CN-121981838-A - Operation decision optimization and insurance charging method under wind farm multi-disaster risk

CN121981838ACN 121981838 ACN121981838 ACN 121981838ACN-121981838-A

Abstract

The invention discloses an operation decision optimization and insurance charging method under the multi-disaster risk of a wind farm, in particular to the technical field of new energy risk management and intelligent operation and maintenance, which comprises the steps of constructing a multi-disaster space-time coupled physical field model and a critical equipment vulnerability curved surface, mapping real-time prediction output of a physical field evolution model as dynamic load to the equipment vulnerability curved surface, calculating the damage probability of a unit, combining geographic coordinates and electric topology to perform spatial interpolation and topology aggregation, generating a unit-level dynamic risk map, identifying a risk unit, splitting and reconstructing a power transmission channel, and performing power self-adaptive redistribution on an online unit according to the reconstruction capacity and the predicted wind speed to form an operation strategy; and aggregating the unit-level damage probability into a comprehensive risk exposure index, fitting the expected loss cost by combining the historical power generation efficiency baseline, dynamically adjusting the basic insurance rate, and generating a differential insurance charging factor linked with an operation strategy.

Inventors

LI DONG
YAN ZIHAN
MA JIAN
LI HAOMING
HUANG TAO
GUO JIAANG
LIAO YIZHEN
WU KEDA
Wu Fanxi

Assignees

福州大学
中国地震局工程力学研究所

Dates

Publication Date: 20260505
Application Date: 20260407

Claims (9)

1. The operation decision optimization and insurance charging method under the risk of multiple disasters of the wind farm is characterized by comprising the following steps: S1, acquiring historical meteorological data and geographic information data of a wind power plant area, taking a disaster causing factor as an input variable, establishing a disaster physical field space-time evolution model based on a partial differential equation set, and constructing a device vulnerability curved surface based on material characteristics and structural dynamics response of a wind turbine generator and a power transmission line; S2, mapping real-time prediction output of the disaster physical field space-time evolution model as dynamic load to an equipment vulnerability curved surface, calculating damage probability of a unit and a line in a future preset time window, and performing spatial interpolation and topology aggregation on the damage probability according to geographic coordinates and electrical connection topology of each unit to generate a dynamic risk map; S3, reading the dynamic risk map, identifying risk units with damage probabilities exceeding a preset tolerance threshold, splitting the risk units according to the electric connection topology, re-planning power output paths of the rest units, generating a power transmission channel reconstruction scheme, avoiding a high risk area, and performing power generation power self-adaptive redistribution on each online unit according to the reconstructed power transmission capacity and the predicted wind speed to form an operation strategy for avoiding multiple disaster risks; And S4, aggregating the unit-level damage probability into a comprehensive risk exposure index of the wind power plant, fitting out expected loss cost considering the dynamic influence of disasters by combining with a historical power generation efficiency baseline of the wind power plant, dynamically adjusting a preset basic insurance rate, generating a differential insurance charging factor, and linking with the operation strategy in real time.
2. The method for optimizing operation decisions and charging insurance under multiple disaster risks of a wind farm according to claim 1 is characterized in that the disaster physical field time-space evolution model is constructed based on a Navier-Stokes equation set with a geographic feature correction term and a device structure feedback term introduced, the geographic feature correction term is obtained by fitting surface roughness parameters with a terrain elevation gradient, and the device structure feedback term is obtained by fitting real-time vibration, strain and load data of a wind turbine generator and a power transmission line.
3. The method for optimizing operation decision and charging insurance under multiple disaster risks of a wind farm according to claim 1, wherein the equipment vulnerability curved surface is a three-dimensional curved surface with disaster factor intensity of a first dimension, equipment operation age of a second dimension and damage probability of a third dimension, the curved surface is generated based on equipment damage sample library by fitting through a support vector machine, and material performance attenuation coefficients obtained through quarter detection are dynamically updated.
4. The method for optimizing operation decision and charging insurance under multiple disaster risks of a wind farm according to claim 1, wherein the calculation mode of the damage probability is that dynamic load output by the physical field space-time evolution model is decomposed into direct load and indirect load, and the direct load and the indirect load are weighted and fused through coupling coefficients fitted by dynamic response data of a device structure to obtain comprehensive dynamic load, and then the comprehensive dynamic load is mapped to the vulnerable curved surface of the device.
5. The method for optimizing operation decision and charging insurance under the risk of multiple disasters of a wind farm according to claim 1 is characterized in that the topological aggregation is realized by adopting a graph neural network algorithm, specifically, the damage probability of a unit on the same power collecting line is weighted and aggregated according to the unit capacity ratio to obtain the damage probability of the power collecting line, and then the damage probabilities of the power collecting line, the main transformer and the sending-out line are aggregated in series to obtain the damage probability of a power transmission channel.
6. The method for optimizing operation decisions and charging insurance under multiple disaster risks of a wind farm according to claim 1 is characterized in that the risk unit adopts a fixed threshold and dynamic threshold collaborative recognition algorithm, the dynamic threshold is adjusted in real time by an LSTM prediction model according to the current disaster evolution trend, the tolerance threshold is lowered when the disaster intensity is aggravated, and the tolerance threshold is raised when the disaster intensity is weakened.
7. The method for optimizing operation decision and charging insurance under multiple disaster risks of a wind farm according to claim 1, wherein the self-adaptive redistribution of the generated power is realized based on deep Q network agents, a state space comprises predicted wind speed, reconstructed transmission capacity and real-time running state of units, and a reward function aims at maximizing total yield of the wind farm and minimizing equipment damage probability, and power adjustment values of the units are output.
8. The method for optimizing operation decisions and charging insurance under multiple disaster risks of a wind farm according to claim 1, wherein the comprehensive risk exposure index is calculated through a weighted aggregation algorithm, and the weight distribution is set to be set damage probability weight 0.4, power transmission channel damage probability weight 0.3 and global disaster intensity weight 0.3 calculated based on the ratio of the maximum value of disaster causing factors predicted by a physical farm to the historical extreme value.
9. The method for optimizing operation decision and charging insurance under multiple disaster risks of a wind farm according to claim 1, wherein the real-time linkage is characterized in that when the comprehensive risk exposure index of the wind farm is increased, the operation strategy automatically strengthens risk prevention and control measures to inhibit further increase of the comprehensive risk exposure index of the wind farm, and when the operation strategy effectively reduces the comprehensive risk exposure index of the wind farm, the differentiated insurance charging factors are synchronously adjusted downwards to form closed loop feedback of risk prevention and control and insurance cost.

Description

Operation decision optimization and insurance charging method under wind farm multi-disaster risk Technical Field The invention relates to the technical field of new energy risk management and intelligent operation and maintenance, in particular to an operation decision optimization and insurance charging method under the risk of multiple disasters of a wind farm. Background The wind power plant is usually located in mountain areas, coasts or gobi areas and other areas far away from the load center, and is subjected to various natural disasters such as typhoons, storm winds, freezing and thunderstorms for a long time, and the space-time superposition and coupling effects of the multiple disasters can cause equipment damages such as cracking of blades of the wind power plant, deformation of towers, broken lines of power transmission lines and the like, and can also cause power grid fluctuation and power generation loss, so that the operation safety and economic benefits of the wind power plant are seriously influenced. The implementation process of the prior art comprises the steps of acquiring forecast data of extreme weather such as typhoons, high winds and the like through early warning information of an access area weather station, triggering full-field or regional shutdown according to a forecast threshold value, manually inspecting equipment conditions after a disaster passes, recovering operation, and estimating loss and paying according to the shutdown time during the disaster, an equipment damage list recorded by manual inspection and historical average generating capacity data in a insurance claim settlement link, wherein insurance charging mainly depends on post-nuclear loss and static annual rate. However, when the method is actually used, the method still has some defects, such as single disaster risk assessment dimension, only dependence on external weather early warning, lack of fine modeling on the micro topography of a wind power plant, equipment structure response and multi-disaster coupling effect, incapability of accurately positioning the differential risk of each unit, excessively passive and extensive operation strategies, huge power generation benefit loss caused by unified shutdown of the whole plant, incapability of dynamically reconstructing a power transmission channel and power distribution according to real-time risks, lag and rigidification of an insurance charging mechanism, long post-nuclear damage period and more disputes, incapability of reflecting real-time risk differences under different disaster scenes by annual fixed rates, and difficulty in realizing collaborative optimization of risk prevention and control and insurance cost. Disclosure of Invention In order to overcome the defects in the prior art, the embodiment of the invention provides an operation decision optimization and insurance charging method under the risk of multiple disasters of a wind farm, and solves the problems in the background art through the following scheme. The method comprises the following steps of S1, acquiring historical meteorological data and geographic information data of a wind power plant area, taking disaster factors as input variables, establishing a disaster physical field space-time evolution model based on partial differential equation sets, and constructing a device vulnerability curved surface based on material characteristics and structural dynamics response of a wind turbine generator and a power transmission line; S2, mapping real-time prediction output of the physical field space-time evolution model as dynamic load to the vulnerable curved surface of the equipment, calculating damage probability of the units and the circuits in a future preset time window, and performing spatial interpolation and topology aggregation on the damage probability according to geographic coordinates and electrical connection topology of each unit to generate a dynamic risk map; S3, reading the dynamic risk map, identifying risk units with damage probability exceeding a preset tolerance threshold, splitting the risk units according to the electric connection topology, re-planning power output paths of the rest units, generating a power transmission channel reconstruction scheme, avoiding a high risk area, and performing power generation power self-adaptive redistribution on each online unit according to the reconstructed power transmission capacity and the predicted wind speed to form an operation strategy for avoiding multiple disaster risks; And S4, aggregating the unit-level damage probability into a comprehensive risk exposure index of the wind power plant, fitting out expected loss cost considering the dynamic influence of disasters by combining with a historical power generation efficiency baseline of the wind power plant, dynamically adjusting a preset basic insurance rate, generating a differential insurance charging factor, and linking with the operation strategy in real time. The invent