An extreme weather new energy output response evaluation system and method based on mechanical inertia

Abstract

The application discloses a new energy output response evaluation system and method under extreme weather based on mechanical inertia, and relates to the technical field of power system dispatching support. The system carries out time alignment verification on multi-source data in the evaluation time window and generates confidence weight; identifies extreme weather types and key meteorological parameters as prediction constraints; determines response lag and upper limit of climbing rate based on equivalent moment of inertia on the wind power side, outputs output prediction and climbing element index; on the photovoltaic side, remote sensing is used to reverse irradiance, and prediction, measurement and historical output are fused to obtain output prediction and climbing element index; further, three-level climbing quantitative indexes of stations, clusters and regions are constructed and classified, and preventive dispatching suggestions are generated and checked in combination with real-time operation constraints. The application realizes reliable characterization and weighted utilization under a unified time base, forms consistent extreme weather scenario constraints, and improves the consistency of cross-scale risk assessment and the executability of dispatching suggestions.

Description

Technical Field

[0001] This invention relates to the field of power system dispatch support technology, and in particular to a new energy output response evaluation system and method based on mechanical inertia under extreme weather conditions. Background Technology

[0002] With the continuous growth of new energy installed capacity, wind power and photovoltaic output are significantly affected by meteorological factors, exhibiting strong volatility and high uncertainty. Under extreme weather conditions such as cold waves, strong winds, and severe convection, new energy output is more prone to rapid increases or decreases in a short period of time, which in turn impacts operational objectives such as grid reserve configuration, power flow control at key sections, and frequency security.

[0003] Existing technologies typically rely on multi-source data, including meteorological information, remote sensing information, power station monitoring information, and power grid operation information, for assessment and dispatch support. However, they generally suffer from the following problems: heterogeneity and differences exist among the multi-source data, including inconsistencies in time references, sampling periods, transmission delays, and data quality, leading to unstable assessment inputs and consequently compromising the stability and reliability of assessment conclusions; differences exist in the response mechanisms of wind power and photovoltaics, and existing assessment methods often fail to account for the changing characteristics of these two types of new energy under extreme weather conditions, resulting in increased bias in depicting rapid changes and thus failing to accurately reflect risk levels; inconsistencies exist in risk characterization across multiple spatial scales, with risk manifestations and resource requirements differing at the power station, cluster, and regional scales. The lack of a unified standard leads to conflicting conclusions at different scales, making it impossible to establish consistent risk quantification and classification criteria; and the conversion of assessment results into dispatch recommendations suffers from low efficiency and insufficient executability, often relying on fixed rules or experience, making it difficult to adapt to changes in power grid operation status in a timely manner and unable to quickly output actionable preventative recommendations under safe operation constraints.

[0004] Therefore, a new energy output response assessment and dispatch support scheme for extreme weather conditions is needed to improve the reliability of the assessment in the face of objective differences in multi-source data, accurately depict the rapid change process, form consistent risk quantification and classification results across multiple scales, and output executable preventive dispatch suggestions under the premise of meeting the constraints of grid safety operation, thereby improving the safety and dispatch efficiency of high-proportion new energy grids under extreme weather conditions. Summary of the Invention

[0005] To address the aforementioned issues, the present invention aims to provide a renewable energy output response assessment system and method based on mechanical inertia under extreme weather conditions: enhancing the stability and reliability of the assessment under multi-source information conditions; improving the identification and risk expression capabilities of rapid change processes under the differentiated change characteristics of wind power and photovoltaics; forming consistent quantitative and hierarchical results for different spatial scales such as power stations, clusters, and regions; and transforming the assessment results into preventive dispatching suggestions that meet the constraints of grid safety operation, thereby improving the timeliness and feasibility of dispatching support.

[0006] To achieve the above objectives, the present invention employs the following technical solution:

[0007] On the one hand, this invention provides a new energy output response evaluation system based on mechanical inertia under extreme weather conditions, comprising:

[0008] The data acquisition and verification module is used to collect multi-source data within the assessment time window of the assessment area, and to perform time stamping, time alignment and data verification on the multi-source data, and generate confidence weights for weighted input.

[0009] The extreme weather identification module is used to identify extreme weather events based on meteorological elements in multi-source data that has been aligned and verified, and outputs the extreme weather type and a set of key meteorological parameters.

[0010] The wind power ramping early warning module is used to construct a wind power output physical response model under the constraints of extreme weather types and key meteorological parameter sets, combined with the equivalent rotational inertia of the wind turbine rotor and transmission system, to determine the response lag and the upper limit of the ramping rate. Based on the wind power output sequence, it outputs wind power ramping event element indicators and wind power output prediction results.

[0011] The photovoltaic inversion and early warning module is used to invert the surface irradiance based on meteorological satellite remote sensing data and irradiance inversion model, and to fuse the inverted irradiance with numerical weather prediction data, surface meteorological measurement data and historical photovoltaic power station output sequences to obtain photovoltaic output prediction results, and output photovoltaic ramp-up event element indicators.

[0012] The multi-scale index classification module is used to construct and calculate quantitative indicators of new energy output ramping based on wind power ramping event element indicators and photovoltaic ramping event element indicators, and to classify the ramping impact intensity based on the index thresholds, and output classification results including ramping intensity levels.

[0013] The scheduling suggestion generation module is used to generate preventive scheduling suggestions and output scheduling action suggestions based on classification results and real-time power grid operation data, combined with power grid safety operation constraints. Before outputting, it performs a feasibility check on power grid safety operation constraints and corrects the scheduling action suggestions based on the feasibility check results.

[0014] As a preferred embodiment of the present invention, the data acquisition and verification module includes:

[0015] Alignment units are used to perform uniform temporal granularity transformation and time drift correction with the evaluation time window boundary as the alignment anchor point;

[0016] The boundary alignment unit is used to locate the ramping event boundary based on the wind power output sequence and the photovoltaic power output sequence, and to perform secondary alignment on numerical weather forecast data, meteorological satellite remote sensing data, new energy station operation data and power grid dispatch plan data within the neighborhood of the ramping event boundary. The neighborhood of the ramping event boundary is a preset time range or a preset number of sampling points before and after the boundary.

[0017] The secondary alignment includes: calculating the cross-source cross-correlation function within the neighborhood of the climbing event boundary, and determining the cross-source delay estimate of the cross-source correlation peak within a preset delay search window; performing reliability discrimination on the cross-source correlation peak, the reliability discrimination including peak amplitude threshold discrimination and main peak to secondary peak ratio discrimination; when the reliability discrimination fails, fixing the global delay of the alignment anchor point or using the cross-source delay estimate of the previous evaluation time window; and performing secondary alignment based on the cross-source delay estimate.

[0018] The consistency verification and weighting unit is used to perform cross-source consistency verification on the multi-source data after secondary alignment and generate data quality labels and confidence weights. The confidence weights are determined by the cross-source consistency residuals through a monotonic mapping rule, which is either a piecewise linear mapping or a lookup table mapping. The confidence weights are used to perform weighted input on the input data entering the extreme weather identification module, the wind power ramp-up early warning module, and the photovoltaic inversion early warning module.

[0019] As a preferred embodiment of the present invention, the extreme weather identification module is used to identify extreme weather events within the evaluation time window and output the extreme weather type and key meteorological parameter set. The key meteorological parameter set serves as the prediction constraint input for the wind power ramp-up early warning module and the photovoltaic inversion early warning module. Furthermore, the extreme weather type is used to select the parameter configuration scheme of the wind power output physical response model and the input feature set of the irradiance inversion model. The parameter configuration scheme is determined by the preset parameter configuration table according to the extreme weather type index, and the input feature set is determined by the preset feature template table according to the extreme weather type index.

[0020] As a preferred embodiment of the present invention, the wind power output physical response model of the wind power ramping early warning module determines the response lag and the upper limit of the achievable ramping rate using the equivalent moment of inertia, and uses the response lag and the upper limit of the achievable ramping rate as constraints for ramping event prediction; the equivalent moment of inertia is determined by the wind turbine and transmission system parameter table and updated with the operating status of the wind turbine generator set, including the pitch angle status, speed status and grid connection control mode status;

[0021] The response lag is determined by the equivalent moment of inertia and the speed state through a preset lag mapping function or a preset lag lookup table relationship; the achievable ramp rate upper limit is determined by the equivalent moment of inertia, the pitch angle state, and the grid-connected control mode state through a preset upper limit mapping function or a preset upper limit lookup table relationship; the achievable ramp rate upper limit is limited to not exceeding the power change rate limit allowed by the unit control system.

[0022] As a preferred embodiment of the present invention, the wind power ramping event element indicators include ramping direction, ramping amplitude, maximum ramping rate, and duration; the maximum ramping rate and duration are calculated based on the input data that has been double-aligned within the boundary neighborhood of the ramping event; the maximum ramping rate is the maximum value of the difference ratio of power change to time within the boundary neighborhood, and the duration is the length of time during which the power change continuously maintains the same ramping direction and exceeds a preset amplitude threshold.

[0023] As a preferred embodiment of the present invention, the photovoltaic inversion early warning module determines the weather model category based on extreme weather types and key meteorological parameter sets, and selects the input feature set and parameter configuration of the irradiance inversion model according to the weather model category; the selection rule for the input feature set is to call the feature template table according to the weather model category or to call the segmentation rule table according to the weather model category; the photovoltaic inversion early warning module inverts the surface irradiance based on meteorological satellite remote sensing data, estimates the cloud movement characteristics based on remote sensing time series changes, performs correction processing on the irradiance abrupt change caused by cloud movement, and generates a corrected irradiance sequence.

[0024] As a preferred embodiment of the present invention, the photovoltaic inversion early warning module integrates the corrected irradiance sequence with numerical weather forecast data, surface meteorological measurement data, and historical photovoltaic power station output sequences to obtain photovoltaic output prediction results. The integration modeling selects a subset of integration input features and configures integration weights according to the weather model category; and the integration weight configuration is obtained by normalizing the confidence weights or by mapping the confidence weights through a lookup table; and outputs photovoltaic ramping event element indicators based on the photovoltaic output prediction results. The photovoltaic ramping event element indicators include ramping direction, ramping amplitude, maximum ramping rate, and duration.

[0025] As a preferred embodiment of the present invention, the index threshold of the multi-scale index classification module includes a basic threshold and an adaptive correction amount. The adaptive correction amount is determined by the margin index of the real-time operation constraint of the power grid through a piecewise function. The margin index includes frequency security margin, rapid reserve margin, and power flow margin of critical sections. The classification includes performing cross-scale consistency verification on the classification results of the station scale, cluster scale, and regional scale, and performing write-back correction on the lower scale classification results based on the higher scale classification results when there is inconsistency.

[0026] As a preferred embodiment of the present invention, the scheduling suggestion generation module will increase the reserve requirement when generating preventive scheduling suggestions. As a fast backup constraint, the target quantile level With increased reserve requirements satisfy:

[0027] ;

[0028] In the formula: To underestimate the marginal cost of risk; To overestimate the marginal cost of risk; For operational risk ratio; For quantile operators; The equivalent power mismatch prediction value for the evaluation area is constructed based on wind power output prediction results and photovoltaic power output prediction results, combined with grid dispatch plan data. The measured values ​​of equivalent power mismatch in the evaluation area are constructed based on real-time power grid operation data. and The parameters are configured by the market or regulatory strategy table and can be updated at any time; the dispatch suggestion generation module performs a feasibility check of power grid safety operation constraints before outputting preventive dispatch suggestions. The feasibility check includes one or more of the following: reserve constraint check and critical section power flow constraint check; the execution receipt corresponding to the dispatch action suggestion and the real-time power grid operation data after execution are used for updating. The prediction error sample set is used to continuously update and adjust the reserve requirement. .

[0029] On the other hand, the present invention also provides a method for evaluating the output response of new energy sources under extreme weather conditions based on mechanical inertia, which is applied to the aforementioned evaluation system for evaluating the output response of new energy sources under extreme weather conditions based on mechanical inertia, and includes the following steps:

[0030] S1. Collect multi-source data within the evaluation time window of the evaluation area, perform time stamping, time alignment and data verification on the multi-source data, and generate confidence weights for weighted input;

[0031] S2. Identify extreme weather events based on meteorological elements in multi-source data that have undergone alignment verification, and output the extreme weather type and key meteorological parameter set;

[0032] S3. Under the constraints of the extreme weather type and key meteorological parameter set, a wind power output physical response model is constructed by combining the equivalent rotational inertia of the wind turbine rotor and transmission system to determine the response lag and the upper limit of the achievable ramp rate. Based on the wind power output sequence, the wind power ramp event element indicators and wind power output prediction results are output.

[0033] S4. Based on meteorological satellite remote sensing data and irradiance inversion model, the surface irradiance is inverted, and the inverted irradiance is fused with numerical weather prediction data, surface meteorological measurement data and historical photovoltaic power station output sequence to obtain photovoltaic output prediction results, and output photovoltaic ramp-up event element indicators.

[0034] S5. Based on the wind power ramping event element indicators and the photovoltaic ramping event element indicators, construct and calculate the quantitative indicators of new energy output ramping for three spatial scales: power station, cluster and region, and classify the ramping influence intensity based on the indicator threshold, and output the classification results including the ramping intensity level.

[0035] S6. Based on the classification results and real-time power grid operation data, generate preventive scheduling suggestions and output scheduling action suggestions in combination with power grid safety operation constraints; perform a feasibility check of power grid safety operation constraints before outputting, and revise the scheduling action suggestions based on the feasibility check results.

[0036] The beneficial effects of this invention are as follows: Addressing the problem of inconsistent time bases for multi-source data under extreme weather conditions, and the instability of assessment inputs due to cross-source delays and quality differences, this invention performs time consistency processing and verification on multi-source data within the assessment time window of the assessment area, generating confidence weights for weighted inputs. This ensures that multi-source data possesses comparability and reliability under the same time base, improving the stability and consistency of the assessment input. The system identifies extreme weather events based on aligned and verified meteorological elements and outputs extreme weather types and key meteorological parameter sets, enabling wind power and photovoltaic assessments to be conducted under unified scenario constraints, reducing deviations introduced by inconsistent constraints under extreme conditions. On the wind power side, the system constructs a physical response model for wind power output by combining the equivalent rotational inertia of the wind turbine and transmission system to determine the response lag and the upper limit of the achievable ramp rate. The response lag and the upper limit of the achievable ramp rate are used as prediction constraints to output wind power ramp event element indicators and wind power output prediction results, giving ramp risk assessment clear achievable boundaries and consistent constraint caliber. On the photovoltaic side, the system obtains surface irradiance based on meteorological satellite remote sensing data and an irradiance inversion model. It then integrates this irradiance with numerical weather prediction data, surface meteorological measurements, and historical photovoltaic power plant output sequences to model and output photovoltaic output prediction results and photovoltaic ramp-up event indicators, improving the ability to characterize output changes under abrupt scenarios. The system constructs and calculates quantitative indicators for renewable energy output ramp-up at three spatial scales: power plant, cluster, and region, and outputs ramp-up intensity levels accordingly, enabling risk expression to be aligned across scales and supporting tiered handling. Based on classification results and real-time grid operation data, the system generates preventative dispatch recommendations in conjunction with grid safety operation constraints. Feasibility checks and corrections are performed before output, ensuring that dispatch recommendations are consistent with safety constraints such as reserve and critical sections, improving the executability and engineering feasibility of dispatch recommendations, and enhancing risk prevention and control capabilities under extreme weather conditions. Attached Figure Description

[0037] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the 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. Wherein:

[0038] Figure 1 This is a schematic diagram of the modular structure of the system of the present invention; Figure 2 This is a flowchart of the method of the present invention. Detailed Implementation

[0039] To make the objectives, technical solutions, and beneficial effects of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the described embodiments are only a part of the embodiments of the present invention, and not all of them; other embodiments obtained by those skilled in the art without departing from the concept of the present invention should all fall within the protection scope of the present invention.

[0040] like Figure 1 The illustration shows an embodiment of the present invention, which provides a new energy output response evaluation system based on mechanical inertia under extreme weather conditions. To facilitate a consistent description of data organization, field meanings, and processing methods among the modules in this embodiment, the following conventions are made regarding the field definitions of several key objects. The fields of the following objects must at least include the listed fields and can be expanded according to engineering implementation needs; the field names of each object can use equivalent names, but should maintain semantic consistency.

[0041] Evaluation time window: includes at least the window start time, window end time, and time step.

[0042] Climbing event boundary: includes at least the start boundary time, the end boundary time, and the boundary neighborhood configuration, which can be represented by a minute range or a range of sampling points.

[0043] Confidence weight: includes at least a weight value (0-1), a quality label, and a mapping rule version or effective period; the quality label is used to identify the data status, and the mapping rule version or effective period is used to support result reproduction and auditing.

[0044] Key meteorological parameter set: includes at least the start and end of the event, duration and intensity indicators, and may optionally include fields for impact level or impact range; the key meteorological parameter set may carry a data version identifier.

[0045] Weather mode category: It should include at least the category code and a summary of the parameters used to determine the criteria, and may also include a template or rule version identifier.

[0046] Wind power / solar power output prediction results: at least include prediction timestamp sequences and prediction power sequences; wind power output prediction results may carry constraint parameter versions (response lag / ramp limit), and solar power output prediction results may carry fused weight versions.

[0047] Three-level scale quantification metrics: These include at least scale labels, amplitude metrics, and rate metrics, and optionally a duration field.

[0048] Slope intensity level: includes at least a level code and a threshold field, wherein the threshold field includes at least a base threshold and an adaptive correction amount, and may carry a piecewise function version identifier.

[0049] Scheduling action suggestions: These should include at least the action type, target value, applicable time window, and verification results, and may also include fields for the number of corrections and status labels.

[0050] Version unification rules: In this embodiment, the configuration items such as parameter configuration table, feature template table, segmentation rule table, segmentation function and mapping table can carry at least one of the following: version number, generation time or effective period, to support reproduction and auditing; the above version information can be output along with the corresponding output results or associated by configuration reference.

[0051] (1) Data acquisition and verification module

[0052] In this embodiment, the data acquisition and verification module performs unified organization, time consistency processing, and quality control on multi-source data within the evaluation time window to provide stable and comparable input data to subsequent modules.

[0053] This system is used to collect multi-source data within the assessment time window of the evaluation area, and performs time stamping, time alignment, and data verification on the multi-source data, generating confidence weights for weighted input. The multi-source data may include numerical weather prediction data, meteorological satellite remote sensing data, new energy power plant operation data, and power grid dispatch plan data, etc. Time stamping includes at least a unified timestamp and data source identifier; data verification includes at least the identification of missing, out-of-bounds, and anomalous jumps, and outputs corresponding data quality labels and confidence weights.

[0054] The data acquisition and verification module includes:

[0055] The alignment unit is used to perform uniform time granularity transformation and time drift correction with the evaluation time window boundary as the alignment anchor point; uniform time granularity transformation is used to resample data from different sampling periods to a uniform time step; time drift correction is used to eliminate the overall time offset of different data sources.

[0056] A boundary alignment unit is used to locate the ramp-up event boundary based on the wind power output sequence and the photovoltaic power output sequence. Within the neighborhood of the ramp-up event boundary, it performs secondary alignment on numerical weather prediction data, meteorological satellite remote sensing data, new energy power station operation data, and power grid dispatch plan data. The neighborhood of the ramp-up event boundary is a preset time range or a preset number of sampling points before and after the boundary. In one embodiment, the differential rate of change is calculated for the power output sequence. When the absolute value of the rate of change continuously exceeds a threshold and reaches a preset duration or preset number of points, a start boundary is determined. When the rate of change falls back to within the threshold and continues to reach a preset duration or preset number of points, an end boundary is determined. The boundary neighborhood can be 5-30 minutes or 10-60 points before and after the boundary.

[0057] Secondary alignment includes: calculating the cross-source cross-correlation function within the neighborhood of the climbing event boundary, and determining the cross-source delay estimate of the cross-source correlation peak within a preset delay search window; performing reliability discrimination on the cross-source correlation peak, the reliability discrimination including peak amplitude threshold discrimination and main peak to secondary peak ratio discrimination; when the reliability discrimination fails, fixing the global delay of the alignment anchor point or using the cross-source delay estimate of the previous evaluation time window; performing secondary alignment based on the cross-source delay estimate; the preset delay search window can be configured to ±1 to ±10 sampling intervals; when the reliability discrimination fails, using the global delay or the delay estimate of the previous evaluation time window to ensure the stability of the alignment result.

[0058] The consistency verification and weighting unit performs cross-source consistency verification on the multi-source data after secondary alignment and generates data quality labels and confidence weights. The confidence weights are determined by the cross-source consistency residuals through a monotonic mapping rule, which can be a piecewise linear mapping or a lookup table mapping. The confidence weights are used to weight the input data entering the extreme weather identification module, wind power ramp-up early warning module, and photovoltaic inversion early warning module. The cross-source consistency residuals can be the absolute difference or squared difference of cross-source features at the same time point or within a sliding window; the monotonic mapping satisfies the condition that a decrease in residuals leads to an increase in weights. The output can include the version number or effective period identifier of the confidence weight mapping rule for reproduction and auditing purposes.

[0059] (2) Extreme weather identification module

[0060] The extreme weather identification module receives meteorological elements from multi-source data that has been aligned and verified, identifies extreme weather events within the evaluation time window, outputs the extreme weather type and a set of key meteorological parameters, and provides the results to the subsequent prediction module as a constraint input and a basis for configuration selection.

[0061] This tool is used to identify extreme weather events based on meteorological elements in multi-source data that has been aligned and verified, and outputs extreme weather types and sets of key meteorological parameters.

[0062] The extreme weather identification module is used to identify extreme weather events within an evaluation time window and output the extreme weather type and a set of key meteorological parameters. This set of key meteorological parameters serves as the prediction constraint input for the wind power ramp-up early warning module and the photovoltaic inversion early warning module. In one embodiment, window statistics are calculated for selected meteorological elements and compared with preset thresholds. When the threshold is met and the event persists for a preset duration, an extreme weather event is determined to have occurred. Adjacent time segments that meet the conditions are merged, and short-term isolated segments are downgraded or removed according to rules. The set of key meteorological parameters includes at least the event start and end times, duration, and intensity index, and can be expanded to include impact level or impact range fields.

[0063] Furthermore, extreme weather types are used to select the parameter configuration scheme for the wind power output physical response model and the input feature set for the irradiance inversion model. The parameter configuration scheme is determined by the preset parameter configuration table indexed by extreme weather type, and the input feature set is determined by the preset feature template table indexed by extreme weather type. The output can include the version number or effective period identifier of the parameter configuration table and feature template table to support reproduction.

[0064] (3) Wind power ramp-up early warning module

[0065] Under the constraints of extreme weather types and key meteorological parameters, the wind power ramping early warning module constructs a physical response model of wind power output by combining the equivalent rotational inertia of the wind turbine rotor and transmission system. This model is used to determine the response lag and the upper limit of the achievable ramping rate, and outputs wind power output prediction results and ramping event element indicators accordingly.

[0066] This model is used to construct a physical response model for wind power output under the constraints of extreme weather types and key meteorological parameter sets, combining the equivalent rotational inertia of the wind turbine rotor and transmission system to determine the response lag and the upper limit of the achievable ramp rate. Based on the wind power output sequence, it outputs wind power ramp event element indicators and wind power output prediction results.

[0067] The wind power output physical response model of the wind power ramping early warning module determines the response lag and the upper limit of the achievable ramping rate using the equivalent rotational inertia, and uses the response lag and the upper limit of the achievable ramping rate as constraints for ramping event prediction; the equivalent rotational inertia is determined by the wind turbine and transmission system parameter table and updated with the operating status of the wind turbine generator set, including the pitch angle status, speed status and grid connection control mode status.

[0068] The response lag is determined by the equivalent moment of inertia and speed state through a preset lag mapping function or a preset lag lookup table relationship; the achievable ramp rate upper limit is determined by the equivalent moment of inertia, pitch angle state, and grid-connected control mode state through a preset upper limit mapping function or a preset upper limit lookup table relationship; the achievable ramp rate upper limit is limited to not exceeding the power change rate limit allowed by the unit control system. During ramp event prediction, the response lag and the achievable ramp rate upper limit are applied to the prediction output, ensuring that the effective change starting point of the prediction sequence is delayed by no less than the response lag relative to the current time, and that the power change rate at any adjacent time does not exceed the achievable ramp rate upper limit. Simultaneously, an upper bound truncation is applied to the achievable ramp rate upper limit to satisfy the power change rate limit allowed by the unit control system.

[0069] In one embodiment, an unconstrained prediction sequence can be obtained by first performing persistent prediction or sliding linear extrapolation on the wind power output sequence, and then the aforementioned constraints can be applied to obtain the wind power output prediction result. The equivalent moment of inertia parameter table, hysteresis mapping / upper limit mapping, or lookup table relationship can all be configured as parameter tables and accompanied by a version number or effective period identifier.

[0070] The wind power ramping event element indicators include ramping direction, ramping amplitude, maximum ramping rate, and duration. The maximum ramping rate and duration are calculated based on the input data that has been double-aligned within the boundary neighborhood of the ramping event. The maximum ramping rate is the maximum value of the difference ratio of power change to time within the boundary neighborhood, and the duration is the length of time during which the power change continuously maintains the same ramping direction and exceeds a preset amplitude threshold.

[0071] (4) Photovoltaic inversion early warning module

[0072] The photovoltaic inversion and early warning module inverts surface irradiance based on meteorological satellite remote sensing data and corrects for abrupt changes in irradiance caused by cloud movement. Then, it integrates numerical weather prediction data, surface meteorological measurement data, and historical photovoltaic power station output sequences to obtain photovoltaic output prediction results and outputs photovoltaic ramp-up event element indicators.

[0073] This is used to invert surface irradiance based on meteorological satellite remote sensing data and irradiance inversion model, and to fuse the inverted irradiance with numerical weather prediction data, surface meteorological measurement data and historical photovoltaic power station output sequences to obtain photovoltaic output prediction results, and output photovoltaic ramp-up event element indicators.

[0074] The photovoltaic inversion and early warning module determines the weather model category based on extreme weather types and key meteorological parameter sets, and selects the input feature set and parameter configuration of the irradiance inversion model according to the weather model category. The selection rule for the input feature set is to call the feature template table or the segmentation rule table according to the weather model category. The photovoltaic inversion and early warning module inverts the surface irradiance based on meteorological satellite remote sensing data, estimates cloud movement characteristics based on remote sensing time-series changes, performs correction processing on irradiance abrupt changes caused by cloud movement, and generates a corrected irradiance sequence. The weather model category can be obtained by mapping extreme weather types or by combining key meteorological parameter sets. Abrupt change correction can detect the rate of change of inverted irradiance within a short time window. When the rate of change exceeds the abrupt change threshold and is consistent with the cloud movement characteristics, the abrupt change amplitude is corrected or a smooth transition is performed according to the correction rules, and the corrected irradiance sequence is output.

[0075] The photovoltaic (PV) inversion and early warning module integrates the corrected irradiance sequence with numerical weather prediction data, surface meteorological measurement data, and historical PV power plant output sequences to obtain PV output prediction results. The fusion modeling selects a subset of fusion input features and configures fusion weights according to weather model categories. The fusion weights are obtained by normalizing confidence weights or by mapping confidence weights through a lookup table. Based on the PV output prediction results, the module outputs PV ramp-up event element indicators, including ramp-up direction, ramp-up amplitude, maximum ramp-up rate, and duration. In one embodiment, the fusion modeling can employ weighted linear regression. The inputs are corrected irradiance and forecast, measurement, and historical output features. The fusion weights are obtained by normalizing confidence weights or by mapping through a lookup table and participate in the weighted solution, outputting the PV output prediction results. The feature template table, segmentation rule table, and fusion weight mapping table can all include version numbers or effective period identifiers.

[0076] If the irradiance inversion model or fusion model is built using machine learning, the training and model version can be completed offline, and only the inference output can be executed during the online operation phase to meet the real-time requirements of the scheduling business. Offline training includes data alignment and cleaning, feature construction, training / validation partitioning, loss minimization and model solidification.

[0077] (5) Multi-scale index classification module

[0078] The multi-scale index classification module aggregates wind power ramping event element indicators and photovoltaic ramping event element indicators according to the affiliation relationship of site-cluster-region, constructs a three-level spatial scale quantitative index of new energy output ramping, and classifies the ramping impact intensity based on thresholds, outputting classification results.

[0079] This tool is used to construct and calculate quantitative indicators for renewable energy output ramping at three spatial scales: power plant, cluster, and region, based on wind power ramping event element indicators and photovoltaic ramping event element indicators. It also classifies the ramping impact intensity based on indicator thresholds, outputting classification results including ramping intensity levels. The power plant-scale quantitative indicators are obtained by combining elements such as wind and solar ramping amplitude, maximum ramping rate, and duration at a single power plant. The cluster-scale quantitative indicators are obtained by aggregating indicators from power plants within the cluster according to capacity weights or preset weights. The region-scale quantitative indicators are obtained by further aggregating indicators based on the cluster-scale indicators. The quantitative indicators must include at least amplitude and rate representations.

[0080] The multi-scale index classification module's index thresholds include a base threshold and an adaptive correction amount. The adaptive correction amount is determined by a piecewise function based on the margin indicators of real-time power grid operation constraints. These margin indicators include frequency safety margin, rapid reserve margin, and power flow margin at critical sections. The base threshold is configured according to operating procedures or historical statistics. The adaptive correction amount is obtained by mapping the margin indicators through a piecewise function to adjust the threshold as the operating margin changes. The piecewise function may include a version number or an effective period identifier.

[0081] The classification process includes cross-scale consistency checks on the classification results at the station, cluster, and regional scales. In case of inconsistency, the classification results at lower scales are corrected by writing back the results at higher scales. When a higher scale level is higher than a lower scale level within the same time period, the classification results at lower scales are corrected by writing back the results at higher scales to ensure that the lower scale level is not lower than the higher scale level, or the lowest permissible level at a lower scale is determined by a preset mapping before writing back the results.

[0082] (6) Scheduling suggestion generation module

[0083] The dispatch suggestion generation module generates preventive dispatch suggestions and outputs dispatch action suggestions based on classification results and real-time power grid operation data, combined with power grid safety operation constraints. Before outputting, it performs a feasibility check on power grid safety operation constraints and corrects the dispatch action suggestions based on the check results. At the same time, it updates the prediction error sample set based on the execution receipt and real-time power grid operation data after execution, and continuously updates and adjusts the reserve demand to form a closed loop.

[0084] This tool is used to generate preventive dispatching suggestions and output dispatching action suggestions based on classification results and real-time power grid operation data, combined with power grid safety operation constraints. Before outputting the suggestions, a feasibility check of power grid safety operation constraints is performed, and the dispatching action suggestions are revised based on the feasibility check results.

[0085] This tool is used to generate preventive dispatching suggestions and output dispatching action suggestions based on classification results and real-time power grid operation data, combined with power grid safety operation constraints. Before outputting the suggestions, a feasibility check of power grid safety operation constraints is performed, and the dispatching action suggestions are revised based on the feasibility check results.

[0086] When generating preventative scheduling recommendations, the scheduling recommendation generation module will increase the reserve requirement. As a fast backup constraint, the target quantile level With increased reserve requirements satisfy:

[0087] ;

[0088] In the formula: To underestimate the marginal cost of risk; To overestimate the marginal cost of risk; For operational risk ratio; For quantile operators; The equivalent power mismatch prediction value for the evaluation area is constructed based on wind power output prediction results and photovoltaic power output prediction results, combined with grid dispatch plan data. The measured values ​​of equivalent power mismatch in the evaluation area are constructed based on real-time power grid operation data. and The parameters are configured by the market or regulatory strategy table and can be updated at any time; the dispatch suggestion generation module performs a feasibility check of power grid safety operation constraints before outputting preventive dispatch suggestions. The feasibility check includes one or more of the following: reserve constraint check and critical section power flow constraint check; the execution receipt corresponding to the dispatch action suggestion and the real-time power grid operation data after execution are used for updating. The prediction error sample set is used to continuously update and adjust the reserve requirement. .

[0089] The quantile operator can be implemented using empirical quantiles or weighted quantiles based on confidence weights. The strategy parameter table can include a version number or an effective period identifier; the prediction error sample set can be maintained using a sliding window and the window length and minimum number of valid samples can be recorded.

[0090] Before outputting preventive dispatch recommendations, the dispatch recommendation generation module performs a feasibility check on the constraints of power grid safe operation. The feasibility check includes one or more of the following: backup constraint check and critical section power flow constraint check. If the check fails, the dispatch action recommendation is iteratively corrected until the constraints are met or the maximum number of iterations is reached.

[0091] The execution receipt corresponding to the dispatch action suggestion and the real-time power grid operation data after execution are used for updating. The prediction error sample set is used to continuously update and adjust the reserve requirement. When the number of valid samples is insufficient, the backup requirement can be adjusted by using the valid value of the previous evaluation time window or by adopting a preset conservative value; when the execution receipt is missing or the real-time data quality label is abnormal, the current update can be paused or the newly added samples can be downweighted for update, and the corresponding status label can be output for subsequent traceability.

[0092] like Figure 2 As shown, another embodiment of the present invention also provides a method for evaluating the output response of new energy sources under extreme weather conditions based on mechanical inertia, applied to the aforementioned system for evaluating the output response of new energy sources under extreme weather conditions based on mechanical inertia, comprising the following steps:

[0093] S1. Collect multi-source data within the evaluation time window of the evaluation area, perform time stamping, time alignment and data verification on the multi-source data, and generate confidence weights for weighted input;

[0094] S2. Identify extreme weather events based on meteorological elements in multi-source data that have undergone alignment verification, and output the extreme weather type and key meteorological parameter set;

[0095] S3. Under the constraints of the extreme weather type and key meteorological parameter set, a wind power output physical response model is constructed by combining the equivalent rotational inertia of the wind turbine rotor and transmission system to determine the response lag and the upper limit of the achievable ramp rate. Based on the wind power output sequence, the wind power ramp event element indicators and wind power output prediction results are output.

[0096] S4. Based on meteorological satellite remote sensing data and irradiance inversion model, the surface irradiance is inverted, and the inverted irradiance is fused with numerical weather prediction data, surface meteorological measurement data and historical photovoltaic power station output sequence to obtain photovoltaic output prediction results, and output photovoltaic ramp-up event element indicators.

[0097] S5. Based on the wind power ramping event element indicators and the photovoltaic ramping event element indicators, construct and calculate the quantitative indicators of new energy output ramping for three spatial scales: power station, cluster and region, and classify the ramping influence intensity based on the indicator threshold, and output the classification results including the ramping intensity level.

[0098] S6. Based on the classification results and real-time power grid operation data, generate preventive scheduling suggestions and output scheduling action suggestions in combination with power grid safety operation constraints; perform a feasibility check of power grid safety operation constraints before outputting, and revise the scheduling action suggestions based on the feasibility check results.

[0099] In one embodiment, the multi-source data in step S1 may include one or more of numerical weather prediction data, meteorological satellite remote sensing data, new energy power station operation data, and real-time power grid operation data; time alignment uses the evaluation time window boundary as the alignment anchor point, and cross-source data can be secondarily aligned within the neighborhood of the ramp event boundary to correct cross-source time delay; confidence weights can be obtained from the cross-source consistency residuals through piecewise linear mapping or table lookup mapping, and are used to weight the input data for subsequent identification and prediction.

[0100] In one embodiment, the extreme weather types and key meteorological parameter sets in steps S2-S4 are used to impose unified scenario constraints on the wind power and photovoltaic prediction process. The wind power side determines the response lag and the upper limit of the achievable ramp rate based on the equivalent moment of inertia and uses them as prediction constraints. The photovoltaic side can correct for abrupt changes in the irradiance retrieved by remote sensing and integrate them with forecasts, measurements and historical output to model and output prediction results. In steps S5-S6, ramp quantification indicators are calculated and output hierarchically according to the scale of the station, cluster and region. After the feasibility verification of the safe operation constraints is passed, scheduling action suggestions are output. If the verification fails, the scheduling action suggestions are modified.

[0101] In summary, this invention constructs an integrated technical system for assessing and supporting the dispatch of renewable energy output in extreme weather scenarios, achieving closed-loop linkage from data processing and risk assessment to dispatch recommendation output. This system can provide reproducible and traceable risk quantification results for the assessment area without altering existing dispatching workflows, and output executable dispatch recommendations that meet safety constraints, thereby improving the efficiency of risk management and operational safety assurance under extreme weather conditions. This invention is applicable to provincial and municipal dispatch centers with high proportions of renewable energy integration, renewable energy centralized control platforms, and regional power grid operation analysis systems, and has good engineering promotion value and application prospects.

[0102] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A new energy power output response evaluation system based on mechanical inertia under extreme weather conditions, characterized in that, include: The data acquisition and verification module is used to collect multi-source data within the assessment time window of the assessment area, and to perform time stamping, time alignment and data verification on the multi-source data, and generate confidence weights for weighted input. The extreme weather identification module is used to identify extreme weather events based on meteorological elements in multi-source data that has been aligned and verified, and outputs the extreme weather type and a set of key meteorological parameters. The wind power ramping early warning module is used to construct a wind power output physical response model under the constraints of extreme weather types and key meteorological parameter sets, combined with the equivalent rotational inertia of the wind turbine rotor and transmission system, to determine the response lag and the upper limit of the ramping rate. Based on the wind power output sequence, it outputs wind power ramping event element indicators and wind power output prediction results. The photovoltaic inversion and early warning module is used to invert surface irradiance based on meteorological satellite remote sensing data and irradiance inversion model. It then integrates the inverted irradiance with numerical weather prediction data, surface meteorological measurement data, and historical photovoltaic power station output sequences to obtain photovoltaic output prediction results and output photovoltaic ramp-up event element indicators. The multi-scale index classification module is used to construct and calculate new energy output ramp-up quantitative indicators for three spatial scales: power station, cluster, and region, based on wind power ramp-up event element indicators and photovoltaic ramp-up event element indicators. It then classifies the ramp-up impact intensity based on index thresholds and outputs classification results including ramp-up intensity levels. The scheduling suggestion generation module is used to generate preventive scheduling suggestions and output scheduling action suggestions based on classification results and real-time power grid operation data, combined with power grid safety operation constraints. Before outputting, it performs a feasibility check on power grid safety operation constraints and corrects the scheduling action suggestions based on the feasibility check results.

2. The new energy output response evaluation system based on mechanical inertia under extreme weather conditions according to claim 1, characterized in that, The data acquisition and verification module includes: an alignment unit, used to perform unified time granularity conversion and time drift correction using the evaluation time window boundary as the alignment anchor point; and a boundary alignment unit, used to locate the ramp event boundary based on the wind power output sequence and photovoltaic power output sequence, and to perform secondary alignment on numerical weather prediction data, meteorological satellite remote sensing data, new energy station operation data, and power grid dispatch plan data within the neighborhood of the ramp event boundary, wherein the neighborhood of the ramp event boundary is a preset time range or a preset number of sampling points before and after the boundary; the secondary alignment includes: calculating the cross-source cross-correlation function within the neighborhood of the ramp event boundary, and determining the cross-source time delay estimate of the cross-source correlation peak within a preset time delay search window; and performing a cross-source correlation peak adjustment function. The reliability assessment includes peak amplitude threshold assessment and main peak to secondary peak ratio assessment. When the reliability assessment fails, the global delay of the alignment anchor point or the cross-source delay estimate of the previous evaluation time window is used. Secondary alignment is performed based on the cross-source delay estimate. The consistency verification and weighting unit is used to perform cross-source consistency verification on the multi-source data after secondary alignment and generate data quality labels and confidence weights. The confidence weights are determined by the cross-source consistency residuals through a monotonic mapping rule, which is a piecewise linear mapping or a lookup table mapping. The confidence weights are used to perform weighted input on the input data entering the extreme weather identification module, wind power ramp-up early warning module, and photovoltaic inversion early warning module.

3. The new energy output response evaluation system based on mechanical inertia under extreme weather conditions according to claim 1, characterized in that, The extreme weather identification module is used to identify extreme weather events within the evaluation time window and output the extreme weather type and key meteorological parameter set. The key meteorological parameter set serves as the prediction constraint input for the wind power ramp-up early warning module and the photovoltaic inversion early warning module. Furthermore, the extreme weather type is used to select the parameter configuration scheme of the wind power output physical response model and the input feature set of the irradiance inversion model. The parameter configuration scheme is determined by the preset parameter configuration table according to the extreme weather type index, and the input feature set is determined by the preset feature template table according to the extreme weather type index.

4. The new energy output response evaluation system based on mechanical inertia under extreme weather conditions according to claim 1, characterized in that, The wind power output physical response model of the wind power ramping early warning module determines the response lag and the upper limit of the achievable ramping rate using the equivalent rotational inertia, and uses the response lag and the upper limit of the achievable ramping rate as constraints for ramping event prediction. The equivalent moment of inertia is determined by the wind turbine and transmission system parameter table and updated according to the operating status of the wind turbine generator set. The operating status includes the pitch angle status, speed status, and grid-connected control mode status. The response hysteresis is determined by the equivalent moment of inertia and speed status through a preset hysteresis mapping function or a preset hysteresis lookup table relationship. The achievable ramp rate upper limit is determined by the equivalent moment of inertia, pitch angle status, and grid-connected control mode status through a preset upper limit mapping function or a preset upper limit lookup table relationship. The achievable ramp rate upper limit is limited to not exceeding the power change rate limit allowed by the generator set control system.

5. The new energy output response evaluation system based on mechanical inertia under extreme weather conditions according to claim 2, characterized in that, The wind power ramping event element indicators include ramping direction, ramping amplitude, maximum ramping rate, and duration. The maximum ramping rate and duration are calculated based on the input data that has been double-aligned within the boundary neighborhood of the ramping event. The maximum ramping rate is the maximum value of the difference ratio of power change to time within the boundary neighborhood, and the duration is the length of time during which the power change continuously maintains the same ramping direction and exceeds a preset amplitude threshold.

6. The new energy output response evaluation system based on mechanical inertia under extreme weather conditions according to claim 3, characterized in that, The photovoltaic inversion early warning module determines the weather model category based on extreme weather types and key meteorological parameter sets, and selects the input feature set and parameter configuration of the irradiance inversion model according to the weather model category; The selection rules for the input feature set are to call the feature template table according to the weather pattern category or to call the segmentation rule table according to the weather pattern category; the photovoltaic inversion early warning module inverts the surface irradiance based on meteorological satellite remote sensing data, and estimates the cloud movement characteristics based on remote sensing time series changes, performs correction processing on the irradiance abrupt change caused by cloud movement, and generates a corrected irradiance sequence.

7. The new energy output response evaluation system based on mechanical inertia under extreme weather conditions according to claim 6, characterized in that, The photovoltaic inversion and early warning module integrates the corrected irradiance sequence with numerical weather forecast data, surface meteorological measurement data, and historical photovoltaic power plant output sequences to obtain photovoltaic output prediction results. The integration modeling selects a subset of integration input features and configures integration weights according to the weather model category. The integration weight configuration is obtained by normalizing the confidence weights or by mapping the confidence weights through a lookup table. Based on the photovoltaic output prediction results, the module outputs photovoltaic ramping event element indicators, which include ramping direction, ramping amplitude, maximum ramping rate, and duration.

8. The new energy output response evaluation system based on mechanical inertia under extreme weather conditions according to claim 1, characterized in that, The multi-scale index classification module includes a basic threshold and an adaptive correction amount. The adaptive correction amount is determined by the margin index of the real-time operation constraints of the power grid through a piecewise function. The margin index includes frequency security margin, rapid reserve margin, and power flow margin of critical sections. The classification includes performing cross-scale consistency verification on the classification results of the station scale, cluster scale, and regional scale, and performing write-back correction on the lower scale classification results based on the higher scale classification results when there is inconsistency.

9. The new energy output response evaluation system based on mechanical inertia under extreme weather conditions according to claim 1, characterized in that, When generating preventative scheduling recommendations, the scheduling recommendation generation module will increase the reserve requirement. As a fast backup constraint, the target quantile level With increased reserve requirements satisfy: ； In the formula: To underestimate the marginal cost of risk; To overestimate the marginal cost of risk; For operational risk ratio; For quantile operators; The equivalent power mismatch prediction value for the evaluation area is constructed based on wind power output prediction results and photovoltaic power output prediction results, combined with grid dispatch plan data. The measured values ​​of equivalent power mismatch in the evaluation area are constructed based on real-time power grid operation data. and The parameters are configured by the market or regulatory strategy table and can be updated at any time; the dispatch suggestion generation module performs a feasibility check of power grid safety operation constraints before outputting preventive dispatch suggestions. The feasibility check includes one or more of the following: reserve constraint check and critical section power flow constraint check; the execution receipt corresponding to the dispatch action suggestion and the real-time power grid operation data after execution are used for updating. The prediction error sample set is used to continuously update and adjust the reserve requirement. .

10. A method for evaluating the output response of new energy sources under extreme weather conditions based on mechanical inertia, applied to the new energy output response evaluation system for extreme weather conditions based on mechanical inertia as described in any one of claims 1-9, characterized in that, Includes the following steps: S1. Collect multi-source data within the evaluation time window of the evaluation area, perform time stamping, time alignment and data verification on the multi-source data, and generate confidence weights for weighted input; S2. Identify extreme weather events based on meteorological elements in multi-source data that have undergone alignment verification, and output the extreme weather type and key meteorological parameter set; S3. Under the constraints of the extreme weather types and key meteorological parameter sets, a physical response model for wind power output is constructed by combining the equivalent rotational inertia of the wind turbine rotor and transmission system to determine the response lag and the upper limit of the achievable ramp rate. Based on the wind power output sequence, wind power ramp event element indicators and wind power output prediction results are output. S4. Surface irradiance is inverted based on meteorological satellite remote sensing data and irradiance inversion model. The inverted irradiance is then fused with numerical weather prediction data, surface meteorological measurement data, and historical photovoltaic power station output sequences to obtain photovoltaic output prediction results. Photovoltaic ramp event element indicators are output. S5. Based on the wind power ramp event element indicators and the photovoltaic ramp event element indicators, quantitative indicators for new energy output ramping at three spatial scales (station, cluster, and region) are constructed and calculated. The ramping impact intensity is classified based on the indicator thresholds, and a classification result including ramping intensity level is output. S6. Based on the classification results and real-time power grid operation data, preventive scheduling suggestions are generated and scheduling action suggestions are output, combined with power grid safety operation constraints. Before outputting, a feasibility check of the power grid safety operation constraints is performed, and the dispatch action recommendations are revised based on the feasibility check results.

Images