CN-121997733-A - Wind turbine generator system component life probability prediction and wind power plant energy control method and system

CN121997733ACN 121997733 ACN121997733 ACN 121997733ACN-121997733-A

Abstract

The invention discloses a method and a system for predicting the service life probability of a wind turbine component, and a method and a system for controlling wind power plant energy, wherein the predicting method comprises the steps of constructing a state space model of a degradation process of the wind turbine component; the method comprises the steps of obtaining real-time state information of a component, inputting the real-time state information into a state space model, evaluating the degradation state of the component by adopting a sequential data assimilation algorithm, outputting posterior probability distribution of the degradation state, and based on the posterior probability distribution, adopting Monte Carlo simulation to simulate a future degradation path of the component forwards until the future degradation path exceeds a preset failure threshold value, so as to obtain residual life probability distribution of the component. According to the method, the personalized residual life of each unit can be accurately estimated, and an innovative wind power plant energy management control strategy is created based on the personalized residual life of each unit, so that the absolute maximization of the economic benefit of the whole life cycle of the wind power plant in a complex market environment is realized.

Inventors

HU XUESONG
ZHENG LIWEN
DU CHUN

Assignees

国电联合动力技术有限公司

Dates

Publication Date: 20260508
Application Date: 20260121

Claims (10)

1. The method for predicting the service life probability of the wind turbine generator component is characterized by comprising the following steps of: Constructing a state space model of a degradation process of a component of a wind turbine, wherein the state space model comprises a state equation and an observation equation, the state equation is an evolution rule of a degradation state variable under the driving of an operation load, and the observation equation is a relation between the degradation state variable and an observable health index; Acquiring real-time state information of the component; Inputting the real-time state information into the state space model, evaluating the degradation state of the component by adopting a sequential data assimilation algorithm, and outputting posterior probability distribution of the degradation state; Based on the posterior probability distribution, adopting Monte Carlo simulation to forward simulate a future degradation path of the component until the future degradation path exceeds a preset failure threshold value, and obtaining the residual life probability distribution of the component.
2. The method for predicting the lifetime probability of a wind turbine component according to claim 1, wherein the evaluating the degradation state of the component using a sequential data assimilation algorithm comprises: According to the state equation, carrying out state transition prediction on the initial particle set to obtain prior probability distribution of the degradation state; if a new observation value is received, calculating the weight of each predicted particle according to the observation equation; and normalizing the weights to obtain the posterior probability distribution.
3. The method for predicting the lifetime probability of a wind turbine component according to claim 2, wherein the normalizing the weights to obtain the posterior probability distribution further comprises: resampling the normalized particle set according to the normalized weight to obtain a target particle set, and taking the target particle set as an initial particle set of the next time step.
4. A wind turbine component life probability prediction system, comprising: The system comprises a model construction unit, a model analysis unit and a model analysis unit, wherein the model construction unit is used for constructing a state space model of a degradation process of a component of a wind turbine, the state space model comprises a state equation and an observation equation, the state equation is an evolution rule of a degradation state variable under the driving of an operation load, and the observation equation is a relation between the degradation state variable and an observable health index; an acquisition unit for acquiring real-time status information of the component; A posterior probability distribution output unit for inputting the real-time state information into the state space model, evaluating the degradation state of the component by adopting a sequential data assimilation algorithm, and outputting the posterior probability distribution of the degradation state; And the residual life probability distribution obtaining unit is used for adopting Monte Carlo simulation to forward simulate a future degradation path of the component based on the posterior probability distribution until the future degradation path exceeds a preset failure threshold value so as to obtain the residual life probability distribution of the component.
5. The wind turbine component lifetime probability prediction system of claim 4, wherein the posterior probability distribution output unit is further configured to: According to the state equation, carrying out state transition prediction on the initial particle set to obtain prior probability distribution of the degradation state; if a new observation value is received, calculating the weight of each predicted particle according to the observation equation; and normalizing the weights to obtain the posterior probability distribution.
6. The wind turbine component life probability prediction system of claim 5, further comprising: And the target particle set obtaining unit is used for resampling the normalized particle set according to the normalized weight to obtain a target particle set, and taking the target particle set as an initial particle set of the next time step.
7. A method of controlling energy in a wind farm, comprising: acquiring the residual life probability distribution and the real-time electricity price of the parts of each wind turbine by adopting the wind turbine part life probability prediction method as claimed in any one of claims 1-3; Constructing a multi-objective optimization function by taking maximization of total expected benefits of a full life cycle of a wind power plant as a core objective, wherein the total expected benefits of the full life cycle are inversely proportional to expected failure risk costs, and the expected failure risk costs are directly proportional to real-time failure probabilities calculated by the residual life probability distribution of the components; Solving the multi-objective optimization function to generate an optimization control instruction corresponding to each wind turbine, wherein the optimization control instruction comprises an optimal power set value and a driving operation strategy based on the real-time electricity price; And issuing the optimization control instruction.
8. The method for controlling energy of a wind farm according to claim 7, wherein the solving the multi-objective optimization function to generate the optimization control command corresponding to each wind turbine includes: solving the multi-objective optimization function to obtain a real-time health degree weight factor of each wind turbine, wherein the real-time health degree weight factor is inversely proportional to the failure risk of the wind turbine; And generating the optimal power set value according to the health degree weight factor.
9. The method for controlling energy of a wind farm according to claim 7, wherein the real-time electricity price includes a high electricity price and a low electricity price, and the solving the multi-objective optimization function generates an optimization control command corresponding to each wind turbine, and the method includes: controlling the healthy wind turbine generator to fully emit or super emit in the high electricity price period; and controlling the high-risk wind turbine generator to reduce the load in the low-electricity-price period.
10. A wind farm energy control system, comprising: An obtaining unit, configured to obtain a part remaining life probability distribution and a real-time electricity price of each wind turbine predicted by using the wind turbine part life probability prediction method according to any one of claims 1 to 3; The target optimization function construction unit is used for constructing a multi-target optimization function by taking the maximization of the total expected benefit of the whole life cycle of the wind power plant as a core target, wherein the total expected benefit of the whole life cycle is inversely proportional to the expected failure risk cost, and the expected failure risk cost is directly proportional to the real-time failure probability calculated by the residual life probability distribution of the component; the optimizing control instruction generating unit is used for solving the multi-objective optimizing function and generating an optimizing control instruction corresponding to each wind turbine generator, and the optimizing control instruction comprises an optimal power set value and a driving operation strategy based on the real-time electricity price; And the instruction issuing unit is used for issuing the optimized control instruction.

Description

Wind turbine generator system component life probability prediction and wind power plant energy control method and system Technical Field The invention relates to the technical field of wind turbines, in particular to a method and a system for predicting service life probability of wind turbine components and a method and a system for controlling energy of a wind farm. Background With the advancement of global energy structure transformation and power marketing reform, wind farm operation targets have been shifted from pursuing maximization of generated energy to pursuing maximization of generated energy gain, and the ability to respond to market price fluctuations in real time is required. In this new situation, wind farm operation is faced with an unprecedented challenge. On the one hand, although the wind generating set is a product with standardized design, once the wind generating set is installed in a specific wind farm, the microscopic topography and wind resource conditions of each machine site are obviously different. This immobility of the installation site and non-uniformity of the environmental resources causes a critical problem in that the actual load history and degradation rate of the critical components of even the same model of unit are greatly different. In general, the machine set at the machine site with good wind resources has large output and high generating income, but the mechanical stress born by the parts is larger, the service life decay is quicker, and on the contrary, the machine set at the machine site with poor wind resources has small output and slow decay. This "same-factory-like different-life" phenomenon makes traditional "one-tool" operation and maintenance strategies based on uniform time intervals or average operating hours uneconomical and reasonable. On the other hand, the volatility of the power market requires a wind farm with great flexibility. Ideally, during high electricity rates, the wind farm should add equipment loss to "full power" and even allow the well-conditioned units to "properly overdriven" within the safety margin to capture high value power, while during low electricity rates or zero/negative electricity rates, the operating strategy should be turned to "maintenance equipment" to actively reduce the load to delay component degradation, and the equipment' life loss "is used on the blade. However, implementing such fine-grained, dynamic policy adjustments is not possible by means of conventional manual decisions and simple control logic. Currently, the prior art has the following major limitations in wind farm operation: 1. component state prediction lacks individualization and probabilization: the traditional alarm system based on the fixed threshold value can not cope with individual differences of the same factory with different orders of the wind generating set, and accurate early warning is difficult to carry out. The existing data-driven diagnostic model is difficult to quantify the prediction uncertainty, and the component physical model cannot be effectively fused, so that the personalized residual life (REMAINING USEFUL LIFE, RUL) caused by different running loads cannot be accurately estimated, and the prediction result is difficult to directly use for economic decisions which need to balance short-term power generation benefits with long-term equipment health. The lack of a dynamic update degradation track tracking mechanism has limited prediction accuracy. 2. Wind farm energy management control strategy rigidifies: The current mainstream power distribution strategy completely isolates market price signals and equipment health states, cannot automatically identify high price window periods and excite healthy units to generate more power, and cannot intelligently reduce load in low price periods for realizing service life equalization. The lack of an optimization model for uniformly quantifying electricity price, equipment residual life probability prediction, power generation income and potential fault loss makes it difficult to automatically generate and execute an optimal control instruction according to a preset economic target, highly relies on manual experience to judge, has slow response and is difficult to optimize. In summary, the prior art has core defects that state prediction fails to quantify uncertainty and cannot reflect unit individual difference, insufficient fusion of a physical model and real-time data, stiff control strategy and the like. Disclosure of Invention The invention aims to overcome the defects of the prior art, and provides a method and a system for predicting the service life probability of a wind turbine generator set component, a method and a system for controlling wind farm energy, wherein the method can accurately evaluate the personalized residual service life of each wind turbine generator set, and an innovative wind farm energy management control strategy is created based on the method, so that the abs