New energy energy-saving dispatching and carbon control system based on artificial intelligence

Abstract

The present application relates to the technical field of energy management, in particular to a new energy energy-saving scheduling and carbon control system based on artificial intelligence, comprising an output collection module, a phase correction module, a path mapping module, a scheduling arrangement module and a carbon scheduling generation module.In the present application, the change nodes are extracted by differential analysis of the multi-source output and the load operation state, the corresponding relationship of the output behavior and the load action on the time axis is established, the consistent adjustment interval is constructed in combination with the adjustment instruction interval and the output response offset relationship, the segmented carbon path structure is formed based on the equipment operation state and the energy flow relationship, the energy type and the flow process are coupled and described, the energy flow direction and the duration characteristics are divided according to the time interval, the execution timing matching the carbon path is formed and associated with the emission coefficient and the energy flow direction, the carbon emission correlation relationship covering the whole adjustment process is constructed, the correlation degree between the adjustment behavior and the emission result is strengthened, and the refinement degree of the scheduling decision is improved.

Description

Technical Field

[0001] This invention relates to the field of energy management technology, and in particular to a new energy energy-saving scheduling and carbon control system based on artificial intelligence. Background Technology

[0002] The field of energy management technology covers core aspects such as power system operation and dispatch, load forecasting, energy production and consumption balance control, distributed energy access management, and carbon emission monitoring and accounting. It mainly involves data collection, status analysis, planning and operation adjustment on the generation, transmission and distribution, and energy consumption sides. By integrating historical energy consumption data, grid operation parameters and equipment status information, it achieves energy supply and demand matching and operation coordination. At the same time, by combining carbon emission factors and energy consumption statistics methods, it quantifies and records and accounts for the consumption process of different energy types, thereby forming a management system covering the entire process of energy production, transmission and consumption. Among them, the traditional new energy energy-saving dispatch and carbon management system refers to the dispatch and carbon emission control issues in the process of new energy power generation and consumption. It usually collects data such as the output power of wind turbines, the current and voltage of photovoltaic arrays, the charging and discharging status of energy storage devices, and the load curve data of users. Based on the load forecast results at fixed time intervals and preset dispatch rules, the power generation output plan is allocated in time periods. According to the preset carbon emission coefficient, the consumption of various types of energy is calculated and statistically analyzed. By establishing a daily plan table and a dispatch instruction table, the start-up and shutdown sequence of power generation equipment, the charging and discharging sequence of energy storage, and the load reduction sequence are arranged. At the same time, the statistical energy consumption data is multiplied by the corresponding emission coefficient to obtain the carbon emission amount, which is then summarized and recorded to complete the energy dispatch and carbon emission management process.

[0003] Existing technologies rely on periodic data processing and standardized statistical methods to record energy consumption and emissions during operation, lacking detailed characterization of dynamic behavior during operation. In multi-energy coordinated regulation scenarios, it is difficult to reflect the actual action paths of various devices at different stages, resulting in a lack of direct correlation between regulation behavior and emission results. For example, during frequent load switching or when energy storage participates in regulation, the energy flow status at each stage is not continuously expressed, which can easily lead to ambiguity in emission attribution and fragmentation of process information, thereby affecting the accuracy and traceability of scheduling strategy evaluation and carbon emission analysis. Summary of the Invention

[0004] To address the technical problems existing in the prior art, embodiments of the present invention provide a new energy energy-saving scheduling and carbon management system based on artificial intelligence.

[0005] On the one hand, it provides an artificial intelligence-based new energy energy-saving scheduling and carbon management system, which includes: The power acquisition module acquires the operating status of wind turbines, photovoltaic arrays, energy storage devices and interruptible loads, extracts change nodes by differential power change, locates the load switch change time and corresponding nodes, and obtains the power output load time series. The phase correction module, based on the output load time series, obtains the start time and duration of the adjustment command, calculates the time difference positioning offset interval with the output change node, and obtains the output matching adjustment time interval. The path mapping module, based on the output matching adjustment time interval, obtains the operating status and energy flow direction of the converter station, energy storage converter and photovoltaic power tracking unit, aligns the time interval to divide the flow direction interval, associates the energy type and duration, and obtains the carbon path distribution sequence. The scheduling and orchestration module, based on the carbon path distribution sequence, obtains the operating status of energy storage devices, interruptible loads and wind turbines, divides time periods and matches carbon path relationships to obtain a set of scheduling execution time sequences; The carbon scheduling generation module, based on the scheduling execution time sequence set, reads the energy type and energy flow direction, matches the carbon emission coefficient with the associated scheduling status, and obtains the carbon emission relationship set by corresponding the release direction time period and load adjustment status.

[0006] As a further embodiment of the present invention, the output load time series includes a set of output change nodes, load switching identifiers, time alignment indexes, and associated mapping labels; the output matching adjustment time interval includes adjustment start offset, adjustment window, interval boundary identifier, and time correction scale; the carbon path distribution sequence includes energy flow direction segments, equipment operating status segments, energy type labels, and duration weights; the scheduling execution time sequence set includes schedulable time slices, release direction identifier segments, scheduling priority labels, and execution order indexes; and the carbon emission relationship set includes a carbon emission coefficient mapping table, energy type matching, scheduling status binding pairs, and emission calculation identifiers.

[0007] As a further aspect of the present invention, the change node refers to the time point corresponding to when the rate of change of power, voltage and load state exceeds a preset threshold. The load switch change time refers to the moment when the switch state of the load equipment changes.

[0008] As a further aspect of the present invention, the continuous interval refers to the continuous time period during which the load state remains unchanged from the start to the end of the load adjustment command; The offset interval refers to the time range formed by the time difference between the scheduling instruction time and the actual load response time.

[0009] As a further aspect of the present invention, the power acquisition module includes: The status acquisition submodule acquires the output power status of the wind turbine, the operating voltage status of the photovoltaic array, and the charging and discharging status of the energy storage device. At the same time, it acquires the operating switch status of the industrial load device, aligns the timestamps according to a unified time sequence, and maps each type of status into a discrete sampling sequence to obtain a multi-source status sequence set. The differential identification submodule calculates the power change rate and voltage change rate based on the multi-source state sequence set and compares them with a preset threshold, marks the corresponding time index and extracts the change node, identifies the jump state of the load switch state sequence, and obtains the state change node index. The time-series construction submodule, based on the state change node index, synchronously maps the new energy state change nodes with the load change nodes, reorganizes the node data in ascending order of time, and writes it into a unified sequence structure to obtain the output load time series.

[0010] As a further aspect of the present invention, the phase correction module includes: Based on the output load time series, the instruction extraction submodule obtains the load adjustment instruction start time and duration interval, maps the start time to the time axis index position, divides the duration interval boundary, and obtains the instruction time interval. The offset positioning submodule identifies the corresponding time offset position based on the instruction time interval, uses a time index difference alignment method to identify the corresponding time offset position, records the continuous offset index and divides the interval range to obtain the time offset interval. Based on the time offset interval, the interval correction submodule shifts the time axis to locate the load adjustment start time along the offset direction. At the same time, it extracts the boundary of the continuous interval according to the shifted time position and updates it synchronously to obtain the output matching adjustment time interval.

[0011] As a further aspect of the present invention, the path mapping module includes: Based on the output matching adjustment time interval, the status alignment submodule obtains the operating status of the converter station, the operating status of the energy storage converter and the operating status of the photovoltaic power tracking unit, and simultaneously collects the energy flow direction identifier, aligns the time interval with the operating status time axis of the corresponding time period, and obtains the operating status alignment sequence. The flow direction division submodule identifies the input-side state direction and the output-side state direction in the same time period based on the running state alignment sequence, divides the intervals according to the direction consistency relationship, merges continuous time periods, and divides the corresponding time ranges to obtain a flow direction interval set. The sequence generation submodule, based on the set of flow direction intervals, associates the energy type identifier and duration of the time period, unfolds and arranges them in chronological order, and writes the time period data into the sequence structure to obtain the carbon path distribution sequence.

[0012] As a further aspect of the present invention, the scheduling and orchestration module includes: Based on the carbon path distribution sequence, the state association submodule obtains the operating status of the energy storage device and the operating status of the interruptible load, associates the adjustment time interval with the time mapping of the operating status of the energy storage device and the output status of the wind turbine, and divides the continuous time period according to the time index position to obtain the adjustable time period. The interval filtering submodule matches the time overlap intervals with the time intervals corresponding to the carbon path distribution sequence based on the adjustable time interval, identifies the energy flow direction as the release direction time interval, eliminates non-release direction time intervals, and obtains the release direction time interval. The scheduling sequence generation submodule arranges the time periods according to the release direction in chronological order, writes the state data corresponding to the time periods into a unified sequence structure and expands the index to obtain a set of scheduling execution time sequences.

[0013] As a further aspect of the present invention, the carbon scheduling generation module includes: The identifier matching submodule reads the energy type identifier and energy flow direction identifier corresponding to the time period based on the scheduling execution time sequence set, matches the energy type of the time period with the corresponding carbon emission coefficient segment by segment, and associates the matching result with the time index to obtain the energy carbon emission mapping sequence. The relationship construction submodule, based on the energy carbon emission mapping sequence and the carbon path distribution sequence, corresponds each time period to a segment, binds the time period scheduling status with the corresponding time period position in the carbon path distribution sequence, and obtains the scheduling path corresponding relationship sequence. The result generation submodule identifies the energy flow direction as the release direction time period based on the scheduling path correspondence sequence, binds the release direction time period with the load adjustment status, and reorganizes the correspondence structure according to the time sequence to obtain the carbon emission relationship set.

[0014] As a further aspect of the present invention, during the process of reading the energy type identifier and energy flow direction identifier corresponding to the time period, the scheduling execution time sequence set is continuously divided according to a preset time step, and the energy type identifier and energy flow direction identifier within each time step are synchronously matched. In the process of matching energy types and corresponding carbon emission coefficients for time periods segment by segment, the energy type identifiers are matched one by one according to the preset carbon emission coefficient table, and the time indexes are continuously connected between adjacent time periods when the energy type identifiers switch.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, differential analysis is performed on the multi-source output and load operation status to extract change nodes, establish the correspondence between output behavior and load action on the time axis, construct a consistent adjustment range by combining the adjustment command range and the output response offset relationship, and form a segmented carbon path structure based on the equipment operation status and energy flow relationship. The energy type and flow process are coupled and characterized, and the energy flow direction and continuous characteristics are divided according to the time interval to form an execution sequence that matches the carbon path and associate the emission coefficient with the energy flow direction. This constructs a carbon emission correlation relationship covering the entire adjustment process, strengthens the correlation between control behavior and emission results, and improves the refinement of scheduling decisions. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a system flowchart of the present invention; Figure 2 This is a system block diagram of the present invention; Figure 3 This is a flowchart of the power acquisition module in this invention; Figure 4 This is a flowchart of the phase correction module in this invention; Figure 5 This is a flowchart of the path mapping module in this invention; Figure 6 This is a flowchart of the scheduling and orchestration module in this invention; Figure 7 This is a flowchart of the carbon scheduling generation module in this invention. Detailed Implementation

[0018] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0019] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0020] This invention provides an artificial intelligence-based new energy energy-saving scheduling and carbon management system, such as... Figure 1-2 The diagram shown illustrates a new energy energy-saving scheduling and carbon management system based on artificial intelligence. This system includes: The power acquisition module acquires the output power status of the wind turbine, the operating voltage status of the photovoltaic array, and the charging and discharging status of the energy storage device. At the same time, it acquires the operating switch status of the industrial load device. Each type of status is unfolded in a unified time sequence. Differential judgment is performed on the power status changes at adjacent time points and the change nodes are extracted. At the same time, the load switch status change time is located and the corresponding change nodes are identified. The time of the new energy status change node and the load change node are corresponding to the time of the load change node to obtain the power load time series. The phase correction module, based on the output load time series, obtains the start time and duration of the load adjustment command, judges the time difference between the load adjustment start time and the execution time of the output change node, locates the offset interval, shifts the start time position within the offset interval, and simultaneously extracts the boundary of the adjusted duration interval to obtain the output matching adjustment time interval. The path mapping module, based on the output matching adjustment time interval, obtains the operating status and energy flow direction identifiers of the converter station, energy storage converter and photovoltaic power tracking unit, aligns the adjustment time interval with the operating status of the corresponding time period, analyzes the state direction of the input side and output side in the same time period and divides the flow direction interval, merges time periods with the same direction, associates the energy type identifier and duration of the time period, and unfolds them in chronological order to obtain the carbon path distribution sequence. The scheduling module, based on the carbon path distribution sequence, obtains the operating status of energy storage devices and the operating status of interruptible loads, associates the energy storage status with the wind power output status within the adjustment time interval to divide the adjustable time period, matches the adjustable time period with the corresponding time period of the carbon path distribution sequence, retains the time period marked as the energy flow direction as the release direction, and unfolds it in chronological order to obtain the scheduling execution time sequence set. The carbon scheduling generation module, based on the scheduling execution time sequence set, reads the energy type identifier and energy flow direction identifier corresponding to the time period, matches the energy type and corresponding carbon emission coefficient of the time period segment by segment, and maps the time period scheduling status to the carbon path distribution sequence segment by segment. It also maps the time period with the energy flow direction as the release direction to the load adjustment status to obtain the carbon emission relationship set.

[0021] The output load time series includes a set of output change nodes, load switching identifiers, time alignment indexes, and associated mapping labels. The output matching adjustment time interval includes adjustment start offset, adjustment window, interval boundary identifier, and time correction scale. The carbon path distribution sequence includes energy flow segment, equipment operating status segment, energy type label, and duration weight. The scheduling execution time sequence set includes schedulable time slices, release direction identifier segments, scheduling priority label, and execution order index. The carbon emission relationship set includes a carbon emission coefficient mapping table, energy type matching, scheduling status binding pairs, and emission calculation identifier.

[0022] Specifically, such as Figure 2 , 3As shown, the power acquisition module includes: The status acquisition submodule acquires the output power status of the wind turbine, the operating voltage status of the photovoltaic array, and the charging and discharging status of the energy storage device. At the same time, it acquires the operating switch status of the industrial load device, aligns the timestamps according to a unified time sequence, and maps each type of status into a discrete sampling sequence to obtain a multi-source status sequence set. First, the active power analog signal is retrieved in real time by the remote terminal unit deployed in the wind turbine generator control system. Then, the power acquisition device at the substation control layer converts the secondary side induced current waveform data collected by the voltage transformer and current transformer into digital signals to obtain the output power status of the wind turbine generator. For example, in a wind farm containing 25 wind turbines with a rated power of 2 MW, if the average real-time output of each unit is 1540.02 kW, then the total active power collected at a certain moment is 38500.5 kW. Simultaneously, the DC side electrical parameters are read through the photovoltaic inverter communication interface via the Modbus TCP protocol to obtain the operating voltage status of the photovoltaic array, with a measured value of 720.6 volts. The monitoring unit of the energy storage converter uploads the charging and discharging current vector of the battery cluster in real time to obtain the charging and discharging status of the energy storage device. The current vector direction is set, and if the measured current value is -400.0 amperes, it is determined to be in a discharging state; if it is positive, it is determined to be in a charging state. At the same time, the operating switch status of the industrial load device is collected by reading the feedback signal of the circuit breaker auxiliary contact relay of the interruptible load control cabinet in the industrial field. Its status logic is defined as 1 representing load connection and 0 representing load disconnection. The timing signal from the BeiDou Navigation Satellite System is invoked, and the above-mentioned multi-source heterogeneous data is resampled with a minimum sampling step size of 10 milliseconds. Missing sampling points are filled in by linear interpolation, and median filtering is performed to remove abnormal pulse values ​​that exceed 30% of the arithmetic mean of the first 100 sampling points of the parameter. The timestamps are aligned according to a unified time order, and discretization is performed to map each type of state into a discrete sampling sequence. For example, the wind power fluctuation curve within 1 minute is discretized into 6000 numerical points, and these points are horizontally correlated with the voltage, current, and switch status bits at the same time to obtain a multi-source state sequence set.

[0023] The differential identification submodule is based on a multi-source state sequence set. It calculates the power change rate and voltage change rate respectively and compares them with preset thresholds. It marks the corresponding time index and extracts the change node. It identifies the jump state of the load switch state sequence and obtains the state change node index. First, a first-order difference operation is performed by reading continuous sampling points from the multi-source state sequence set. The difference between the nth point and the (n-1th)th point in the power sequence is calculated to obtain the power change amplitude. For example, if the power is 38500.5 kW at 100 ms and drops to 37850.5 kW at 110 ms, the absolute value of the power change is 650.0 kW. This change value is compared with a preset power change amplitude range, the upper limit of which is set at 1.5% of the unit's rated total power, i.e., 750 kW. Simultaneously, the same subtraction operation is performed on the voltage sequence to obtain the voltage change amplitude, which is then compared with the voltage change amplitude range. The fluctuation threshold is set at 1% of the rated voltage, i.e., approximately 7.21 volts. If the power change exceeds 750 kW or the voltage change exceeds 7.21 volts, the corresponding time index is immediately marked and the change node is extracted. For the load side, the XOR gate principle is used to compare the switching state bits of adjacent time points to identify the load switching state sequence transition state. For example, when the switch state changes abruptly from 1 to 0, the XOR logic output is true, indicating that a load shedding action has occurred. By summarizing the abnormal fluctuations in renewable energy power and voltage with the switching action times of the load, and recording the time points of all trigger thresholds in chronological order, a state change node index is obtained.

[0024] The time series construction submodule is based on the state change node index, synchronously maps the new energy state change nodes with the load change nodes, reorganizes the node data in ascending order of time and writes it into a unified sequence structure to obtain the output load time series. First, all marked feature time points are retrieved from memory, and the new energy status change nodes are synchronously mapped with the load change nodes. For example, if wind power fluctuates beyond the threshold at 10:15:30.500 AM, and interruptible load performs adjustment at 10:15:30.540 AM, these two time points are correlated on the global time axis based on Coordinated Universal Time (UTC). The node data is reorganized in ascending order and written into a unified sequence structure. Between two adjacent change nodes, a zero-order hold is used for step-filling, maintaining the state value of the previous moment until the next change node appears. Specifically, real-time wind power, photovoltaic operating voltage, energy storage charging and discharging current, and load switching status are encapsulated as fields in a structured array. For example, a two-dimensional matrix containing 1000 key change points is generated, where each row represents a timestamp accurate to milliseconds, and each column represents the state physical quantity of an energy device, resulting in the output load time series.

[0025] Specifically, such as Figure 2 , 4 As shown, the phase correction module includes: The instruction extraction submodule obtains the start time and duration interval of the load adjustment instruction based on the output load time series, maps the start time to the time axis index position, divides the boundary of the duration interval, and obtains the instruction time interval. First, by retrieving the level transition characteristics of the load switch status field in the sequence, the start time and duration of the load regulation command are obtained. The first sampling point where the load status bit jumps from 1 to 0 is located as the starting trigger time of the load response command, and its start time is mapped to the time axis index position. Then, the duration of the load status remaining at 0 is detected using the sliding window method. When the load status bit jumps back to 1 or reaches the command termination time preset by the dispatch center, the boundary of the duration interval is defined. For example, if the command starts at the 500th sampling index point and ends at the 3500th index point, and the sampling period is 10 milliseconds, the effective response duration is 30 seconds. Through this time-domain feature extraction method, the actual load regulation response segment is identified, and the command time interval is obtained.

[0026] The offset positioning submodule identifies the corresponding time offset position based on the instruction time interval and uses the time index difference alignment method to identify the corresponding time offset position, records the continuous offset index and divides the interval range to obtain the time offset interval. First, by comparing the baseline time of the logical instruction issuance with the actual time of the load's physical action, the corresponding time offset position is identified using a time index difference alignment method. The timestamp of the control message issued by the dispatch automation system, for example, 10:00:00:000 milliseconds, is retrieved and subtracted from the timestamp of the load switch switching point recorded in the output load time series, which is 10:00:00:120 milliseconds. The calculated offset is 120 milliseconds. This calculation is repeated over multiple instruction cycles, recording continuous offset indices and dividing the interval range. For example, in 10 consecutive adjustment tests, the offset sequence is obtained. If the mean is distributed between 115 milliseconds and 125 milliseconds, the arithmetic mean is calculated to determine the typical system delay as 120 milliseconds, corresponding to 12 sampling index intervals, thus obtaining the time offset interval.

[0027] The interval correction submodule is based on the time offset interval, shifts the time axis, positions the load adjustment start time along the offset direction, and simultaneously extracts the continuous interval boundary based on the shifted time position and updates it synchronously to obtain the output matching adjustment time interval. First, the power output load time series is reverse-shifted, translating the time axis to position the load adjustment start time along the offset direction. For example, if the determined time offset is 120 milliseconds, or 12 sampling intervals, the entire load-side state sequence is shifted 12 index positions to the left of the time axis, synchronizing the load-side response start point with the supply-side renewable energy power fluctuation start point in the calculation dimension. Simultaneously, the continuous interval boundary is extracted based on the shifted time position and updated synchronously. During the shift, invalid filler points at the beginning of the sequence are automatically removed, and data is supplemented at the end based on the last state value, ensuring the total sequence length remains unchanged at 3000 sampling points. For example, the original interval is from index 500 to 3500, and after correction, it is updated to 488 to 3488, resulting in the power output matching adjustment time interval.

[0028] Specifically, such as Figure 2 , 5 As shown, the path mapping module includes: The status alignment submodule acquires the operating status of the converter station, the energy storage converter and the photovoltaic power tracking unit based on the output matching adjustment time interval, and collects the energy flow direction identifier. It aligns the time interval with the operating status time axis of the corresponding time period to obtain the operating status alignment sequence. First, the operating status of the converter station is acquired through the communication gateway at the plant end, including the firing angle of the converter valve and the effective value of the AC bus voltage. The operating status of the energy storage converter is also acquired, reading its current DC-side capacitor voltage, IGBT switching frequency, and real-time temperature data of the main circuit power devices. The operating status of the photovoltaic power tracking unit is acquired, recording its current duty cycle control parameters and PV curve search points. Simultaneously, by reading the power flow calculation results from the energy management system, energy flow direction indicators are collected, defining current flow from the power source to the bus as positive and current flow from the bus to the load as negative. A spatiotemporal alignment operation is performed to align the time interval with the corresponding operating status time axis. For example, within the 500th to 1500th index interval, the converter station's conversion efficiency of 98.5%, the energy storage discharge rate of 0.5C, and the photovoltaic tracking accuracy data are stored as attributes in the corresponding adjustment time segment to obtain the operating status alignment sequence.

[0029] The flow direction division submodule identifies the input-side state direction and the output-side state direction in the same time period based on the running state alignment sequence, divides the intervals according to the consistency of the directions, merges continuous time periods, and divides the corresponding time ranges to obtain a set of flow direction intervals. First, a logical judgment algorithm identifies the input-side and output-side state directions within the same time period. The input side primarily examines the power output vector direction of wind turbines and photovoltaic arrays, while the output side examines the current polarity of energy storage devices and the load switching status. Based on the consistency of direction, intervals are divided. If, within a certain time period, the wind power output vector is positive and energy storage is discharging, it is determined to be a multi-source power supply interval; if the wind power output is positive and the load switching status is 0, it is determined to be an energy substitution regulation interval. Continuous time periods with the same physical meaning are merged and divided into corresponding time ranges. For example, a continuous 120-second period where wind power fully undertakes the regulation of the load gap is merged into a single logical unit, resulting in a set of flow direction intervals.

[0030] The sequence generation submodule is based on the set of flow direction intervals, associates the energy type identifier and duration of time periods, unfolds and arranges them in chronological order, and writes the time period data into the sequence structure to obtain the carbon path distribution sequence. First, the associated device metadata for each interval is retrieved, linking the energy type identifier and duration of the time period. The energy type identifier is obtained through a preset hexadecimal hardware code, such as 0x65 for wind power, 0x66 for photovoltaic, and 0x67 for energy storage. The data is then arranged chronologically and written into a sequence structure, generating a composite time-series structure containing a start index, end index, energy code, average output value, and energy conversion loss rate. For example, if wind power supports the energy source from 0 to 30 seconds, photovoltaic supplements it from 30 to 45 seconds, and energy storage discharges it from 45 to 60 seconds, this information is seamlessly linked using millisecond-level timestamps to obtain the carbon path distribution sequence.

[0031] Specifically, such as Figure 2 , 6 As shown, the scheduling and orchestration module includes: The state association submodule obtains the operating status of energy storage devices and the operating status of interruptible loads based on the carbon path distribution sequence. It associates the adjustment time interval with the time mapping between the operating status of energy storage devices and the output status of wind turbines, and divides continuous time periods according to the time index position to obtain the adjustable time period. First, the energy storage operating status is obtained through the battery management system communication protocol, focusing on reading the battery state of charge (SOC), for example, 78.2% at this point. The grid load operating status is also obtained, reading the real-time load rate of the distribution transformer, for example, 65.4%. The wind turbine output status is obtained, reading the real-time wind speed and rotor speed measured by the ultrasonic anemometer in the nacelle. The adjustment time interval is correlated with the output operating conditions of the energy storage device and the wind turbine. By calculating the maximum discharge power margin that the energy storage device can provide based on the current SOC and the power reserve of the wind turbine based on the wind energy capture curve at each sampling point, continuous time periods are divided according to the time index position. For example, if it is identified that the load is under controlled adjustment within the 10:20 segment and the wind power output surplus is greater than the load adjustment amount, this period is marked as a green adjustment window, resulting in an adjustable time period.

[0032] The interval filtering submodule is based on an adjustable time period. It matches the time overlap interval with the time period corresponding to the carbon path distribution sequence, identifies the energy flow direction as the time period of release direction, eliminates non-release direction time periods, and obtains the release direction time periods. First, based on the adjustable time period, a logical intersection operation is performed with the corresponding time period of the carbon path distribution sequence to match overlapping time intervals. Time periods marked as release directions are identified, focusing on the effective substitution process of energy flowing from the power source to the load side. Specifically, software logic gates are used to filter out segments marked as absorption directions, such as eliminating time periods where energy storage draws power from the grid for charging, as this process represents load increase rather than emission substitution in carbon accounting. Non-release direction time periods are then eliminated, retaining the time intervals where clean energy truly replaces high-carbon thermal power. For example, from a 30-minute observation period, a cumulative effective release interval of 22.5 minutes is precisely selected to obtain the release direction time periods.

[0033] The scheduling sequence generation submodule arranges the time periods according to the release direction in chronological order, writes the status data corresponding to the time periods into a unified sequence structure and expands the index to obtain a set of scheduling execution time sequences. First, the data for executing scheduling instructions is solidified and rearranged chronologically according to the release direction and time period to ensure that the logical relationship of the data frames conforms to the measured trajectory of the power system. The physical state data such as output, voltage, and frequency corresponding to the time period are written into a unified sequence structure and indexed. Each sampling point is assigned a power contribution value and a corresponding real-time carbon emission intensity weight. For example, a three-dimensional tensor is constructed, with dimensions corresponding to the time sampling point, energy composition, and real-time carbon emission intensity weight. By integrating discrete release intervals into continuous digital records, a scheduling execution time sequence set is obtained.

[0034] Specifically, such as Figure 2 , 7As shown, the carbon scheduling generation module includes: The identifier matching submodule reads the energy type identifier and energy flow direction identifier corresponding to the time period based on the set of scheduling execution time sequence, matches the energy type of the time period with the corresponding carbon emission coefficient segment by segment, and associates the matching results with the time index to obtain the energy carbon emission mapping sequence. First, the energy type identifier corresponding to the time period is read, and emission factors for different energy types are retrieved from a pre-set life cycle assessment carbon emission database. The energy type for each time period is then matched segment by segment with its corresponding carbon emission coefficient. For example, the carbon emission intensity coefficient for wind power is set at 11 g CO2 / kWh, for photovoltaic power at 20 g CO2 / kWh, and for thermal power at a baseline emission factor of 850 g CO2 / kWh. The matching results are then associated with a time index, and the energy carbon emission mapping sequence is obtained by calculating the product of the output value at each sampling point and its corresponding energy carbon intensity coefficient.

[0035] The relationship construction submodule is based on the energy carbon emission mapping sequence and the carbon path distribution sequence. It corresponds to time periods segment by segment, binds the time period scheduling status with the corresponding time period position in the carbon path distribution sequence, and obtains the scheduling path corresponding relationship sequence. First, a deep integration of scheduling behavior and carbon emission reduction logic is implemented. Based on the energy carbon emission mapping sequence and carbon path distribution sequence, each segment is mapped to a time axis. A timestamp alignment algorithm is used to bind the scheduling status of a time period to the physical attributes in the carbon path distribution sequence. Then, correlation operations are performed to integrate load adjustments, real-time increases in clean energy generation, and emissions from thermal power avoided through substitution, all within a unified spatiotemporal scale. For example, when the load decreases by 500 kW and wind power simultaneously increases by 500 kW, the instantaneous emission reduction rate is calculated by subtracting the product of the 500 kW power increase and the thermal power baseline factor from the product of the 500 kW power increase and the wind power's own carbon emission intensity coefficient. This yields the scheduling path correspondence sequence.

[0036] The result generation submodule identifies the energy flow direction as the release direction time period based on the scheduling path correspondence sequence, binds the release direction time period with the load adjustment status, and reorganizes the correspondence structure according to the time order to obtain the carbon emission relationship set. First, identify time periods where energy flow is in the release direction, focusing on effective regulation data that contributes to the actual environment. Bind the release-direction time periods to the load regulation execution logic, and calculate the total carbon emission reduction over the entire scheduling cycle by performing definite integral summation on the instantaneous emission reductions on the time axis. For example, if 2500 kWh of thermal power is replaced by wind-solar-storage coordinated control within a 30-minute regulation cycle, the emission reduction contribution at each sampling point is accumulated to calculate the total emission reduction. Reorganize the corresponding relationship structure according to time sequence, encapsulating the final emission reduction value and the proportion of each energy source involved in regulation into a structured data package to obtain a carbon emission relationship set.

[0037] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A new energy energy-saving scheduling and carbon management system based on artificial intelligence, characterized in that, include: The power acquisition module acquires the operating status of wind turbines, photovoltaic arrays, energy storage devices and interruptible loads, extracts change nodes by differential power change, locates the load switch change time and corresponding nodes, and obtains the power output load time series. The phase correction module, based on the output load time series, obtains the start time and duration of the adjustment command, calculates the time difference positioning offset interval with the output change node, and obtains the output matching adjustment time interval. The path mapping module, based on the output matching adjustment time interval, obtains the operating status and energy flow direction of the converter station, energy storage converter and photovoltaic power tracking unit, aligns the time interval to divide the flow direction interval, associates the energy type and duration, and obtains the carbon path distribution sequence. The scheduling and orchestration module, based on the carbon path distribution sequence, obtains the operating status of energy storage devices, interruptible loads and wind turbines, divides time periods and matches carbon path relationships to obtain a set of scheduling execution time sequences; The carbon scheduling generation module, based on the scheduling execution time sequence set, reads the energy type and energy flow direction, matches the carbon emission coefficient with the associated scheduling status, and obtains the carbon emission relationship set by corresponding release direction time period and load adjustment status. The phase correction module includes: Based on the output load time series, the instruction extraction submodule obtains the load adjustment instruction start time and duration interval, maps the start time to the time axis index position, divides the duration interval boundary, and obtains the instruction time interval. The offset positioning submodule identifies the corresponding time offset position based on the instruction time interval, uses a time index difference alignment method to identify the corresponding time offset position, records the continuous offset index and divides the interval range to obtain the time offset interval. Based on the time offset interval, the interval correction submodule shifts the time axis to locate the load adjustment start time along the offset direction. At the same time, it extracts the boundary of the continuous interval according to the shifted time position and updates it synchronously to obtain the output matching adjustment time interval. The path mapping module includes: Based on the output matching adjustment time interval, the status alignment submodule obtains the operating status of the converter station, the operating status of the energy storage converter and the operating status of the photovoltaic power tracking unit, and simultaneously collects the energy flow direction identifier, aligns the time interval with the operating status time axis of the corresponding time period, and obtains the operating status alignment sequence. The flow direction division submodule identifies the input-side state direction and the output-side state direction in the same time period based on the running state alignment sequence, divides the intervals according to the direction consistency relationship, merges continuous time periods, and divides the corresponding time ranges to obtain a flow direction interval set. The sequence generation submodule, based on the set of flow direction intervals, associates the energy type identifier and duration of the time period, unfolds and arranges them in chronological order, and writes the time period data into the sequence structure to obtain the carbon path distribution sequence.

2. The artificial intelligence-based new energy energy-saving scheduling and carbon management system according to claim 1, characterized in that, The output load time series includes a set of output change nodes, load switching identifiers, time alignment indexes, and associated mapping labels. The output matching adjustment time interval includes adjustment start offset, adjustment window, interval boundary identifier, and time correction scale. The carbon path distribution sequence includes energy flow segment, equipment operating status segment, energy type label, and duration weight. The scheduling execution time sequence set includes schedulable time slices, release direction identifier segments, scheduling priority labels, and execution order index. The carbon emission relationship set includes a carbon emission coefficient mapping table, energy type matching, scheduling status binding pairs, and emission calculation identifiers.

3. The artificial intelligence-based new energy energy-saving scheduling and carbon management system according to claim 1, characterized in that, The change node refers to the time point when the rate of change of power, voltage and load status exceeds a preset threshold. The load switch change time refers to the moment when the switch state of the load equipment changes.

4. The artificial intelligence-based new energy energy-saving scheduling and carbon management system according to claim 1, characterized in that, The continuous interval refers to the continuous time period during which the load status remains unchanged from the start to the end of the load adjustment command; The offset interval refers to the time range formed by the time difference between the scheduling instruction time and the actual load response time.

5. The artificial intelligence-based new energy energy-saving scheduling and carbon management system according to claim 1, characterized in that, The output acquisition module includes: The status acquisition submodule acquires the output power status of the wind turbine, the operating voltage status of the photovoltaic array, and the charging and discharging status of the energy storage device. At the same time, it acquires the operating switch status of the industrial load device, aligns the timestamps according to a unified time sequence, and maps each type of status into a discrete sampling sequence to obtain a multi-source status sequence set. The differential identification submodule calculates the power change rate and voltage change rate based on the multi-source state sequence set and compares them with a preset threshold, marks the corresponding time index and extracts the change node, identifies the jump state of the load switch state sequence, and obtains the state change node index. The time-series construction submodule, based on the state change node index, synchronously maps the new energy state change nodes with the load change nodes, reorganizes the node data in ascending order of time, and writes it into a unified sequence structure to obtain the output load time series.

6. The artificial intelligence-based new energy energy-saving scheduling and carbon management system according to claim 1, characterized in that, The scheduling and orchestration module includes: Based on the carbon path distribution sequence, the state association submodule obtains the operating status of the energy storage device and the operating status of the interruptible load, associates the adjustment time interval with the time mapping of the operating status of the energy storage device and the output status of the wind turbine, and divides the continuous time period according to the time index position to obtain the adjustable time period. The interval filtering submodule matches the time overlap intervals with the time intervals corresponding to the carbon path distribution sequence based on the adjustable time interval, identifies the energy flow direction as the release direction time interval, eliminates non-release direction time intervals, and obtains the release direction time interval. The scheduling sequence generation submodule arranges the time periods according to the release direction in chronological order, writes the state data corresponding to the time periods into a unified sequence structure and expands the index to obtain a set of scheduling execution time sequences.

7. The artificial intelligence-based new energy energy-saving scheduling and carbon management system according to claim 1, characterized in that, The carbon scheduling generation module includes: The identifier matching submodule reads the energy type identifier and energy flow direction identifier corresponding to the time period based on the scheduling execution time sequence set, matches the energy type of the time period with the corresponding carbon emission coefficient segment by segment, and associates the matching result with the time index to obtain the energy carbon emission mapping sequence. The relationship construction submodule, based on the energy carbon emission mapping sequence and the carbon path distribution sequence, corresponds each time period to a segment, binds the time period scheduling status with the corresponding time period position in the carbon path distribution sequence, and obtains the scheduling path corresponding relationship sequence. The result generation submodule identifies the energy flow direction as the release direction time period based on the scheduling path correspondence sequence, binds the release direction time period with the load adjustment status, and reorganizes the correspondence structure according to the time sequence to obtain the carbon emission relationship set.

8. The artificial intelligence-based new energy energy-saving scheduling and carbon management system according to claim 7, characterized in that: During the process of reading the energy type identifier and energy flow direction identifier corresponding to the time period, the scheduling execution time sequence set is continuously divided according to the preset time step, and the energy type identifier and energy flow direction identifier in each time step are synchronously matched. In the process of matching energy types and corresponding carbon emission coefficients for time periods segment by segment, the energy type identifiers are matched one by one according to the preset carbon emission coefficient table, and the time indexes are continuously connected between adjacent time periods when the energy type identifiers switch.

Images