Power system production simulation method based on time sequence feature migration and double-layer hierarchical

Abstract

The application discloses a power system production simulation method based on time sequence feature migration and double-layer hierarchical, and belongs to the technical field of power production simulation. The method comprises the following steps: extracting features of meteorological data and power system output data, and constructing a meteorological output time sequence feature library; generating a planning level year meteorological sequence according to the meteorological output time sequence feature library, and migrating the meteorological output time sequence feature library to a planning year based on a pre-constructed nonlinear mapping model with an introduced migration correction coefficient to obtain a wind-solar-thermal long time sequence output curve; obtaining power balance results and power balance results by solving a pre-constructed double-layer hierarchical production simulation model through double-layer bidirectional iteration according to the wind-solar-thermal long time sequence output curve; and performing multi-dimensional evaluation on the power balance results and the power balance results to obtain power system production simulation results. The application has high time sequence fidelity, rigorous balance logic and excellent solution efficiency, and is suitable for the complex characteristics and planning requirements of a power system.

Description

Technical Field

[0001] This invention relates to the field of power production simulation technology, and in particular to a power system production simulation method based on time-series feature migration and two-level hierarchical structure. Background Technology

[0002] The construction of new energy power systems, primarily based on wind, solar, and solar thermal power, is accelerating. New energy output exhibits strong randomness, volatility, intermittency, and temporal coupling. Coupled with deep coupling between power generation, grid, load, and storage, the power system evolves into a complex, nonlinear, highly stochastic, and rapidly time-varying mega-system. Power system planning and production simulation must simultaneously address multiple optimization objectives, including long-term power sufficiency, short-term power balance, new energy absorption, grid security, and reliability assurance. Output, load, energy storage, and power flow constraints are interconnected across different time periods, making it difficult for traditional methods to accurately characterize long-term dynamic characteristics.

[0003] Traditional power system production simulation methods mainly suffer from two types of defects: one is the output construction method based on typical days / typical weeks, which cannot retain the correlation between meteorological conditions and output time series, extreme ramp-up and continuous extreme value characteristics, has poor adaptability to high-proportion renewable energy scenarios, and has significant errors in absorption assessment and adequacy analysis; the other is the single-layer balance production simulation method, which calculates power balance and electricity balance independently, separates long-term total volume from short-term scheduling, has a large solution scale and slow convergence speed, and cannot balance accuracy and efficiency.

[0004] Furthermore, existing methods lack mechanisms for transferring meteorological and power output characteristics and a two-tiered hierarchical solution architecture, making it difficult to meet the requirements of long-term, highly reliable, and fast-solving production simulations at the planning level. Therefore, the power grid urgently needs a power system production simulation method that can balance time-series accuracy, coordination, and high-efficiency solution, while also meeting the complex characteristics and planning requirements under the "dual-carbon" objective of the new power system. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a power system production simulation method based on time-series feature migration and two-level hierarchical structure. This method has high time-series fidelity, rigorous balance logic, and excellent solution efficiency, which is in line with the complex characteristics and planning requirements of new power systems.

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

[0007] This invention provides a power system production simulation method based on time-series feature migration and a two-level hierarchical structure, comprising:

[0008] Acquire meteorological data and power system output data;

[0009] Extract features from the meteorological data and power system output data to construct a meteorological output time series feature library;

[0010] Based on the meteorological output time series feature library, a meteorological sequence for the planning level year is generated. Based on the meteorological sequence for the planning level year, the meteorological output time series feature library is transferred to the planning year using a pre-constructed nonlinear mapping model that introduces a migration correction coefficient, resulting in long-term wind, solar and thermal output curves.

[0011] Based on the long-term power output curves of wind, solar and thermal power, the pre-constructed two-layer hierarchical production simulation model is solved by two-layer bidirectional iteration to obtain the power balance result and the electricity balance result; the two-layer hierarchical production simulation model is constructed with power balance as the guide and electricity balance as the main factor.

[0012] The power balance results and the power balance results are evaluated from multiple dimensions to obtain the power system production simulation results.

[0013] Optionally, the meteorological data includes wind speed, irradiance, ambient temperature, cloud cover, and air pressure;

[0014] The power system output data includes wind power output data, photovoltaic power output data, and solar thermal power output data.

[0015] Optionally, the meteorological output time series feature database includes time series autocorrelation coefficient, climbing rate, extreme value duration and weather transition probability;

[0016] The time-series autocorrelation coefficient is expressed as:

[0017] ;

[0018] The climbing rate is expressed as:

[0019] ;

[0020] The duration of the extreme value is expressed as follows:

[0021] ;

[0022] The weather shift probability is expressed as:

[0023] ;

[0024] in, Indicates a delay of The time series autocorrelation coefficient; Indicates the total time; They represent time, Weather / output sequence values ​​at any given time; Represents the mean of the sequence; express Constantly adjust the climbing rate; They represent time, New energy sources are always contributing their power; Indicates the time step; Indicates the number of hours the extreme output lasts; These represent the low output threshold and the high output threshold, respectively. Indicates an indicator function; They represent time, Real-time weather conditions; Indicates weather conditions Total number of occurrences; Indicates weather conditions arrive The number of transfers; This indicates the probability of a weather state transition.

[0025] Optionally, the planned horizontal annual meteorological sequence is represented as follows:

[0026] ;

[0027] in, They represent time, Time-based planning of horizontal annual meteorological sequences; This represents the weather transfer matrix.

[0028] Optionally, the nonlinear mapping model that introduces the migration correction coefficient includes a wind power output time series characteristic migration model, a photovoltaic power output model, and a solar thermal system model;

[0029] The wind power output time-series feature transfer model is expressed as follows:

[0030] ;

[0031] ;

[0032] The photovoltaic power output model is expressed as follows:

[0033] ;

[0034] The photothermal system model is represented as follows:

[0035] ;

[0036] ;

[0037] in, express Wind power output at all times; Indicates the rated power of the wind turbine unit; express Wind speed at all times; Indicates the cut-in wind speed; Indicates the rated wind speed; express Temporal feature migration correction coefficient; This represents the autocorrelation coefficient of wind power; express Wind power ramp rate at any time; Indicates the historical average rate of ascent; All represent weighting coefficients; express Photovoltaic power output at all times; Indicates the rated power of the photovoltaic system; express irradiance at any given time; Indicates standard test irradiance; express Ambient temperature at all times; Indicates the standard test environment temperature; Indicates the power temperature coefficient; express Real-time heat collection power; Indicates the area of ​​the heat collection field; Indicates heat collection efficiency; They represent time, Real-time thermal storage capacity; Indicates the time step; express Real-time solar thermal power generation capacity; This indicates the efficiency of concentrated solar power generation.

[0038] Optionally, the two-layer hierarchical production simulation model includes an upper-layer power balance objective function, a lower-layer power balance objective function, monthly power balance constraints, hourly power balance constraints, power generation equipment capacity constraints, unit ramp-up constraints, grid security constraints, and spinning reserve constraints.

[0039] Optionally, the objective function for upper-level power balance is expressed as:

[0040] ;

[0041] The objective function for lower-level power balance is expressed as follows:

[0042] ;

[0043] in, These represent the upper-level power balance objective function and the lower-level power balance objective function, respectively. This indicates taking the minimum value; They represent Monthly power generation cost, grid loss cost, curtailment cost, and carbon emission cost; They represent Costs include fuel costs, operation and maintenance costs, unit start-up and shutdown costs, and network loss costs.

[0044] Optionally, the monthly power balance constraint is expressed as:

[0045] ;

[0046] The hourly power balance constraint is expressed as follows:

[0047] ;

[0048] The capacity constraint of the power generation equipment is expressed as follows:

[0049] ;

[0050] The unit ramping constraint is expressed as follows:

[0051] ;

[0052] The power grid security constraints are expressed as follows:

[0053] ;

[0054] ;

[0055] The rotational spare constraint is expressed as follows:

[0056] ;

[0057] in, They represent time, Regular units contribution; They represent Real-time wind power output, photovoltaic power output, solar thermal power generation, inter-provincial power exchange capacity, and grid loss capacity; Indicates the time step; express Constant load; They represent Monthly network loss electricity, inter-provincial exchange electricity; They represent the generating units. Minimum output, maximum output; Indicates the unit Maximum climbing speed; express Time grid node Voltage amplitude; Representing power grid nodes Lower voltage limit; express Timetable Transmission power; Indicates the line Thermal stability limit; express Constantly rotate reserve capacity; This represents the reserve factor.

[0058] Optionally, based on the long-term power output curves of wind, solar, and thermal power, a pre-constructed two-layer hierarchical production simulation model is solved through a two-layer bidirectional iterative process to obtain the power balance results and the electricity balance results, including:

[0059] Obtain the current monthly power consumption plan and the new information for the next month; the current monthly power consumption plan includes initial output, load, new energy forecast, and exchange plan, and the new information for the next month includes load, new energy, and unit status;

[0060] Based on the current monthly electricity consumption plan and the new information for the next month, a search and adjustment is performed in the neighborhood of the current monthly electricity consumption plan to obtain the current adjusted monthly plan. A greedy strategy is then used to reorder the current adjusted monthly plan to generate a new monthly electricity consumption solution. The greedy strategy is formulated based on the monthly scale and sorted in ascending order according to the unit electricity cost.

[0061] Based on the new month's electricity solution, calculate the objective function value of the upper-level electricity balance until the preset termination condition is met, and obtain the upper-level electricity balance solution;

[0062] Using the upper-level power balance solution as a constraint, the hourly dispatching scheme is encoded into a subpopulation according to the number of power grid nodes or lines;

[0063] Calculate the fitness value of each individual in each subpopulation, expressed as:

[0064] ;

[0065] in, Indicates the individual's fitness value; This represents the objective function for lower-level power balance. Indicates the penalty coefficient; express Constantly constrain the amount of violations;

[0066] Tournament selection, adaptive crossover, and Gaussian mutation are performed sequentially on each subpopulation. Excellent individuals are migrated between subpopulations every fixed number of generations until the preset maximum number of iterations or the preset fitness value is reached. The subpopulation with the highest fitness is used as the objective function value of the lower-level power balance to obtain the lower-level power balance solution.

[0067] The lower-level power balance solution is fed back to the upper level to correct the upper-level power balance solution until the change in the upper-level power balance objective function is less than the first threshold and the change in the lower-level power balance objective function is less than the second threshold, thus completing the convergence and obtaining the power balance result and the power balance result.

[0068] Optionally, the power balance results and the electricity balance results are evaluated from multiple dimensions to obtain power system production simulation results, including:

[0069] By using a weighted average to fuse the energy balance results and the power balance results, the power system production simulation results are obtained, as follows:

[0070] ;

[0071] in, express Real-time power system production simulation results; They represent Real-time power balance results; All of these represent weighting coefficients.

[0072] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0073] This invention integrates time-series feature transfer and two-level hierarchical optimization. The former achieves high-fidelity output generation, while the latter enables coordinated optimization of long-cycle and short-term periods. Together, they break through the triple bottlenecks of time-series fidelity, balance accuracy, and solution speed, meeting the needs of power system planning and simulation. It overcomes the bottlenecks of traditional output construction time-series distortion and power balance disconnection, taking into account both long-cycle sufficiency and short-term security, significantly improving solution accuracy and computational efficiency. It aligns with the high proportion of renewable energy consumption and safe and stable operation requirements under the "dual-carbon" goal of the new power system, providing reliable support for long-term planning, operation verification, and consumption analysis in the power system. In addition, it provides precise adaptation from the source, grid, load, and storage sides, offering strong support for long-term planning, renewable energy consumption, and safe and stable operation in the power system. Attached Figure Description

[0074] Figure 1 This is a flowchart illustrating the power system production simulation method based on time-series feature migration and two-level hierarchical structure provided in an embodiment of the present invention. Detailed Implementation

[0075] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations thereof. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0076] The term "and / or" simply describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. Additionally, the character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0077] Example 1

[0078] This embodiment introduces a power system production simulation method based on time-series feature migration and a two-level hierarchical structure, including:

[0079] Acquire meteorological data and power system output data;

[0080] Extract features from the meteorological data and power system output data to construct a meteorological output time series feature library;

[0081] Based on the meteorological output time series feature library, a meteorological sequence for the planning level year is generated. Based on the meteorological sequence for the planning level year, the meteorological output time series feature library is transferred to the planning year using a pre-constructed nonlinear mapping model that introduces a migration correction coefficient, resulting in long-term wind, solar and thermal output curves.

[0082] Based on the long-term power output curves of wind, solar and thermal power, the pre-constructed two-layer hierarchical production simulation model is solved by two-layer bidirectional iteration to obtain the power balance result and the electricity balance result; the two-layer hierarchical production simulation model is constructed with power balance as the guide and electricity balance as the main factor.

[0083] The power balance results and the power balance results are evaluated from multiple dimensions to obtain the power system production simulation results.

[0084] This embodiment integrates time-series feature transfer and two-level hierarchical optimization. The former achieves high-fidelity output generation, while the latter enables collaborative optimization of long-cycle and short-term periods. Together, they break through the triple bottlenecks of time-series fidelity, balance accuracy, and solution speed, meeting the needs of power system planning and simulation. It overcomes the bottlenecks of traditional output construction time-series distortion and power balance disconnection, taking into account both long-cycle sufficiency and short-term security, significantly improving solution accuracy and computational efficiency. It aligns with the requirements of high-proportion renewable energy consumption and safe and stable operation of new power systems, providing reliable support for long-term planning, operation verification, and consumption analysis in power systems. In addition, it provides precise adaptation from the source, grid, load, and storage sides, offering strong support for long-term planning, renewable energy consumption, and safe and stable operation in power systems.

[0085] Example 2

[0086] like Figure 1 As shown in the figure, this embodiment introduces a power system production simulation method based on time-series feature migration and two-level hierarchical structure, including the following steps:

[0087] Step 1: Extract features and construct a meteorological output time-series feature library, specifically:

[0088] We comprehensively analyzed historical hourly meteorological data to obtain meteorological data and power system output data. The meteorological data included wind speed, irradiance, ambient temperature, cloud cover, and air pressure; the power system output data included wind power output data, photovoltaic power output data, and solar thermal power output data.

[0089] Features of meteorological data and power system output data are extracted to construct a meteorological output time series feature library. At the same time, data such as station parameters, unit characteristics, and terrain boundaries are collected to provide support for the construction of output curves.

[0090] The meteorological output time series feature database includes time series autocorrelation coefficient, climbing rate, duration of extreme values, and weather transition probability.

[0091] The time series autocorrelation coefficient is expressed as:

[0092] ;

[0093] in, Indicates a delay of The time series autocorrelation coefficient; Indicates the total time; They represent time, Weather / output sequence values ​​at any given time; This represents the mean of the sequence.

[0094] The rate of ascent is expressed as:

[0095] ;

[0096] in, express Constantly adjust the climbing rate; They represent time, New energy sources are always contributing their power; Indicates the time step.

[0097] The duration of extreme values ​​is expressed as:

[0098] ;

[0099] in, Indicates the number of hours the extreme output lasts; These represent the low output threshold and the high output threshold, respectively. This indicates an indicator function, which is 1 if the condition is met and 0 if the condition is not met.

[0100] The probability of weather shift is expressed as:

[0101] ;

[0102] in, They represent time, Real-time weather conditions; Indicates weather conditions Total number of occurrences; Indicates weather conditions arrive The number of transfers; This indicates the probability of a weather state transition.

[0103] The meteorological power output time series feature database is the core basis for generating long-term power output curves for wind, solar and thermal power.

[0104] Based on the weather transition probabilities in the meteorological output time series feature library, a Markov chain combined with time-preserving Monte Carlo is used to generate a planning-level annual meteorological sequence that is consistent with historical time series features.

[0105] Based on the fingerprint database, a nonlinear mapping model is established, and a migration correction coefficient is introduced to fully transfer the fluctuations, slopes, and extreme values ​​of historical wind, solar and thermal energy to the planning year.

[0106] The final output curves for wind power, photovoltaic power, and solar thermal power are generated over a long period of 8760 hours per year, solving the problem of time-series distortion in traditional methods.

[0107] The high-fidelity output curve is input into the two-layer hierarchical production simulation model as the core boundary condition to ensure simulation accuracy.

[0108] Step 2: Based on the meteorological output time series feature database, generate the meteorological sequence for the planned horizontal year, specifically as follows:

[0109] Based on the weather transition probability in the meteorological output time series feature database, a time series preservation method combining Markov chains and Monte Carlo is adopted to generate the planned level annual meteorological sequence, ensuring that the seasonal characteristics, intra-day periodicity, and continuous daily correlation are consistent with the history, and retaining key scenarios such as extreme light winds, continuous cloudy and rainy weather, strong radiation, and strong winds, thus avoiding time series distortion by traditional methods.

[0110] The meteorological sequence for the planned horizontal year is represented as follows:

[0111] ;

[0112] in, They represent time, Time-based planning of horizontal annual meteorological sequences; This represents the weather transfer matrix.

[0113] To generate 8,760 hours of meteorological data for the entire planning year, the weather for the next year must be simulated before wind power generation, photovoltaic power generation, and solar thermal power generation can be calculated.

[0114] Step 3: Transfer the meteorological output time series feature database to the planning year to generate long-term wind, solar, and thermal output curves, specifically:

[0115] A nonlinear mapping model of wind power, photovoltaic, solar thermal and meteorological factors is established. A time-series feature migration correction coefficient is introduced to transfer the fluctuation, ramp-up and continuous extreme value characteristics of historical output to the planning year as a whole. The output curve of 8760h throughout the year is generated by mapping time period by time. The output sets of multiple scenarios such as benchmark, extreme and high absorption are constructed to meet the planning simulation needs of multiple scenarios.

[0116] Based on the meteorological sequence of the planning year, and using a pre-constructed nonlinear mapping model with a migration correction coefficient, the meteorological output time series feature library is migrated to the planning year to obtain the long-term output curves of wind, solar and thermal power.

[0117] Nonlinear mapping models that incorporate migration correction coefficients include wind power output time-series characteristic migration models, photovoltaic power output models, and solar thermal system models.

[0118] The wind power output time-series feature transfer model is represented as follows:

[0119] ;

[0120] ;

[0121] in, express Wind power output at all times; Indicates the rated power of the wind turbine unit; express Wind speed at all times; Indicates the cut-in wind speed; Indicates the rated wind speed; express Temporal feature migration correction coefficient; This represents the autocorrelation coefficient of wind power; express Wind power ramp rate at any time; Indicates the historical average rate of ascent; All of these represent weighting coefficients.

[0122] The photovoltaic power output model is expressed as follows:

[0123] ;

[0124] in, express Photovoltaic power output at all times; Indicates the rated power of the photovoltaic system; express irradiance at any given time; Indicates standard test irradiance; express Ambient temperature at all times; Indicates the standard test environment temperature; This represents the power temperature coefficient.

[0125] The photothermal system model is represented as follows:

[0126] ;

[0127] ;

[0128] in, express Real-time heat collection power; Indicates the area of ​​the heat collection field; Indicates heat collection efficiency; They represent time, Real-time thermal storage capacity; express Real-time solar thermal power generation capacity; This indicates the efficiency of concentrated solar power generation.

[0129] Step 4: Construct a two-tiered hierarchical production simulation model, prioritizing power balance and guided by overall power balance principles. Specifically:

[0130] A two-tiered hierarchical production simulation model is constructed, with power balance as the leading factor and power balance as the dominant factor, and the objective function and constraints are clearly defined.

[0131] The pilot layer is the annual / monthly power balancing layer, used to anchor the total power volume over a long period, the target for renewable energy consumption, and the power exchange between provinces, ensuring the long-term sufficiency of the system; the main layer is the hourly power balancing layer, which optimizes unit start-up and shutdown, output allocation, energy storage scheduling, and grid power flow under the constraint of total power volume, ensuring short-term peak safety and equipment constraints.

[0132] Determine the optimization objectives: Based on the power system planning requirements, select a two-tiered hierarchical optimization objective. The pilot layer focuses on minimizing the annual / monthly electricity cost and maximizing the absorption of new energy sources, and can construct an objective function through weighted averaging. The dominant layer focuses on minimizing the hourly generation cost, grid loss, abandoned electricity, and carbon emissions, and can also construct an objective function through weighted averaging. This forms the two-tiered objective function.

[0133] The two-level hierarchical production simulation model includes an upper-level power balance objective function, a lower-level power balance objective function, monthly power balance constraints, hourly power balance constraints, power generation equipment capacity constraints, unit ramp-up constraints, grid security constraints, and spinning reserve constraints.

[0134] The objective function for upper-level power balance is expressed as:

[0135] ;

[0136] in, This represents the objective function for balancing the power supply at the upper level. This indicates taking the minimum value; They represent Monthly power generation cost, grid loss cost, curtailment cost, and carbon emission cost.

[0137] The objective function for lower-level power balance is expressed as:

[0138] ;

[0139] in, This represents the objective function for lower-level power balance. They represent Costs include fuel costs, operation and maintenance costs, unit start-up and shutdown costs, and network loss costs.

[0140] Clearly define the constraints, including physical constraints, operational constraints, and balance constraints:

[0141] The monthly electricity balance constraint is expressed as:

[0142] ;

[0143] in, express Regular units contribution; Indicates the time step, with a value of 1 hour; express Constant load; They represent Monthly network power loss and inter-provincial power exchange.

[0144] The hourly power balance constraint is expressed as:

[0145] ;

[0146] in, They represent Constantly exchange power and network loss power across provinces.

[0147] The capacity constraint of power generation equipment is expressed as:

[0148] ;

[0149] in, They represent the generating units. Minimum output, maximum output.

[0150] The unit ramp-up constraint is expressed as:

[0151] ;

[0152] in, express Regular units contribution; Indicates the unit Maximum climbing speed.

[0153] Power grid security constraints are expressed as follows:

[0154] ;

[0155] ;

[0156] in, express Time grid node Voltage amplitude; Representing power grid nodes Lower voltage limit; express Timetable Transmission power; Indicates the line Thermal stability limit.

[0157] The spin-off reserve constraint is expressed as:

[0158] ;

[0159] in, express Constantly rotate reserve capacity; This represents the reserve factor, with a value ranging from 0.05 to 0.10.

[0160] Step 5: Solve the two-level hierarchical production simulation model through two-level bidirectional iteration, specifically as follows:

[0161] The model must know about wind power, photovoltaic power, and solar thermal power in order to arrange thermal power, energy storage, and grid dispatch.

[0162] Therefore, based on the long-term output curves of wind, solar and thermal power, as well as load data, unit parameters and grid constraints, the pre-constructed two-level hierarchical production simulation model is solved through two-level bidirectional iterative solution to obtain the power balance results and the electricity balance results.

[0163] A hybrid algorithm is employed, consisting of an upper-level temporal greedy anchoring algorithm and a lower-level parallel adaptive genetic algorithm, to perform iterative solutions for power balance and grid balance, respectively. The upper-level power balance algorithm first calculates the total monthly power generation to ensure long-term power availability, while the lower-level grid balance algorithm performs hourly scheduling, power allocation, and grid security calculations to ensure short-term operational stability.

[0164] The upper-level temporal greedy anchoring algorithm is based on the idea of ​​convergence from local optima to global optima. Using a monthly scale, it formulates a greedy strategy in ascending order of unit electricity cost to quickly satisfy electricity balance and absorption constraints, generating long-cycle electricity allocation schemes. This provides boundary conditions for lower-level optimization and has the advantages of low computational cost, fast convergence, and high feasibility.

[0165] Obtain the current monthly power consumption plan and new information for the next month; the current monthly power consumption plan includes initial output, load, renewable energy forecast, and power exchange plan, and the new information for the next month includes load, renewable energy, and unit status;

[0166] Based on the current monthly electricity consumption plan and the new information for the next month, a search and adjustment is performed in the neighborhood of the current monthly electricity consumption plan to obtain the current adjusted monthly plan. A greedy strategy is used to reorder the current adjusted monthly plan to generate a new monthly electricity consumption solution. The greedy strategy is formulated on a monthly scale and sorted in ascending order according to the unit electricity cost, which clarifies the electricity balance decision objective and determines the unit electricity allocation variables.

[0167] Based on the monthly electricity output solution, the objective function value of the upper-level electricity balance is calculated. The constraints of electricity balance, absorption, and exchange are checked until the preset termination condition is met, yielding the upper-level electricity balance solution. The upper-level electricity balance solution refers to the monthly power generation per unit, the amount of renewable energy absorbed, and the total monthly cost.

[0168] The greedy strategy sorts all generator units in order of cost per kilowatt-hour from low to high, prioritizing the lowest-cost units to generate more electricity to fill the monthly load and renewable energy consumption targets, until the monthly power balance is met, at which point it stops.

[0169] The lower-level parallel adaptive genetic algorithm uses the upper-level power balance solution as a constraint, encodes the hourly scheduling scheme into individuals, divides it into multiple subpopulations for parallel computation, and improves the optimization efficiency through adaptive selection, crossover, and mutation. Excellent individuals are periodically exchanged between subpopulations to avoid getting trapped in local optima, ultimately outputting the globally optimal hourly power balance solution, i.e.:

[0170] Using the upper-level power balance solution as a constraint, the hourly dispatching scheme is encoded into a subpopulation according to the number of power grid nodes or lines;

[0171] Calculate the fitness value of each individual in each subpopulation, expressed as:

[0172] ;

[0173] in, Indicates the individual's fitness value; The lower-level power balance objective function of the two-level hierarchical production simulation model represents the total operating cost of hourly power balance, which is used to minimize the hourly generation and dispatch-related costs. Indicates the penalty coefficient; express Constantly constrain the amount of violations;

[0174] Tournament selection, adaptive crossover, and Gaussian mutation are performed sequentially on each subpopulation. Excellent individuals are migrated between subpopulations every fixed number of generations until the preset maximum number of iterations or the preset fitness value is reached. The subpopulation with the highest fitness is used as the objective function value of the lower-level power balance to obtain the lower-level power balance solution.

[0175] The lower-level power balance solution is fed back to the upper level to correct the upper-level power balance solution until the change in the upper-level power balance objective function is less than the first threshold and the change in the lower-level power balance objective function is less than the second threshold, thus completing the convergence and obtaining the power balance result and the power balance result.

[0176] The two-layer bidirectional iterative mechanism refers to the upper layer outputting power boundary constraints to optimize the lower layer, and the lower layer's actual power generation feedback to correct the upper layer's power.

[0177] The total monthly power generation calculated by the upper layer becomes the hard upper limit of the lower layer. The total power generation scheduled by the lower layer on an hourly basis cannot exceed the number given by the upper layer. The lower layer must perform hourly optimization within this total amount. After the lower layer calculates the actual power generation, it is sent back to the upper layer. The upper layer then fine-tunes the power allocation for the next month based on the actual power generation, and iterates repeatedly until convergence.

[0178] Step Six: Evaluate and compare the two-level hierarchical solutions from multiple dimensions, and optimize them by fusion based on their superiority to obtain the power system production simulation results, specifically as follows:

[0179] In terms of objective function values, the optimization effects of the two-layer solution are compared, with the upper-layer electricity cost, absorption rate and the lower-layer power generation cost, curtailment rate and carbon emissions as the core indicators.

[0180] Constraint satisfaction dimensions: statistically analyze the number and extent of violations of energy balance, power balance, unit, grid, and energy storage constraints in the two-layer solution to measure the feasibility of the scheme;

[0181] Calculate the time dimension, record the time taken by the upper and lower layer algorithms from startup to convergence, and evaluate the overall solution efficiency;

[0182] To assess the stability of the solution, the algorithm is run multiple times, and the standard deviation of the objective function is calculated. The smaller the standard deviation, the more stable the result.

[0183] The power balance results and electricity balance results are evaluated from multiple dimensions to obtain the power system production simulation results, namely, comparing the objective function, constraint satisfaction, computation time, and result stability. If the two-level hierarchical solution is significantly better than the single-level solution in multiple dimensions and satisfies all constraints, the hierarchical solution is directly selected as the final result, or a weighted average is used to fuse the power balance results and electricity balance results to obtain the power system production simulation results, expressed as follows:

[0184] ;

[0185] in, express Real-time power system production simulation results; They represent Real-time power balance results; All of these represent weighting coefficients.

[0186] The power system production simulation results are a complete hourly power system dispatching scheme.

[0187] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0188] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0189] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0190] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0191] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A power system production simulation method based on time-series feature transfer and two-level hierarchical structure, characterized in that, include: Acquire meteorological data and power system output data; Extract features from the meteorological data and power system output data to construct a meteorological output time series feature library; Based on the meteorological output time series feature library, a meteorological sequence for the planning level year is generated. Based on the meteorological sequence for the planning level year, the meteorological output time series feature library is transferred to the planning year using a pre-constructed nonlinear mapping model that introduces a migration correction coefficient, resulting in long-term wind, solar and thermal output curves. Based on the long-term output curves of wind, solar and thermal power, the pre-constructed two-layer hierarchical production simulation model is solved by two-layer bidirectional iterative solution to obtain the power balance result and the electricity balance result. The two-level hierarchical production simulation model is constructed with power balance as the guide and power balance as the main factor. The power balance results and the power balance results are evaluated from multiple dimensions to obtain the power system production simulation results.

2. The power system production simulation method based on time-series feature migration and two-level hierarchical structure according to claim 1, characterized in that, The meteorological data includes wind speed, irradiance, ambient temperature, cloud cover, and air pressure. The power system output data includes wind power output data, photovoltaic power output data, and solar thermal power output data.

3. The power system production simulation method based on time-series feature migration and two-level hierarchical structure according to claim 1, characterized in that, The meteorological output time series feature database includes time series autocorrelation coefficient, climbing rate, extreme value duration and weather transition probability; The time-series autocorrelation coefficient is expressed as: ； The climbing rate is expressed as: ； The duration of the extreme value is expressed as follows: ； The weather shift probability is expressed as: ； in, Indicates a delay of The time series autocorrelation coefficient; Indicates the total time; They represent time, Weather / output sequence values ​​at any given time; Represents the mean of the sequence; express Constantly adjust the climbing rate; They represent time, New energy sources are always contributing their power; Indicates the time step; Indicates the number of hours the extreme output lasts; These represent the low output threshold and the high output threshold, respectively. Indicates an indicator function; They represent time, Real-time weather conditions; Indicates weather conditions Total number of occurrences; Indicates weather conditions arrive The number of transfers; This indicates the probability of a weather state transition.

4. The power system production simulation method based on time-series feature migration and two-level hierarchical structure according to claim 1, characterized in that, The planned annual meteorological sequence is represented as follows: ； in, They represent time, Time-based planning of horizontal annual meteorological sequences; This represents the weather transfer matrix.

5. The power system production simulation method based on time-series feature migration and two-level hierarchical structure according to claim 1, characterized in that, The nonlinear mapping models that introduce migration correction coefficients include wind power output time-series characteristic migration models, photovoltaic power output models, and solar thermal system models; The wind power output time-series feature transfer model is expressed as follows: ； ； The photovoltaic power output model is expressed as follows: ； The photothermal system model is represented as follows: ； ； in, express Wind power output at all times; Indicates the rated power of the wind turbine unit; express Wind speed at all times; Indicates the cut-in wind speed; Indicates the rated wind speed; express Temporal feature migration correction coefficient; This represents the autocorrelation coefficient of wind power; express Wind power ramp rate at any time; Indicates the historical average rate of ascent; All represent weighting coefficients; express Photovoltaic power output at all times; Indicates the rated power of the photovoltaic system; express irradiance at any given time; Indicates standard test irradiance; express Ambient temperature at all times; Indicates the standard test environment temperature; Indicates the power temperature coefficient; express Real-time heat collection power; Indicates the area of ​​the heat collection field; Indicates heat collection efficiency; They represent time, Real-time thermal storage capacity; Indicates the time step; express Real-time solar thermal power generation capacity; This indicates the efficiency of concentrated solar power generation.

6. The power system production simulation method based on time-series feature migration and two-level hierarchical structure according to claim 1, characterized in that, The two-layer hierarchical production simulation model includes an upper-layer power balance objective function, a lower-layer power balance objective function, monthly power balance constraints, hourly power balance constraints, power generation equipment capacity constraints, unit ramp-up constraints, grid security constraints, and spinning reserve constraints.

7. The power system production simulation method based on time-series feature migration and two-level hierarchical structure according to claim 6, characterized in that, The objective function for upper-level power balance is expressed as: ； The objective function for lower-level power balance is expressed as follows: ； in, These represent the upper-level power balance objective function and the lower-level power balance objective function, respectively. This indicates taking the minimum value; They represent Monthly power generation cost, grid loss cost, curtailment cost, and carbon emission cost; They represent Costs include fuel costs, operation and maintenance costs, unit start-up and shutdown costs, and network loss costs.

8. The power system production simulation method based on time-series feature migration and two-level hierarchical structure according to claim 6, characterized in that, The monthly electricity balance constraint is expressed as: ； The hourly power balance constraint is expressed as follows: ； The capacity constraint of the power generation equipment is expressed as follows: ； The unit ramping constraint is expressed as follows: ； The power grid security constraints are expressed as follows: ； ； The rotational spare constraint is expressed as follows: ； in, They represent time, Regular units contribution; They represent Real-time wind power output, photovoltaic power output, solar thermal power generation, inter-provincial power exchange capacity, and grid loss capacity; Indicates the time step; express Constant load; They represent Monthly network loss electricity, inter-provincial exchange electricity; They represent the generating units. Minimum output, maximum output; Indicates the unit Maximum climbing speed; express Time grid node Voltage amplitude; Representing power grid nodes Lower voltage limit; express Timetable Transmission power; Indicates the line Thermal stability limit; express Constantly rotate reserve capacity; This represents the reserve factor.

9. The power system production simulation method based on time-series feature migration and two-level hierarchical structure according to claim 6, characterized in that, Based on the aforementioned long-term power output curves of wind, solar, and thermal power, a pre-constructed two-layer hierarchical production simulation model is solved through a two-layer bidirectional iterative process to obtain the power balance results and the electricity balance results, including: Obtain the current monthly power consumption plan and the new information for the next month; the current monthly power consumption plan includes initial output, load, new energy forecast, and exchange plan, and the new information for the next month includes load, new energy, and unit status; Based on the current monthly electricity consumption plan and the new information for the next month, a search and adjustment is performed in the neighborhood of the current monthly electricity consumption plan to obtain the current adjusted monthly plan. A greedy strategy is then used to reorder the current adjusted monthly plan to generate a new monthly electricity consumption solution. The greedy strategy is formulated based on the monthly scale and sorted in ascending order according to the unit electricity cost. Based on the new month's electricity solution, calculate the objective function value of the upper-level electricity balance until the preset termination condition is met, and obtain the upper-level electricity balance solution; Using the upper-level power balance solution as a constraint, the hourly dispatching scheme is encoded into a subpopulation according to the number of power grid nodes or lines; Calculate the fitness value of each individual in each subpopulation, expressed as: ； in, Indicates the individual's fitness value; This represents the objective function for lower-level power balance. Indicates the penalty coefficient; express Constantly constrain the amount of violations; Tournament selection, adaptive crossover, and Gaussian mutation are performed sequentially on each subpopulation. Excellent individuals are migrated between subpopulations every fixed number of generations until the preset maximum number of iterations or the preset fitness value is reached. The subpopulation with the highest fitness is used as the objective function value of the lower-level power balance to obtain the lower-level power balance solution. The lower-level power balance solution is fed back to the upper level to correct the upper-level power balance solution until the change in the upper-level power balance objective function is less than the first threshold and the change in the lower-level power balance objective function is less than the second threshold, thus completing the convergence and obtaining the power balance result and the power balance result.

10. The power system production simulation method based on time-series feature migration and two-level hierarchical structure according to claim 1, characterized in that, The power balance results and the electricity balance results are evaluated from multiple dimensions to obtain the power system production simulation results, including: By using a weighted average to fuse the energy balance results and the power balance results, the power system production simulation results are obtained, as follows: ； in, express Real-time power system production simulation results; They represent Real-time power balance results; All of these represent weighting coefficients.

Images