water-saving and carbon reduction assessment report.

[0306] Example 3

[0307] This embodiment discloses an electronic device, which may include one or more processors and one or more memories. The memories store computer-readable codes, which, when executed by the one or more processors, may execute the above-mentioned method for calculating carbon emissions of urban water supply plants and water-saving and carbon-reduction for large water users.

[0308] The method or system according to the implementation mode of the present application can also be implemented with the help of the architecture of an electronic device. The electronic device may include a bus, one or more CPUs, a read-only memory (ROM), a random access memory (RAM), a communication port connected to a network, an input / output component, a hard disk, etc. The storage device in the electronic device, such as a ROM or a hard disk, can store the urban water supply plant carbon emissions and the water-saving and carbon reduction accounting method for large water users provided in this application. The urban water supply plant carbon emissions and the water-saving and carbon reduction accounting method for large water users may, for example, include: real-time collection of first data of the water supply plant, first data of large water users and itemized electricity consumption data to form second data; preprocessing the second data to obtain third data; extracting features from the third data to obtain artificially designed features; based on the artificially designed features, obtaining machine learning features; fusing the artificially designed features with the machine learning features to construct fourth data; based on the fourth data, calculating the carbon footprint of the water supply plant to obtain first carbon emission data; based on the fourth data, calculating the carbon emissions of large water users to obtain the second carbon emission data; input the first carbon emission data and the second carbon emission data into a preset carbon emission accounting model to obtain a total carbon emission accounting value; obtain a total carbon emission forecast value based on the total carbon emission accounting value, as well as a pre-constructed short-term carbon emission prediction model and a long-term carbon emission prediction model; set a total carbon emission accounting threshold and target value, and when the total carbon emission accounting value exceeds the accounting threshold or the total carbon emission forecast value exceeds the target value, implement water-saving transformation of the water supply system; regularly evaluate the changes in carbon emissions before and after the water-saving transformation of the water supply system, and quantify the carbon reduction effect of the water-saving transformation.

[0309] Furthermore, the electronic device may also include a user interface. Of course, the architecture disclosed in the present invention is only exemplary, and when implementing different devices, one or more components in the electronic device disclosed in the present invention may be omitted according to actual needs.

[0310] Example 4

[0311] This embodiment discloses a computer-readable storage medium, on which computer-readable instructions are stored. When the computer-readable instructions are executed by a processor, the carbon emissions of urban water supply plants and the water-saving and carbon reduction accounting method for large water users according to the implementation of the present application can be executed. The storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory (cache). Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0312] In addition, according to the implementation mode of the present application, the process described in the above reference flowchart can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium, the non-transitory machine-readable storage medium stores machine-readable instructions, and the machine-readable instructions can be executed by a processor to execute instructions corresponding to the method steps provided in the present application, for example: real-time collection of first data of the water supply plant, first data of large water users and itemized electricity consumption data to form second data; pre-processing the second data to obtain third data; extracting features from the third data to obtain artificially designed features; based on the artificially designed features, obtaining machine learning features; fusing the artificially designed features with the machine learning features to construct fourth data; and calculating the carbon footprint of the water supply plant based on the fourth data. , obtain the first carbon emission data; according to the fourth data, calculate the carbon emissions of large water users to obtain the second carbon emission data; input the first carbon emission data and the second carbon emission data into the preset carbon emission accounting model to obtain the total carbon emission accounting value; according to the total carbon emission accounting value, and the pre-built short-term carbon emission prediction model and long-term carbon emission prediction model, obtain the total carbon emission prediction value; set the total carbon emission accounting threshold and target value, when the total carbon emission accounting value exceeds the accounting threshold or the total carbon emission prediction value exceeds the target value, implement water-saving transformation of the water supply system; regularly evaluate the carbon emission changes before and after the water-saving transformation of the water supply system, and quantify the carbon reduction effect of the water-saving transformation. When the computer program is executed by the central processing unit (CPU), the above functions defined in the method of this application are executed.

[0313] The methods, systems, and devices of the present application may be implemented in many ways. For example, the methods, systems, and devices of the present application may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above order of steps for the method is for illustration only, and the steps of the method of the present application are not limited to the order specifically described above unless otherwise specified. In addition, in some embodiments, the present application may also be implemented as programs recorded in a recording medium, which include machine-readable instructions for implementing the method according to the present application. Therefore, the present application also covers recording media storing programs for executing the method according to the present application.

[0314] In addition, the parts of the above-mentioned technical solutions provided in the embodiments of the present application that are consistent with the implementation principles of the corresponding technical solutions in the prior art are not described in detail to avoid excessive redundancy.

[0315] The specific implementation modes as described above further describe the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation mode of the present invention and is not intended to limit the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. The method for calculating carbon emissions from urban water supply plants and water conservation and carbon reduction for large water users is characterized by: The method comprises: Collect the first data of the water supply plant, the first data of the large water user and the electricity consumption data by item in real time to form the second data; pre-process the second data to obtain the third data; extract features from the third data to obtain artificial design features; obtain machine learning features based on the artificial design features; integrate the artificial design features with the machine learning features to construct the fourth data; According to the fourth data, the carbon footprint of the water supply plant is calculated to obtain the first carbon emission data; according to the fourth data, the carbon emissions of large water users are calculated to obtain the second carbon emission data; the first carbon emission data and the second carbon emission data are input into a preset carbon emission accounting model to obtain a total carbon emission accounting value; according to the total carbon emission accounting value, as well as the pre-constructed short-term carbon emission prediction model and the long-term carbon emission prediction model, a total carbon emission prediction value is obtained; Set accounting thresholds and target values ​​for total carbon emissions. When the accounting value of total carbon emissions exceeds the accounting threshold or the predicted value of total carbon emissions exceeds the target value, implement water-saving transformation of the water supply system; regularly evaluate the changes in carbon emissions before and after the water-saving transformation of the water supply system, and quantify the carbon reduction effect of the water-saving transformation; The artificial design characteristics include energy consumption intensity characteristic indicators of water supply plants and water use pattern characteristic indicators of large water users; the energy consumption intensity characteristic indicators of water supply plants include unit water supply electricity consumption, net water consumption and pipe network leakage rate; the water use pattern characteristic indicators of large water users include peak water consumption, water balance coefficient and minimum water consumption at night; The method of obtaining machine learning features based on artificially designed features includes: The artificially designed features are centralized, and the mean of each feature is shifted to 0 to obtain the original feature matrix X; the covariance matrix of the centralized features is calculated; the covariance matrix is ​​eigenvalue decomposed to obtain eigenvalues ​​and eigenvectors; the eigenvalues ​​are sorted from large to small, and the eigenvectors corresponding to the first k largest eigenvalues ​​are selected to form the transformation matrix P; The transformation matrix P is used to multiply the original feature matrix X after centralization on the left to obtain the principal component feature matrix Z=PX after dimensionality reduction; the variance contribution rate of each principal component feature is calculated according to the eigenvalue; feature selection is performed on the principal component feature matrix Z, and the principal component features whose variance contribution rate is lower than the preset variance contribution rate threshold are eliminated to obtain machine learning features.

2. The method for calculating carbon emissions from urban water supply plants and water saving and carbon reduction for large water users according to claim 1 is characterized in that: The calculating of the carbon footprint of the water supply plant according to the fourth data to obtain the first carbon emission data includes: The carbon emissions of the electricity use of the water supply plant are calculated based on the unit water supply power consumption of the water supply plant in the fourth data and the real-time carbon emission factor of the power grid; the carbon emissions of the electricity use of the water supply plant are equal to the product of the unit water supply power consumption and the water supply and the real-time carbon emission factor of the power grid; According to the water consumption of the water supply plant in the fourth data, calculate the carbon emissions of the water supply plant in the water intake link; According to the pipe network leakage rate of the water supply plant in the fourth data, calculate the carbon emissions of the water supply plant's transmission and distribution links; The carbon emissions from the electricity usage, water intake and transmission and distribution links of the water supply plant are summed up to obtain the carbon footprint of the water supply plant, forming the first carbon emission data.

3. The method for calculating carbon emissions from urban water supply plants and water saving and carbon reduction for large water users according to claim 2 is characterized in that: The calculation of the carbon emissions of the water supply plant's transmission and distribution link according to the pipe network leakage rate of the water supply plant in the fourth data includes: The process life cycle assessment method is used to calculate the total carbon emissions of the pipeline network throughout its life cycle. The total carbon emissions of the pipeline network throughout its life cycle are divided by the water supply to obtain the carbon intensity per unit water supply, which is then multiplied by the pipeline leakage rate to obtain the carbon emissions of the water supply plant's transmission and distribution links.

4. The method for calculating carbon emissions from urban water supply plants and water saving and carbon reduction for large water users according to claim 3 is characterized in that: The second carbon emission data obtained by calculating the carbon emissions of large water users based on the fourth data includes: Based on the peak water consumption of large water users in the fourth data, calculate the carbon emissions of transmission and distribution during peak hours; According to the water balance coefficient of large water users in the fourth data, calculate the balanced carbon emissions of transmission and distribution energy consumption; Based on the minimum nighttime water consumption of large water users in the fourth data, calculate the background leakage carbon emissions of the pipe network; By adding together the carbon emissions from transmission and distribution during peak hours, the balanced carbon emissions from transmission and distribution energy consumption, and the background leakage carbon emissions of the pipeline network, we can obtain the carbon emission footprint of large water users and form the second carbon emission data.

5. The method for calculating carbon emissions from urban water supply plants and water saving and carbon reduction for large water users according to claim 4 is characterized in that: The calculation of the carbon emissions during peak hours based on the peak water consumption of large water users in the fourth data includes: Determine the peak load regulation operation mode of the water supply network; the peak load regulation operation mode of the water supply network includes two types: variable frequency speed regulation and pump start and stop; According to the peak load regulation operation mode of the water supply network and the peak water consumption of large water users, the network energy consumption during peak hours is calculated; under the variable frequency speed regulation operation mode, the network energy consumption during peak hours is proportional to the square of the peak water consumption of large water users; under the start-stop pump operation mode, the network energy consumption during peak hours is proportional to the peak water consumption of large water users; According to the real-time carbon emission factor of the power grid during peak hours, the energy consumption of the pipeline network during peak hours is converted into the transmission and distribution carbon emissions during peak hours.

6. The method for calculating carbon emissions from urban water supply plants and water saving and carbon reduction for large water users according to claim 1 is characterized in that: The predicted value of total carbon emissions includes: Obtain weather data and holiday data, build a short-term carbon emission prediction model, and predict the daily carbon emissions in the future short-term period based on the total carbon emission accounting value, weather data, holiday data and the short-term carbon emission prediction model; Obtaining long-term forecast-related data, building a long-term carbon emission forecast model, and forecasting monthly carbon emissions in the future long-term period based on the total carbon emission accounting value, long-term forecast-related data and the long-term carbon emission forecast model; the long-term forecast-related data includes seasonal regularity data, statutory holiday data and facility maintenance plans; The daily carbon emissions and monthly carbon emissions are superimposed and combined to form a forecast value of total carbon emissions.

7. A system for calculating carbon emissions from urban water supply plants and water-saving and carbon reduction for large water users, which is used to implement the method for calculating carbon emissions from urban water supply plants and water-saving and carbon reduction for large water users according to any one of claims 1 to 6, characterized in that the system comprises: Data acquisition module: used for real-time collection of first data of water supply plants, first data of large water users and itemized electricity consumption data to form second data; pre-processing the second data to obtain third data; Feature extraction module: used to extract features from the third data to obtain artificially designed features; obtain machine learning features based on the artificially designed features; and fuse the artificially designed features with the machine learning features to construct fourth data; Carbon emission accounting module: according to the fourth data, the carbon footprint of the water supply plant is calculated to obtain the first carbon emission data; according to the fourth data, the carbon emission of large water users is calculated to obtain the second carbon emission data; the first carbon emission data and the second carbon emission data are input into the preset carbon emission accounting model to obtain the total carbon emission accounting value; Carbon emission prediction module: obtains the total carbon emission prediction value based on the total carbon emission accounting value, as well as the pre-built short-term carbon emission prediction model and long-term carbon emission prediction model; Water-saving and carbon reduction assessment module: used to set the accounting threshold and target value of total carbon emissions. When the calculated value of total carbon emissions exceeds the accounting threshold or the predicted value of total carbon emissions exceeds the target value, water-saving transformation of the water supply system is implemented; regular assessment of carbon emission changes before and after the water-saving transformation of the water supply system is conducted to quantify the carbon reduction effect of the water-saving transformation.

Images