An energy consumption optimization method and system based on dynamic carbon constraints

Abstract

This invention provides an energy consumption optimization method and system based on dynamic carbon constraints. It constructs carbon emission fluctuation maps at multiple time points based on the carbon emission time-series data stream of the target region to build a multi-region minimum carbon emission demand distribution map. Based on the real-time operating status parameters of energy generation equipment within the target region, it performs multi-cycle output change analysis and short-term capacity filtering prediction to obtain an energy capacity prediction curve. It then performs nonlinear implicit feature analysis and modeling on the carbon emission time-series data stream of the target region to construct a carbon emission nested energy structure collaborative graph. Based on the carbon emission constraints, it couples the multi-region minimum carbon emission demand distribution map and the energy capacity prediction curve through supply and demand matching, and then constructs an intelligent energy allocation strategy based on the carbon emission nested energy structure collaborative graph, performing iterative allocation optimization to obtain the energy consumption optimization result. This method can address the dynamic evolution and high-dimensional complexity of the system to improve the energy consumption optimization effect.

Description

Technical Field

[0001] This invention relates to the field of energy optimization technology, and in particular to an energy consumption optimization method and system based on dynamic carbon constraints. Background Technology

[0002] As global climate change becomes increasingly severe, new energy sources such as wind, solar, and hydropower are playing an increasingly important role in renewable energy systems, smart grids, and the construction of new power systems. However, due to factors such as changes in the natural environment, uneven distribution of resources, and technological bottlenecks, the utilization rate of new energy sources still faces many challenges. In particular, during the dispatch and operation of power systems, problems such as insufficient new energy absorption capacity and wind and electricity curtailment remain prominent. Therefore, improving the maximum utilization rate of new energy sources has become an important research topic for the current intelligent development of energy systems.

[0003] However, in the actual dispatch and operation of new energy sources, power systems are often subject to multiple constraints, especially carbon emission constraints. This leads to problems such as insufficient performance and poor adaptability of traditional dispatch optimization methods in practical applications. Carbon constraints, as a key indicator for measuring the environmental friendliness of energy systems, will significantly improve the environmental friendliness and sustainability of system operation when introduced into optimization models, promoting the transformation of the energy structure towards a low-carbon direction. However, in the multi-source, time-varying, and uncertain power system environment, a single optimization algorithm is difficult to cope with the dynamic evolution and high-dimensional complexity of the system. Especially when facing the requirements of dispatch and rapid response, traditional optimization strategies are inadequate. Therefore, there is an urgent need for a more effective energy consumption optimization method. Summary of the Invention

[0004] This invention provides an energy consumption optimization method and system based on dynamic carbon constraints, which aims to improve the energy consumption optimization effect in order to cope with the dynamic evolution and high-dimensional complexity of the system.

[0005] To achieve the above objectives, the present invention provides an energy consumption optimization method based on dynamic carbon constraints, comprising:

[0006] Step 1: Calculate the total carbon emissions based on the carbon emission time series data stream of the target area, and perform multi-time point abrupt change fitting based on the total carbon emissions to construct a carbon emission fluctuation map at multiple time points;

[0007] Step 2: Calculate the minimum carbon emission demand for multiple time periods based on the carbon emission fluctuation map, and conduct regional demand distribution analysis to construct a multi-regional minimum carbon emission demand distribution map.

[0008] Step 3: Based on the real-time operating status parameters of energy power generation equipment in the target area, perform multi-cycle output change analysis and short-term capacity filtering prediction to obtain the energy capacity prediction curve.

[0009] Step 4: Based on real-time operating status parameters, perform nonlinear implicit feature analysis and modeling on the carbon emission time series data stream of the target area, and construct a carbon emission nested energy structure collaborative graph;

[0010] Step 5: Based on the carbon emission constraints, the minimum carbon emission demand distribution map of multiple regions and the energy production capacity prediction curve are coupled for supply and demand matching to obtain potential supply and demand conflict points. Based on the carbon emission nested energy structure coordination diagram, the potential supply and demand conflict points are allocated to construct an intelligent energy allocation strategy.

[0011] Step 6: Based on the intelligent energy allocation strategy, perform carbon emission scheduling and carbon emission feedback information collection to obtain the carbon efficiency deviation curve, and iteratively optimize the intelligent energy allocation strategy based on the carbon efficiency deviation curve to obtain the energy consumption optimization result.

[0012] Furthermore, step 1 includes:

[0013] Carbon emission data of the target area is monitored based on a distributed multi-source carbon emission sensor network to obtain a carbon emission time series data stream;

[0014] Dynamic time-frequency decomposition of carbon emission time-series data stream yields multiple carbon emission monitoring time windows;

[0015] The total carbon emissions are calculated for the carbon emission time series data stream within each carbon emission monitoring time window, thus obtaining the total carbon emissions for each carbon emission monitoring time window;

[0016] For each carbon emission monitoring time window, the total carbon emissions are fitted with a time-series fluctuation to generate a total carbon emission waveform curve;

[0017] Multi-time-point abrupt change fitting was performed on the total carbon emission fluctuation curve to construct a carbon emission fluctuation map at multiple time points.

[0018] Furthermore, the total carbon emissions for each carbon emission monitoring time window are fitted with a time-series fluctuation to generate a total carbon emission waveform curve, including:

[0019] The total carbon emissions for each carbon emission monitoring time window are used as input to the fitting model;

[0020] The parameters of the fitted model are estimated using the least squares method, and the best fitting parameters of the fitted model are calculated.

[0021] A waveform curve of total carbon emissions is generated based on the best-fit parameters.

[0022] Furthermore, multi-time-point abrupt change fitting is performed on the total carbon emission fluctuation curve to construct carbon emission fluctuation maps at multiple time points, including:

[0023] The slope change and derivative analysis methods were used to identify and mark abrupt changes in carbon emissions on the total carbon emission waveform curve.

[0024] The carbon emission mutation points are calculated to obtain the gradient change rate used to quantify the intensity of the mutation point change and the mutation duration used to reflect the duration of the mutation effect, and the timestamp of the occurrence of the carbon emission mutation point is extracted.

[0025] Based on the occurrence timestamp, the carbon emission monitoring time window is located, and the window distribution of the carbon emission monitoring time window is analyzed to obtain the distribution characteristics of the abrupt change point window;

[0026] Based on the distribution characteristics of mutation point windows, gradient change rate, and mutation duration, carbon emission mutation point features are mined to obtain multi-dimensional carbon emission mutation point features.

[0027] Based on the characteristics of multi-dimensional carbon emission mutation points, the waveform curve of total carbon emissions is fitted with multiple time-point mutations to construct a carbon emission fluctuation map at multiple time points.

[0028] Furthermore, step 2 includes:

[0029] Based on the carbon emission fluctuation map, the carbon emission trend changes in different regions were analyzed to obtain the characteristics of carbon emission trend changes in different regions.

[0030] The evolution of carbon emission trends in all regions is predicted, resulting in carbon emission trend evolution prediction curves.

[0031] Based on the carbon emission status evolution prediction curve, multi-regional carbon emission demand analysis is conducted to generate carbon emission demand data for multiple regions.

[0032] The minimum carbon emission demand for each region is calculated over multiple time periods to obtain the minimum carbon emission demand value for each region over multiple time periods.

[0033] The minimum carbon emission demand values ​​for multiple time periods in all regions are fitted with regional demand distribution to construct a multi-regional minimum carbon emission demand distribution map.

[0034] Furthermore, step 3 includes:

[0035] Monitor the real-time operating status parameters of each energy power generation device within the target area;

[0036] The real-time operating status parameters are calculated to obtain the energy production cycle, power generation of each energy power generation device, energy inventory and operating efficiency. The power generation, energy inventory and operating efficiency of each energy power generation device are then fitted with a time series to obtain a resource dynamic data sequence.

[0037] Based on the energy production cycle, multi-cycle output change analysis is performed on the dynamic data sequence of resources to obtain the output fluctuation characteristics of each cycle.

[0038] Short-term capacity filtering forecasts are performed on the output fluctuation characteristics of each cycle to obtain the energy capacity forecast curve.

[0039] Furthermore, step 4 includes:

[0040] The dispatchable energy resources for multiple time windows are calculated based on real-time operating status parameters.

[0041] A time-series carbon budget allocation status analysis was performed on the carbon emission time-series data stream of the target area to obtain the carbon budget allocation status at multiple time points.

[0042] Nonlinear implicit feature analysis is performed on the carbon budget allocation status and dispatchable energy at each time point to identify the implicit correlation between the carbon budget and dispatchable energy.

[0043] Based on the correlation features, a nested correlation model of carbon emissions and energy is constructed to build a carbon emission nested energy structure correlation graph.

[0044] Furthermore, step 5 includes:

[0045] Based on the carbon emission fluctuation maps at multiple time points, an adaptive carbon emission upper limit constraint is defined to obtain the carbon emission constraint conditions for the target area.

[0046] Calculate the maximum energy utilization rate based on the energy production capacity forecast curve;

[0047] Based on carbon emission constraints, a spatiotemporal energy supply and demand matching model is constructed by coupling supply and demand to maximize energy utilization.

[0048] Multi-scenario simulations were conducted on the spatiotemporal energy supply and demand matching model to identify potential supply and demand conflict points.

[0049] Based on the carbon emission nested energy structure coordination diagram, a multi-objective function intelligent allocation process is performed on potential supply and demand conflict points to construct an intelligent energy allocation strategy.

[0050] Furthermore, step 6 includes:

[0051] Carbon emission scheduling and execution operations are carried out based on intelligent energy allocation strategies, and carbon emission feedback information is collected.

[0052] Based on carbon emission feedback information, the power grid carbon emission optimization efficiency is calculated to obtain the global carbon efficiency vector of the power grid.

[0053] The carbon emission efficiency deviation is calculated based on the preset carbon budget target vector to obtain the carbon efficiency deviation curve.

[0054] Based on the carbon efficiency deviation curve, the energy intelligent allocation strategy is iteratively optimized and subjected to global long-term evolution learning to obtain energy consumption optimization results.

[0055] This invention also provides an energy consumption optimization system based on dynamic carbon constraints, comprising:

[0056] The mutation fitting module is used to calculate the total carbon emissions based on the carbon emission time series data stream of the target area, and to perform multi-time point mutation fitting based on the total carbon emissions to construct a carbon emission fluctuation map at multiple time points.

[0057] The demand analysis module is used to calculate the minimum carbon emission demand for multiple time periods based on the carbon emission fluctuation map, and to perform regional demand distribution analysis to construct a multi-region minimum carbon emission demand distribution map.

[0058] The capacity forecasting module is used to perform multi-cycle output change analysis and short-term capacity filtering forecast based on the real-time operating status parameters of energy power generation equipment in the target area, and to obtain the energy capacity forecasting curve.

[0059] The feature analysis module is used to perform nonlinear implicit feature analysis and modeling on the carbon emission time series data stream of the target area based on real-time operating status parameters, and to construct a carbon emission nested energy structure collaborative graph.

[0060] The allocation module is used to couple the minimum carbon emission demand distribution map and energy capacity forecast curve of multiple regions according to carbon emission constraints, obtain potential supply and demand conflict points, and allocate potential supply and demand conflict points based on the carbon emission nested energy structure coordination diagram to construct an intelligent energy allocation strategy.

[0061] The allocation optimization module is used to perform carbon emission scheduling and execution operations and collect carbon emission feedback information based on the intelligent energy allocation strategy to obtain the carbon efficiency deviation curve. Based on the carbon efficiency deviation curve, the intelligent energy allocation strategy is iteratively optimized to obtain the energy consumption optimization result.

[0062] The above-described solution of the present invention has the following beneficial effects:

[0063] This invention calculates total carbon emissions based on time-series carbon emission data streams of the target region, and performs multi-time-point abrupt change fitting based on the total carbon emissions to construct carbon emission fluctuation maps at multiple time points for calculating minimum carbon emission demand over multiple time periods. It also performs regional demand distribution analysis to construct a multi-regional minimum carbon emission demand distribution map. Based on real-time operating status parameters of energy generation equipment within the target region, it performs multi-cycle output change analysis and short-term capacity filtering prediction to obtain an energy capacity prediction curve. Based on real-time operating status parameters, it performs nonlinear implicit feature analysis and modeling on the time-series carbon emission data streams of the target region to construct a carbon emission nested energy structure coordination diagram. According to carbon emission constraints, it couples the multi-regional minimum carbon emission demand distribution map and the energy capacity prediction curve for supply and demand matching to obtain potential supply and demand conflict points, and then analyzes these potential supply and demand conflict points based on the carbon emission nested energy structure coordination diagram. This invention constructs an intelligent energy allocation strategy; based on this strategy, carbon emission scheduling is performed and carbon emission feedback information is collected to obtain a carbon efficiency deviation curve. The intelligent energy allocation strategy is then iteratively optimized based on this curve to achieve energy consumption optimization. Compared to existing technologies, this invention effectively extracts information on carbon emission trend changes, abnormal disturbances, and periodic fluctuations through multi-time-point mutation fitting. By constructing a multi-regional minimum carbon emission demand distribution map and using the minimum carbon emission demand as a boundary condition, the optimization strategy is ensured to be feasible within the carbon reduction target range. Based on carbon emission constraints, the multi-regional minimum carbon emission demand distribution map and energy production capacity prediction curve are coupled for supply and demand matching. Iterative optimization of the intelligent energy allocation strategy based on the carbon efficiency deviation curve can address the dynamic evolution and high-dimensional complexity of the system, thereby improving energy consumption optimization.

[0064] Other beneficial effects of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0065] Figure 1 This is a flowchart illustrating an embodiment of the present invention. Detailed Implementation

[0066] To make the technical problems, solutions, and advantages of this invention clearer, a detailed description will be provided below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0067] Furthermore, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0068] This invention addresses existing problems by providing an energy consumption optimization method and system based on dynamic carbon constraints.

[0069] like Figure 1 As shown, an embodiment of the present invention provides an energy consumption optimization method based on dynamic carbon constraints, comprising:

[0070] Step 1: Calculate the total carbon emissions based on the carbon emission time series data stream of the target area, and perform multi-time point abrupt change fitting based on the total carbon emissions to construct a carbon emission fluctuation map at multiple time points;

[0071] Step 2: Calculate the minimum carbon emission demand for multiple time periods based on the carbon emission fluctuation map, and conduct regional demand distribution analysis to construct a multi-regional minimum carbon emission demand distribution map.

[0072] Step 3: Based on the real-time operating status parameters of energy power generation equipment in the target area, perform multi-cycle output change analysis and short-term capacity filtering prediction to obtain the energy capacity prediction curve.

[0073] Step 4: Based on real-time operating status parameters, perform nonlinear implicit feature analysis and modeling on the carbon emission time series data stream of the target area, and construct a carbon emission nested energy structure collaborative graph;

[0074] Step 5: Based on the carbon emission constraints, the minimum carbon emission demand distribution map of multiple regions and the energy production capacity prediction curve are coupled for supply and demand matching to obtain potential supply and demand conflict points. Based on the carbon emission nested energy structure coordination diagram, the potential supply and demand conflict points are allocated to construct an intelligent energy allocation strategy.

[0075] Step 6: Based on the intelligent energy allocation strategy, perform carbon emission scheduling and carbon emission feedback information collection to obtain the carbon efficiency deviation curve, and iteratively optimize the intelligent energy allocation strategy based on the carbon efficiency deviation curve to obtain the energy consumption optimization result.

[0076] Specifically, step 1 includes:

[0077] Carbon emission data of the target area is monitored based on a distributed multi-source carbon emission sensor network to obtain a carbon emission time series data stream;

[0078] Dynamic time-frequency decomposition of carbon emission time-series data stream yields multiple carbon emission monitoring time windows;

[0079] The total carbon emissions are calculated for the carbon emission time series data stream within each carbon emission monitoring time window, thus obtaining the total carbon emissions for each carbon emission monitoring time window;

[0080] For each carbon emission monitoring time window, the total carbon emissions are fitted with a time-series fluctuation to generate a total carbon emission waveform curve;

[0081] Multi-time-point abrupt change fitting was performed on the total carbon emission fluctuation curve to construct a carbon emission fluctuation map at multiple time points.

[0082] Before performing dynamic time-frequency decomposition on the carbon emission time-series data stream, this embodiment of the invention further includes: performing preliminary cleaning on the data to remove outliers and missing values; if the carbon emission in a certain hour is abnormally higher than 300 tons, it needs to be manually reviewed and processed; and storing the cleaned data in the database to form a structured carbon emission time-series data stream.

[0083] In this embodiment of the invention, the distributed multi-source carbon emission sensing network includes CO2 sensors, gas flow meters, etc., to ensure accurate monitoring of various carbon emission data. CO2 sensors and gas flow meters are deployed in the target area. In this embodiment of the invention, 50 sensors are set up to cover 5 areas including cities, industrial zones, and transportation hubs as target areas. Each type of sensor records carbon emission data once per minute, including parameters such as CO2 concentration, gas flow rate, and temperature. Assuming that each sensor generates one data record per minute, the target area can generate 72,000 carbon emission data points as a carbon emission time-series data stream within 24 hours.

[0084] It should be noted that the purpose of dynamic time-frequency decomposition is to convert the acquired carbon emission time-series data stream into multiple time windows. In this embodiment of the invention, a suitable window function (such as Hanning window or Hamming window) is selected, and the window size is set to 10 minutes with an overlap rate of 50%. The carbon emission time-series data stream is dynamically decomposed using methods such as short-time Fourier transform or wavelet transform to obtain multiple carbon emission monitoring time windows. It is assumed that 12 time windows are generated within 24 hours, and each window corresponds to 10 minutes of carbon emission time-series data.

[0085] This invention, in its embodiment, iterates through each time window, extracts the carbon emission time-series data within that window, sums the carbon emission time-series data within each time window, and calculates the total carbon emissions. For example, if the extracted carbon emission time-series data within a certain window is [10, 12, 11, 14, 13] (unit: tons), then the total carbon emissions for that window are 10 + 12 + 11 + 14 + 13 = 60 tons.

[0086] In this embodiment of the invention, polynomial fitting is selected. The order of the polynomial is set (such as second or third order) to fit the carbon emission time series fluctuation of the total carbon emission for each carbon emission monitoring time window, and a total carbon emission waveform curve is generated to ensure that the total carbon emission waveform curve can capture the changes in the data better.

[0087] Specifically, the total carbon emissions for each carbon emission monitoring time window are fitted with a time-series fluctuation to generate a total carbon emission waveform curve, including:

[0088] The total carbon emissions for each carbon emission monitoring time window are used as the fitting model y = ax 2 Inputting +bx+c;

[0089] The parameters of the fitted model are estimated using the least squares method, and the best fitting parameters a, b, and c of the fitted model are calculated.

[0090] A waveform curve of total carbon emissions is generated based on the best-fit parameters to reflect the changing trend of carbon emission time-series data.

[0091] Specifically, the total carbon emission fluctuation curve is fitted with multi-time-point abrupt changes to construct carbon emission fluctuation maps at multiple time points, including:

[0092] The slope change and derivative analysis methods were used to identify and mark abrupt changes in carbon emissions on the total carbon emission waveform curve.

[0093] The carbon emission mutation points are calculated to obtain the gradient change rate used to quantify the intensity of the mutation point change and the mutation duration used to reflect the duration of the mutation effect, and the timestamp of the occurrence of the carbon emission mutation point is extracted.

[0094] Based on the occurrence timestamp, the carbon emission monitoring time window is located, and the window distribution of the carbon emission monitoring time window is analyzed to obtain the distribution characteristics of the abrupt change point window;

[0095] Based on the distribution characteristics of mutation point windows, gradient change rate, and mutation duration, carbon emission mutation point features are mined to obtain multi-dimensional carbon emission mutation point features.

[0096] Based on the characteristics of multi-dimensional carbon emission mutation points, the waveform curve of total carbon emissions is fitted with multiple time-point mutations to construct a carbon emission fluctuation map at multiple time points.

[0097] The purpose of identifying abrupt change points in this invention is to find significant change points in the time series data of carbon emissions by analyzing the waveform curve of total carbon emissions. These change points may be related to abnormal times or operational changes. Therefore, slope change and derivative analysis methods are used to identify abrupt change points. The rate of change of carbon emissions is obtained by calculating the derivative of the carbon emission waveform curve. Assuming that the time series data of the waveform curve is T and the total carbon emissions are C, the derivative D can be expressed as: D(t)=(C(t+Δt)-C(t)) / Δt; where Δt is the time interval. A threshold for slope change is set to determine when the derivative is considered as an abrupt change point. When the absolute value of the derivative exceeds a certain set value (such as 5%), it is considered that an abrupt change has occurred, and the point is marked to generate an abrupt change point marking map for direct and intuitive display of the waveform curve and its abrupt change points.

[0098] This invention calculates the carbon emission abrupt change point and obtains the formula for calculating the gradient change rate:

[0099] Gradient change rate = (mean after mutation - mean before mutation) / Δt

[0100] If Δt is 20 minutes, then the gradient change rate is 1 ton / minute;

[0101] The duration of the carbon emission mutation is calculated. For example, if the carbon emission value remains above 90 tons after the mutation until it drops to 70 tons after 40 minutes, the duration of the mutation is 10 minutes.

[0102] In this embodiment of the invention, a carbon emission monitoring time window is located at the occurrence of a timestamp. For example, if the mutation point time is 30 minutes, the corresponding time window is 20-30 minutes. Statistical analysis is performed on all windows to calculate the frequency distribution of mutation points within each window. If multiple mutation points are found to be concentrated in the 20-30 minute and 40-50 minute windows during the monitoring period, it can be considered that the carbon emission fluctuations in these windows are relatively significant. The mutation point window distribution characteristics are obtained to show the distribution of mutation points in different time windows.

[0103] In this embodiment of the invention, the number and distribution pattern of mutation points within each monitoring window are statistically analyzed to form a feature matrix. If three mutation points appear within 20-30 minutes of the window, their number and related characteristics are recorded. The gradient change rate and mutation duration are combined with the window features to form a multi-dimensional carbon emission mutation point feature.

[0104] In this embodiment of the invention, a suitable fitting method is selected, such as piecewise linear regression or polynomial fitting, to model the extracted multi-dimensional carbon emission mutation point features. The parameters of the fitting model are set, and the feature information of each mutation point is considered. During the fitting process, the focus is on the time points with significant features. Through the fitting algorithm, the total carbon emission waveform curve is combined with the extracted multi-dimensional carbon emission mutation point features. For example, the weights of the fitting model are adjusted according to the gradient change rate of the mutation points to construct carbon emission fluctuation maps at multiple time points, showing the carbon emission change trend at different time points.

[0105] Specifically, step 2 includes:

[0106] Based on the carbon emission fluctuation map, the carbon emission trend changes in different regions were analyzed to obtain the characteristics of carbon emission trend changes in different regions.

[0107] The evolution of carbon emission trends in all regions is predicted, resulting in carbon emission trend evolution prediction curves.

[0108] Based on the carbon emission status evolution prediction curve, multi-regional carbon emission demand analysis is conducted to generate carbon emission demand data for multiple regions.

[0109] The minimum carbon emission demand for each region is calculated over multiple time periods to obtain the minimum carbon emission demand value for each region over multiple time periods.

[0110] The minimum carbon emission demand values ​​for multiple time periods in all regions are fitted with regional demand distribution to construct a multi-regional minimum carbon emission demand distribution map.

[0111] This invention decomposes the carbon emission fluctuation map by region, extracting carbon emission fluctuation data for each region. Assuming the target region includes five regions: A, B, C, D, and E, statistical analysis is performed on the carbon emission fluctuations of each region, including mean, standard deviation, maximum, and minimum values. Region A has a mean carbon emission of 70 tons, a standard deviation of 10 tons, a maximum of 100 tons, and a minimum of 50 tons. The dynamic change rate of carbon emissions in each region is calculated, i.e., the change in carbon emissions per unit time. Assuming region B has a carbon emission change rate of 5 tons / hour over a certain period, the carbon emission characteristics of each region are displayed through visualization methods, such as drawing carbon emission trend maps for each region, allowing for a direct comparison of the carbon emission trends in different regions. The characteristics of the carbon emission trend changes in each region are recorded, providing a basis for subsequent trend evolution prediction.

[0112] Assume that the historical carbon emission time series data for region A includes the total monthly carbon emissions for the past 12 months. A suitable prediction model is selected; assuming the ARIMA model is used. First, the model order (p, d, q) is determined using autocorrelation and partial autocorrelation plots. The model is then fitted to the carbon emission time series data for each region to obtain the model parameters. The parameters of the ARIMA model for region A are assumed to be (0.5, 0, 0.3). Using the fitted model, future carbon emissions are predicted, generating a carbon emission trend evolution prediction curve. The predicted carbon emissions for region A in the next three months are 75 tons, 78 tons, and 80 tons, respectively. The carbon emission trend evolution prediction curve is plotted and compared with historical data to verify the accuracy of the prediction model and ensure the reliability of the prediction.

[0113] Specifically, based on the carbon emission trend evolution prediction curve, multi-regional carbon emission demand analysis is conducted to generate carbon emission demand data for multiple regions, including:

[0114] Assume that the future carbon emission demand in region A is 1.1 times the predicted value, and in region B it is 1.2 times. Considering economic growth factors, assume that the GDP growth rate in region A is 3% and in region B it is 2%. Therefore, the carbon emission demand data is adjusted to reflect economic changes. If the predicted carbon emissions in region A are 80 tons, then the adjusted demand is 80*(1+0.03)=82.4 tons. This generates carbon emission demand data for multiple regions, such as region A with a demand of 82.4 tons, region B with a demand of 90 tons, region C with a demand of 75 tons, and so on.

[0115] Specifically, the minimum carbon emission demand for each region is calculated over multiple time periods to obtain the minimum carbon emission demand value for each region over multiple time periods, including:

[0116] The carbon emission demand data for each region is categorized by time period, such as by hour, day, or month. For example, the data is divided into daily demand data. For each region, a calculation model for the minimum carbon emission demand is established. It is assumed that a linear programming model is used to solve for the minimum demand value, with the objective function being to minimize the carbon emission demand of each region. Constraints are set, such as meeting the needs of economic growth and policy-mandated carbon emission caps. For example, if the carbon emission cap for region A is 90 tons, an optimization algorithm is used to solve for the minimum carbon emission demand, resulting in the minimum carbon emission demand values ​​for each region across multiple time periods.

[0117] Specifically, the minimum carbon emission demand values ​​for multiple time periods in all regions are fitted with regional demand distributions to construct a multi-regional minimum carbon emission demand distribution map, including:

[0118] Collect the minimum carbon emission demand values ​​for multiple time periods for all regions to form a dataset. The minimum demand value for region A is 70 tons, for region B it is 85 tons, and for region C it is 75 tons.

[0119] Choose a suitable distribution fitting model, assuming a Gaussian distribution is chosen for fitting: First, calculate the mean and standard deviation of the dataset, and use the least squares method to fit the Gaussian distribution parameters to obtain the fitted demand distribution curve, with a mean of 76 tons and a standard deviation of 7 tons; Based on the fitted demand distribution curve, draw a multi-region minimum carbon emission demand distribution map to show the carbon emission demand distribution characteristics of different regions for intuitive comparison. The map shows the demand distribution of regions A, B, and C.

[0120] Specifically, step 3 includes:

[0121] Monitor the real-time operating status parameters of each energy power generation device within the target area;

[0122] The real-time operating status parameters are calculated to obtain the energy production cycle, power generation of each energy power generation device, energy inventory and operating efficiency. The power generation, energy inventory and operating efficiency of each energy power generation device are then fitted with a time series to obtain a resource dynamic data sequence.

[0123] Based on the energy production cycle, multi-cycle output change analysis is performed on the dynamic data sequence of resources to obtain the output fluctuation characteristics of each cycle.

[0124] Short-term capacity filtering forecasts are performed on the output fluctuation characteristics of each cycle to obtain the energy capacity forecast curve.

[0125] This invention embodiment installs necessary sensors, such as power sensors, temperature sensors, and flow meters, on each energy generation device to ensure the collection of real-time operating status parameters. For wind turbines, wind speed sensors and power generation modules are installed, a data acquisition system is configured, and the data acquisition frequency is set to once per minute to update the device operating status in real time. Assuming there are 10 energy generation devices in the monitoring area, a total of 10 real-time operating status parameters will be collected every minute. This data should include operating power, device temperature, wind speed (for wind power), and light intensity (for solar power). In 24 hours, it is assumed that 14,400 real-time operating status parameters can be collected.

[0126] The real-time status parameters of each power generation device are calculated to obtain its power generation capacity. For example, if a wind turbine has a power of 150kW at a specific time, its operating efficiency is calculated based on wind speed and equipment characteristics.

[0127] Calculate energy inventory, assuming each device has an initial inventory value, which is dynamically updated based on power generation and consumption. If the initial inventory of a device is 1000kWh, and it generates 150kWh of power in 1 hour, then the new inventory is 850kWh.

[0128] The operating efficiency of the equipment is calculated using the formula: Operating efficiency = Actual power generation / Theoretical power generation × 100%. Assuming the theoretical power generation is 200kW, the efficiency is 75%.

[0129] The calculated power generation, energy inventory, and operating efficiency data are arranged in chronological order to construct a dynamic resource data sequence.

[0130] Time series analysis methods, such as linear regression or exponential smoothing, are used to fit time series data to eliminate noise and fluctuations, generating smooth resource dynamic data sequences.

[0131] Collect historical power generation data from energy generation equipment, determine the preliminary range of the production cycle, analyze power generation data from the past three months, and hypothesize that the equipment's power generation capacity differs between weekends and weekdays.

[0132] Based on the real-time operating status parameters, a reasonable calculation cycle of 24 hours is set, and the power generation within 24 hours is analyzed. In each cycle, the total power generation and average power generation of the equipment are calculated to determine the energy output cycle. If the total power generation in a 24-hour period is 3600kWh, then the average power generation is 150kW.

[0133] Based on the energy production cycle, the resource dynamic data sequence is divided into multiple cycles. Each device is divided into 24-hour cycles to form multiple cycle datasets. The data of each cycle is analyzed to calculate the total power generation, mean, standard deviation and rate of change for each cycle. The power generation of a certain device in different cycles is [1500, 1800, 1200] kWh, the mean is calculated to be 1500 kWh and the standard deviation is 300 kWh.

[0134] Extract the output fluctuation characteristics for each period, such as the maximum value, minimum value, and fluctuation amplitude within the period;

[0135] Assuming a linear regression model is chosen as the prediction model, historical data is used for model training. Using a dataset from the past 30 days, the model learns the relationship between electricity generation changes and power generation conditions. Cross-validation is used to evaluate the model, ensuring its generalization ability and accuracy. The model's root mean square error (RMSE) is assumed to be 5 kWh, indicating good predictive ability. The trained model is then used for short-term capacity forecasting, generating a prediction curve for the next three days' electricity generation. The prediction results are compared with actual operating data to verify the model's accuracy, and the model parameters are adjusted accordingly. Extracted output fluctuation features are input into the trained model for short-term capacity filtering forecasting, resulting in the energy capacity forecast curve.

[0136] Specifically, step 4 includes:

[0137] The dispatchable energy resources for multiple time windows are calculated based on real-time operating status parameters.

[0138] A time-series carbon budget allocation status analysis was performed on the carbon emission time-series data stream of the target area to obtain the carbon budget allocation status at multiple time points.

[0139] Nonlinear implicit feature analysis is performed on the carbon budget allocation status and dispatchable energy at each time point to identify the implicit correlation between the carbon budget and dispatchable energy.

[0140] Based on the correlation features, a nested correlation model of carbon emissions and energy is constructed to build a carbon emission nested energy structure correlation graph.

[0141] In this embodiment of the invention, the resource dynamic data sequence obtained based on real-time running status parameters is divided into time windows. For example, each time window is set to 1 hour, and 24 time windows are divided within 24 hours.

[0142] Within each time window, the dispatchable power generation is calculated. This involves determining the actual power generation of each device within that window based on its historical power generation capacity and real-time operating status. For example, if a wind turbine has a power generation capacity of 100kW and an actual power generation of 80kW within a window, then the dispatchable energy is 80kW. The total dispatchable energy of all power generation devices within each time window is then calculated. Assuming there are 5 devices in the region within a certain window, with dispatchable energy values ​​of [80, 70, 60, 90, 85]kW respectively, then the total dispatchable energy is 385kW.

[0143] Collect time-series carbon emission data streams from multiple regions to ensure data integrity and accuracy, assuming the data covers daily carbon emissions from five regions over the past month;

[0144] Set a total carbon budget and allocate it based on the historical emissions and future projected emissions of each region. Assuming the total carbon budget is 1,000 tons and the historical emissions of region A account for 20%, then the budget for region A is 200 tons.

[0145] A time-series analysis of carbon budget allocation in each region was conducted to assess budget changes at different points in time. For example, on a certain day, the carbon emissions of region A were 50 tons, and its budget utilization rate was 25%.

[0146] The carbon budget allocation status and dispatchable energy at multiple time points are integrated to form a feature dataset, assuming that the dataset contains the carbon budget, dispatchable energy and its corresponding carbon emissions at each time point;

[0147] Choose a suitable nonlinear analysis model, such as random forest, for training and validation. Divide the feature dataset into training and test sets to ensure the accuracy of the model. Assume that the training set contains data from the past 30 days and the test set contains data from the last 7 days. Use the trained model to identify the implicit correlation between carbon budget and dispatchable energy.

[0148] Based on the results of implicit structure analysis, the relationship between carbon budget, dispatchable energy and carbon emission intensity was identified;

[0149] A collaborative graph is constructed using graph theory, where nodes represent different variables (such as carbon budget, dispatchable energy, and carbon emissions), and edges represent the relationships between variables. Based on the analysis, the constructed graph is assumed to contain 5 nodes and 6 edges.

[0150] Network analysis is performed to calculate indicators such as centrality and connectivity of each node, generating a carbon emission nested energy structure coordination graph to show the relationship between carbon emissions and energy and their dynamic changes. The graph shows how dispatchable energy responds when the carbon budget increases.

[0151] Specifically, step 5 includes:

[0152] Based on the carbon emission fluctuation maps at multiple time points, an adaptive carbon emission upper limit constraint is defined to obtain the carbon emission constraint conditions for the target area.

[0153] Calculate the maximum energy utilization rate based on the energy production capacity forecast curve;

[0154] Based on carbon emission constraints, a spatiotemporal energy supply and demand matching model is constructed by coupling supply and demand to maximize energy utilization.

[0155] Multi-scenario simulations were conducted on the spatiotemporal energy supply and demand matching model to identify potential supply and demand conflict points.

[0156] Based on the carbon emission nested energy structure coordination diagram, a multi-objective function intelligent allocation process is performed on potential supply and demand conflict points to construct an intelligent energy allocation strategy.

[0157] This invention identifies carbon emission patterns and trends in various regions by analyzing carbon emission fluctuation maps at multiple time points. It is assumed that carbon emissions in region A reach 120 tons during peak periods and 60 tons during trough periods.

[0158] Adaptive constraints are set, and the constraints are dynamically adjusted based on the standard deviation and mean of historical data to obtain the carbon emission constraints for the target area. If the historical carbon emission mean of area A is 80 tons and the standard deviation is 15 tons, then its carbon emission limit can be set to the mean plus one standard deviation, i.e., 95 tons. For different time periods and seasons, the carbon emission limit is dynamically adjusted. In summer, due to peak electricity consumption, the carbon emission limit of area B can be set to 100 tons, while in winter it is set to 80 tons.

[0159] Analyze the energy production capacity forecast curve to determine the available energy in different time periods. Assume that the predicted energy production capacity is 200 kWh in a certain time period.

[0160] The maximum utilization rate is calculated under a given carbon emission constraint. The formula is: Utilization rate = Actual energy consumption / Available energy × 100%;

[0161] The optimization target is to maximize the actual energy consumption while meeting carbon emission constraints. If the available energy is 200 kWh and the actual consumption is 150 kWh, the utilization rate is 75%.

[0162] Using linear programming or other optimization methods, calculate the optimal energy utilization strategy based on energy availability and carbon emission constraints, and record the maximized energy utilization rate.

[0163] Define the relevant variables of the supply and demand matching model, including the supply, demand, carbon emissions and dispatchable resources of each region. Assume that the demand of region A is 300 kWh and the energy supply is 250 kWh. Set the model constraints, including carbon emission constraints and energy balance conditions. The carbon emissions of region A cannot exceed 95 tons, and the supply must meet the demand.

[0164] Using optimization algorithms, such as mixed integer linear programming (MILP), a supply and demand matching model is constructed to solve the optimal supply and demand allocation strategy in different time periods. The model solution results are recorded, and supply and demand matching schemes for each region in different time periods are generated and visualized to help understand the supply and demand relationship.

[0165] By setting different scenario variables, such as weather changes, demand fluctuations and policy adjustments, and considering the scenario of a 10% decrease in energy production capacity, the spatiotemporal supply and demand matching model is simulated to evaluate the supply and demand matching effect under different scenarios. The simulation results show that under a certain scenario, the supply in region A is insufficient, which may lead to supply and demand conflicts.

[0166] Record the supply and demand matching results under each scenario, paying special attention to the supply and demand conflict points. In a certain scenario, the energy supply of region B is 80kWh, while the demand is 100kWh, resulting in a supply and demand gap of 20kWh.

[0167] Considering potential supply and demand conflicts, a multi-objective function is constructed, with objectives including minimizing carbon emissions, maximizing energy efficiency, and meeting energy demand.

[0168] Choose a suitable optimization algorithm, such as genetic algorithm or particle swarm optimization (PSO), to calculate the intelligent allocation strategy;

[0169] The optimization objective and constraints are set in the model, and iterative calculations are performed to obtain the optimal energy allocation strategy. The optimization results may show that, within a certain time period, 80% of the energy should be allocated to region A and 20% to region B.

[0170] Specifically, step 6 includes:

[0171] Carbon emission scheduling and execution operations are carried out based on intelligent energy allocation strategies, and carbon emission feedback information is collected.

[0172] Based on carbon emission feedback information, the power grid carbon emission optimization efficiency is calculated to obtain the global carbon efficiency vector of the power grid.

[0173] The carbon emission efficiency deviation is calculated based on the preset carbon budget target vector to obtain the carbon efficiency deviation curve.

[0174] Based on the carbon efficiency deviation curve, the energy intelligent allocation strategy is iteratively optimized and subjected to global long-term evolution learning to obtain energy consumption optimization results.

[0175] In this embodiment of the invention, a specific scheduling plan is formulated based on an intelligent energy allocation strategy. Energy is prioritized for scheduling during peak periods to ensure that the power demand of each region is met. The power grid scheduling system is activated to execute the scheduling plan and monitor the power supply and carbon emissions of each region in real time. Assuming that during the scheduling process, the energy supply of region A is 300 kWh and the carbon emissions are 50 tons, sensors and data acquisition equipment are used to record the carbon emission data and power consumption of each region in real time. Assuming that 1,000 carbon emission feedback messages are collected during the entire scheduling process, including timestamps, regions, supply and emissions.

[0176] Based on carbon emission feedback information, the carbon emission efficiency of each region is calculated. Carbon emission efficiency can be defined as the amount of carbon emissions generated per unit of electricity. The calculation formula is: Carbon efficiency = Carbon emissions / Electricity supply. If the carbon emissions of region A are 50 tons and the electricity supply is 300 kWh, then the carbon efficiency of region A is: Carbon efficiency A = 50 tons / 300 kWh ≈ 0.167 tons / kWh. The carbon efficiency of each region is recorded and a global carbon efficiency vector is formed. Assume that there are 5 regions with carbon efficiencies of [0.167, 0.12, 0.15, 0.18, 0.14] tons / kWh.

[0177] Set a preset carbon budget target vector, with the target vector set as [0.15, 0.13, 0.14, 0.15, 0.12] tons / kWh. Calculate the carbon efficiency deviation using the formula: Deviation = Actual Carbon Efficiency - Target Carbon Efficiency. Calculate the deviation for each region. Assuming the actual carbon efficiency of region A is 0.167 tons / kWh and the target carbon efficiency is 0.15 tons / kWh, the deviation is: Deviation A = 0.167 - 0.15 = 0.017 tons / kWh. Record the carbon efficiency deviation for each region and calculate the carbon efficiency deviation curve.

[0178] Analyze the carbon efficiency deviation curve to identify areas with large deviations. For example, if the deviation of area A is 0.017 tons / kWh, decide to prioritize adjusting the resource allocation of that area in the next allocation.

[0179] The energy allocation strategy is iteratively optimized using optimization algorithms (such as genetic algorithms or particle swarm optimization). The objective function is set as minimizing carbon emission deviation, and the constraints are ensuring electricity demand and carbon emission constraints. Multiple iterative calculations are performed, and the optimization results of each iteration are recorded to ensure that the carbon emission deviation gradually decreases in each optimization. The optimization results are input into a deep learning model for long-term evolutionary learning to train the model to adapt to different carbon emission and energy demand scenarios. Historical data is used as the training set, and real-time feedback is used as the validation set to construct a global carbon-constrained energy consumption optimization model. The energy consumption optimization results are obtained based on the global carbon-constrained energy consumption optimization model.

[0180] This invention calculates total carbon emissions based on the time-series carbon emission data stream of the target region, and performs multi-time-point abrupt change fitting based on the total carbon emissions to construct carbon emission fluctuation maps at multiple time points for calculating minimum carbon emission demand over multiple time periods. It also performs regional demand distribution analysis to construct a multi-regional minimum carbon emission demand distribution map. Based on the real-time operating status parameters of energy generation equipment within the target region, it performs multi-cycle output change analysis and short-term capacity filtering prediction to obtain an energy capacity prediction curve. Based on the real-time operating status parameters, it performs nonlinear implicit feature analysis and modeling on the time-series carbon emission data stream of the target region to construct a carbon emission nested energy structure coordination diagram. According to carbon emission constraints, it couples the multi-regional minimum carbon emission demand distribution map and the energy capacity prediction curve for supply and demand matching to obtain potential supply and demand conflict points, and then analyzes these potential supply and demand conflict points based on the carbon emission nested energy structure coordination diagram. This invention constructs an intelligent energy allocation strategy; based on this strategy, carbon emission scheduling is performed and carbon emission feedback information is collected to obtain a carbon efficiency deviation curve. The intelligent energy allocation strategy is then iteratively optimized based on this curve to achieve energy consumption optimization. Compared to existing technologies, this invention effectively extracts information on carbon emission trend changes, abnormal disturbances, and periodic fluctuations through multi-time-point mutation fitting. By constructing a multi-regional minimum carbon emission demand distribution map and using the minimum carbon emission demand as a boundary condition, the optimization strategy is ensured to be feasible within the carbon reduction target range. Based on carbon emission constraints, the multi-regional minimum carbon emission demand distribution map and energy production capacity prediction curve are coupled for supply and demand matching. Iterative optimization of the intelligent energy allocation strategy based on the carbon efficiency deviation curve addresses the dynamic evolution and high-dimensional complexity of the system, thereby improving energy consumption optimization.

[0181] This invention also provides an energy consumption optimization system based on dynamic carbon constraints, comprising:

[0182] The mutation fitting module is used to calculate the total carbon emissions based on the carbon emission time series data stream of the target area, and to perform multi-time point mutation fitting based on the total carbon emissions to construct a carbon emission fluctuation map at multiple time points.

[0183] The demand analysis module is used to calculate the minimum carbon emission demand for multiple time periods based on the carbon emission fluctuation map, and to perform regional demand distribution analysis to construct a multi-region minimum carbon emission demand distribution map.

[0184] The capacity forecasting module is used to perform multi-cycle output change analysis and short-term capacity filtering forecast based on the real-time operating status parameters of energy power generation equipment in the target area, and to obtain the energy capacity forecasting curve.

[0185] The feature analysis module is used to perform nonlinear implicit feature analysis and modeling on the carbon emission time series data stream of the target area based on real-time operating status parameters, and to construct a carbon emission nested energy structure collaborative graph.

[0186] The allocation module is used to couple the minimum carbon emission demand distribution map and energy capacity forecast curve of multiple regions according to carbon emission constraints, obtain potential supply and demand conflict points, and allocate potential supply and demand conflict points based on the carbon emission nested energy structure coordination diagram to construct an intelligent energy allocation strategy.

[0187] The allocation optimization module is used to perform carbon emission scheduling and execution operations and collect carbon emission feedback information based on the intelligent energy allocation strategy to obtain the carbon efficiency deviation curve. Based on the carbon efficiency deviation curve, the intelligent energy allocation strategy is iteratively optimized to obtain the energy consumption optimization result.

[0188] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. An energy consumption optimization method based on dynamic carbon constraints, characterized in that, include: Step 1: Calculate the total carbon emissions based on the carbon emission time series data stream of the target area, and perform multi-time point abrupt change fitting based on the total carbon emissions to construct a carbon emission fluctuation spectrum at multiple time points; Step 2: Calculate the minimum carbon emission demand for multiple time periods based on the carbon emission fluctuation map, and conduct regional demand distribution analysis to construct a multi-regional minimum carbon emission demand distribution map. Step 3: Based on the real-time operating status parameters of the energy power generation equipment in the target area, perform multi-cycle output change analysis and short-term capacity filtering prediction to obtain the energy capacity prediction curve. Step 4: Based on the real-time operating status parameters, perform nonlinear implicit feature analysis and modeling on the carbon emission time-series data stream of the target area to construct a carbon emission nested energy structure coordination diagram, including: Based on the real-time operating status parameters, the schedulable energy for multiple time windows is calculated; A time-series carbon budget allocation status analysis was performed on the carbon emission time-series data stream of the target area to obtain the carbon budget allocation status at multiple time points; Nonlinear implicit feature analysis is performed on the carbon budget allocation status and the dispatchable energy at each time point to identify the implicit correlation features between the carbon budget and the dispatchable energy. Based on the aforementioned correlation features, a nested correlation model of carbon emissions and energy is constructed to create a carbon emission nested energy structure synergy diagram. Step 5: Based on the carbon emission constraints, couple the minimum carbon emission demand distribution map of the multi-regional areas with the energy production capacity prediction curve to obtain potential supply-demand conflict points. Then, based on the carbon emission nested energy structure coordination diagram, allocate these potential supply-demand conflict points to construct an intelligent energy allocation strategy, including: Based on the carbon emission fluctuation maps at the multiple time points, an adaptive carbon emission upper limit constraint is defined to obtain the carbon emission constraint conditions for the target region. Calculate the maximum energy utilization rate based on the energy production capacity prediction curve; Based on the carbon emission constraints, a spatiotemporal energy supply and demand matching model is constructed by coupling supply and demand for maximizing energy utilization efficiency. Multi-scenario simulations were performed on the spatiotemporal energy supply and demand matching model to obtain potential supply and demand conflict points; Based on the carbon emission nested energy structure coordination diagram, the potential supply and demand conflict points are intelligently allocated using a multi-objective function to construct an intelligent energy allocation strategy. Step 6: Based on the energy intelligent allocation strategy, perform carbon emission scheduling and carbon emission feedback information collection to obtain the carbon efficiency deviation curve, and iteratively optimize the energy intelligent allocation strategy based on the carbon efficiency deviation curve to obtain energy consumption optimization results.

2. The energy consumption optimization method based on dynamic carbon constraints according to claim 1, characterized in that, Step 1 includes: Carbon emission data of the target area is monitored based on a distributed multi-source carbon emission sensor network to obtain a carbon emission time series data stream; The carbon emission time-series data stream is dynamically decomposed into multiple carbon emission monitoring time windows. The total carbon emissions are calculated for the carbon emission time series data stream within each carbon emission monitoring time window, thus obtaining the total carbon emissions for each carbon emission monitoring time window; For each carbon emission monitoring time window, the total carbon emissions are fitted with a time-series fluctuation to generate a total carbon emission waveform curve; The total carbon emission fluctuation curve is fitted with multiple time-point abrupt changes to construct a carbon emission fluctuation map at multiple time points.

3. The energy consumption optimization method based on dynamic carbon constraints according to claim 2, characterized in that, For each carbon emission monitoring time window, the total carbon emissions are fitted with a time-series fluctuation to generate a total carbon emission waveform curve, including: The total carbon emissions for each carbon emission monitoring time window are used as input to the fitting model; The parameters of the fitted model are estimated using the least squares method, and the optimal fitting parameters of the fitted model are calculated. A waveform curve of total carbon emissions is generated based on the optimal fitting parameters.

4. The energy consumption optimization method based on dynamic carbon constraints according to claim 2, characterized in that, The total carbon emission fluctuation curve is fitted with multi-time-point abrupt changes to construct carbon emission fluctuation maps at multiple time points, including: The slope change and derivative analysis methods were used to identify and mark abrupt carbon emission change points on the total carbon emission waveform curve. The carbon emission mutation points are calculated to obtain the gradient change rate used to quantify the intensity of the mutation point change and the mutation duration used to reflect the duration of the mutation effect, and the timestamp of the occurrence of the carbon emission mutation point is extracted. Based on the timestamp, the carbon emission monitoring time window is located, and the window distribution analysis of the carbon emission monitoring time window is performed to obtain the distribution characteristics of the mutation point window; Based on the distribution characteristics of the mutation point window, the gradient change rate, and the duration of the mutation, carbon emission mutation point features are mined to obtain multi-dimensional carbon emission mutation point features. Based on the multi-dimensional carbon emission mutation point characteristics, the total carbon emission waveform curve is fitted with multi-time-point mutations to construct a carbon emission fluctuation spectrum at multiple time points.

5. The energy consumption optimization method based on dynamic carbon constraints according to claim 4, characterized in that, Step 2 includes: Based on the carbon emission fluctuation map, the carbon emission trend changes in different regions are analyzed to obtain the characteristics of carbon emission trend changes in different regions. The evolution of carbon emission trends in all regions is predicted, resulting in carbon emission trend evolution prediction curves. Based on the carbon emission status evolution prediction curve, a multi-regional carbon emission demand analysis is performed to generate carbon emission demand data for multiple regions. The minimum carbon emission demand for each region is calculated over multiple time periods to obtain the minimum carbon emission demand value for each region over multiple time periods. The minimum carbon emission demand values ​​for multiple time periods in all regions are fitted with regional demand distribution to construct a multi-regional minimum carbon emission demand distribution map.

6. The energy consumption optimization method based on dynamic carbon constraints according to claim 5, characterized in that, Step 3 includes: Monitor the real-time operating status parameters of each energy power generation device within the target area; The real-time operating status parameters are calculated to obtain the energy production cycle, the power generation of each energy power generation device, the energy inventory and the operating efficiency. The power generation of each energy power generation device, the energy inventory and the operating efficiency are then fitted with a time series sequence to obtain a resource dynamic data sequence. Based on the energy production cycle, a multi-cycle output change analysis is performed on the resource dynamic data sequence to obtain the output fluctuation characteristics of each cycle. Short-term capacity filtering forecasts are performed on the output fluctuation characteristics of each cycle to obtain the energy capacity forecast curve.

7. The energy consumption optimization method based on dynamic carbon constraints according to claim 1, characterized in that, Step 6 includes: Based on the aforementioned intelligent energy allocation strategy, carbon emission scheduling and execution operations are performed, and carbon emission feedback information is collected. Based on the carbon emission feedback information, the power grid carbon emission optimization efficiency is calculated to obtain the global carbon efficiency vector of the power grid. The carbon emission efficiency deviation is calculated based on the preset carbon budget target vector to obtain the carbon efficiency deviation curve. Based on the carbon efficiency deviation curve, the energy intelligent allocation strategy is iteratively optimized and subjected to global long-term evolution learning to obtain energy consumption optimization results.

8. An energy consumption optimization system based on dynamic carbon constraints, characterized in that, include: The mutation fitting module is used to calculate the total carbon emissions based on the carbon emission time series data stream of the target area, and to perform multi-time point mutation fitting based on the total carbon emissions to construct a carbon emission fluctuation spectrum at multiple time points. The demand analysis module is used to calculate the minimum carbon emission demand for multiple time periods based on the carbon emission fluctuation map, and to perform regional demand distribution analysis to construct a multi-regional minimum carbon emission demand distribution map. The capacity prediction module is used to perform multi-cycle output change analysis and short-term capacity filtering prediction based on the real-time operating status parameters of energy power generation equipment in the target area, and to obtain the energy capacity prediction curve. The feature analysis module is used to perform nonlinear implicit feature analysis and modeling on the carbon emission time-series data stream of the target region based on the real-time operating status parameters, and to construct a carbon emission nested energy structure collaborative graph, including: Based on the real-time operating status parameters, the schedulable energy for multiple time windows is calculated; A time-series carbon budget allocation status analysis was performed on the carbon emission time-series data stream of the target area to obtain the carbon budget allocation status at multiple time points; Nonlinear implicit feature analysis is performed on the carbon budget allocation status and the dispatchable energy at each time point to identify the implicit correlation features between the carbon budget and the dispatchable energy. Based on the aforementioned correlation features, a nested correlation model of carbon emissions and energy is constructed to create a carbon emission nested energy structure synergy diagram. The allocation module is used to couple the minimum carbon emission demand distribution map of the multi-regional area with the energy production capacity prediction curve according to carbon emission constraints, obtain potential supply and demand conflict points, and allocate the potential supply and demand conflict points based on the carbon emission nested energy structure coordination diagram to construct an intelligent energy allocation strategy, including: Based on the carbon emission fluctuation maps at the multiple time points, an adaptive carbon emission upper limit constraint is defined to obtain the carbon emission constraint conditions for the target region. Calculate the maximum energy utilization rate based on the energy production capacity prediction curve; Based on the carbon emission constraints, a spatiotemporal energy supply and demand matching model is constructed by coupling supply and demand for maximizing energy utilization efficiency. Multi-scenario simulations were performed on the spatiotemporal energy supply and demand matching model to obtain potential supply and demand conflict points; Based on the carbon emission nested energy structure coordination diagram, the potential supply and demand conflict points are intelligently allocated using a multi-objective function to construct an intelligent energy allocation strategy. The allocation optimization module is used to perform carbon emission scheduling and execution operations and collect carbon emission feedback information based on the energy intelligent allocation strategy to obtain a carbon efficiency deviation curve, and to perform iterative allocation optimization on the energy intelligent allocation strategy based on the carbon efficiency deviation curve to obtain energy consumption optimization results.

Images