CN-122021134-A - Wind power tower crane lightweight design method and system based on digital twin and multi-objective optimization

CN122021134ACN 122021134 ACN122021134 ACN 122021134ACN-122021134-A

Abstract

The invention relates to the technical field of computer aided engineering and structural optimization, and discloses a light-weight design method of a wind power generator based on digital twin and multi-objective optimization. By defining the multi-objective optimization problem with the minimum structural quality and the minimum equivalent stress of key parts as cores and solving by adopting a non-dominant sorting mechanism, a series of pareto optimal design schemes which achieve optimal balance between the light weight degree and the structural strength can be automatically generated, so that the unilaterality possibly caused by the traditional single-objective optimization is avoided, and a designer can make decisions in a more comprehensive balance space, so that the stress level is effectively controlled while the structural weight is obviously lightened, and the overall safety and reliability of the tower crane are ensured.

Inventors

HU TING
QU YONGLEI
ZHAO WEIXING
HU BIN
ZHANG JIAN
ZHANG JINGYONG
ZHAO JIAN
MAO TONGTONG
LI WEI
SONG ZIYAO

Assignees

河南金利重工科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260112

Claims (10)

1. The light-weight design method of the wind power generator based on digital twin and multi-objective optimization is characterized by comprising the following steps of: s1, constructing a full-parameterized initial three-dimensional geometric model of a wind power tower crane, and establishing a corresponding initial finite element model in a finite element simulation platform; s2, based on real-time sensor data and historical operation data of a physical entity wind turbine, constructing a multi-dimensional dynamic digital twin model of the wind turbine, wherein the digital twin model is used for mapping and predicting structural response of the wind turbine under various working conditions; S3, defining a multi-objective optimization problem on an initial finite element model based on a prediction result of the digital twin model, wherein the multi-objective optimization problem comprises minimization of the overall structure quality of the tower crane and minimization of the maximum equivalent stress of a key part, and determining design variables and constraint conditions; S4, calling a multi-objective optimization solver, carrying out multi-round iterative adjustment and evaluation on the structural parameters of the initial finite element model under the limitation of design variables and constraint conditions, wherein the multi-objective optimization solver carries out evaluation and screening of candidate schemes based on a non-dominant sorting mechanism, and after each round of iteration, defining a currently obtained candidate scheme set which is mutually not dominant as a pareto front; S5, selecting one or more preferred solutions from the pareto front edge, updating the full-parameterized initial three-dimensional geometric model according to structural parameters corresponding to the preferred solutions, and generating a three-dimensional structural model of the wind power tower crane after light weight design; In the step S2, the multi-dimensional dynamic digital twin model dynamically corrects the local stiffness matrix and the damping coefficient in the initial finite element model by performing online data assimilation on the real-time sensor data and the initial finite element model, so that the errors of the simulation output of the model and the actual measured data of the sensor are converged to be within a preset threshold.
2. The method for designing a lightweight wind power tower machine based on digital twin and multi-objective optimization according to claim 1, wherein in the step S1, the geometric parameters on which the full-parameterized initial three-dimensional geometric model is constructed include the diameter, wall thickness, start-stop position and dimension of variable cross section, position and dimension of internal platform, structural dimension of connecting flange, and interface dimension of nacelle and impeller of each segment of tower barrel of the wind power tower machine, and the initial finite element model is established by performing meshing operation, giving material attribute parameters, defining connection relation, and applying boundary load condition.
3. The wind power tower crane lightweight design method based on digital twin and multi-objective optimization according to claim 1 is characterized in that in the step S2, real-time sensor data comprise stress data, strain data, acceleration data, vibration spectrum data and environment wind speed and direction data which are deployed at different heights of a tower drum, historical operation data comprise structural response time sequence data recorded under different wind speed grades and corresponding load spectrums, the multi-dimensional dynamic digital twin model is constructed through an online data assimilation technology, specifically comprises the steps that real-time sensor data are used as observation vectors, simulation output of an initial finite element model is used as state vectors, optimal estimation of the state vectors is calculated through iteration of a Kalman filtering algorithm, and local stiffness matrix elements Kij and Rayleigh damping coefficients alpha and beta in the finite element model are reversely corrected according to the optimal estimation until root mean square error of the simulation output and measured data is smaller than 5%.
4. The method for designing the wind power tower crane in the light weight mode based on the digital twin and multi-objective optimization according to claim 1 is characterized in that in the step S3, design variables are modifiable size parameter sets in a full-parameterized initial three-dimensional geometric model, specifically include wall thickness values of all segments of a tower barrel, longitudinal transition lengths of all variable-section cone segments, thickness and width of flanges connecting adjacent tower barrel segments and distribution circumference diameters of flange connecting bolts, constraint conditions are limit state limits set based on design specifications and safety requirements, specifically include maximum horizontal displacement limit values of the top end of the tower barrel under rated wind load, absolute values of differences between first-order natural frequencies of the integral structure of the tower barrel and wind wheel excitation frequencies are not smaller than 10% of wind wheel excitation frequencies, and limit values of ratio values of maximum equivalent stress and material yield strength of key areas such as the root of the tower barrel and the variable sections under limit working conditions.
5. The method for designing the lightweight of the wind power tower crane based on digital twin and multi-objective optimization according to claim 1, wherein in the step S4, the multi-objective optimization solver generates a offspring population by using Latin hypercube sampling in the initial population, then generates new candidate schemes of each subsequent generation through standard genetic operation of NSGA-II algorithm, selects parent population according to non-dominant sorting and crowding distance, applies analog binary crossover (SBX) operators (crossover distribution index eta c =20, crossover probability p c =0.9) and Polynomial Mutation (PM) operators (mutation distribution index eta m =20, mutation probability p m =1/n, n is the number of design variables), generates offspring population by using finite element analysis on each candidate design scheme to obtain the corresponding total structure quality and key maximum equivalent stress, compares the superiority of all candidate schemes on two targets according to non-dominant sorting mechanism, and keeps the superiority of the candidate schemes on at least one target and the other candidate schemes not to meet the preset condition until the iteration is finished.
6. The method for designing the wind power tower crane in light weight based on digital twin and multi-objective optimization according to claim 1, wherein in the step S5, the preferred solution is selected from the pareto front in such a way that all solutions on the pareto front are visually displayed according to the corresponding total structural mass and the maximum equivalent stress value, and the final design scheme is manually specified or automatically recommended by the system according to the preset engineering preference decision rule or in combination with the manufacturing process cost factor.
7. The method for designing the lightweight of the wind power tower crane based on digital twin and multi-objective optimization according to claim 1, further comprising the step S6 of inputting structural parameters of a lightweight design model into an updated multi-dimensional dynamic digital twin model, simulating a long-term operation process of the model under a standard wind load spectrum, evaluating the fatigue damage accumulation degree of key parts of the model, and generating a report document comprising structural performance prediction and service life evaluation.
8. A digital twin and multi-objective optimization based wind turbine lightweight design system for implementing the method of any of claims 1-7, comprising: the parameterized modeling and finite element preprocessing module is used for constructing a full parameterized three-dimensional geometric model of the wind power tower crane and automatically generating a corresponding finite element analysis model based on the model; The dynamic digital twin model management module is used for accessing the sensor data stream and the history database of the physical tower crane, constructing and operating a digital twin model updated synchronously with the physical entity, and outputting structural response prediction data; The multi-objective optimization problem configuration module is connected with the dynamic digital twin model management module and is used for setting multi-objective, design variables and constraint conditions of the lightweight design according to the prediction result of the twin model; The optimization solving and iteration control module is connected with the multi-objective optimization problem configuration module and the parameterized modeling and finite element preprocessing module and is used for driving the optimization solver to execute multi-round iteration adjustment, finite element analysis and scheme screening on the structural parameters of the finite element model until the pareto front is generated; The design decision and post-processing module is connected with the optimization solving and iteration control module and is used for displaying the pareto front edge, assisting in selecting the preferred solution and outputting a final lightweight design three-dimensional model and an engineering drawing.
9. The digital twinning and multi-objective optimization-based wind turbine lightweight design system of claim 8, wherein the dynamic digital twinning model management module comprises: The data acquisition and access unit is used for acquiring multi-source heterogeneous data from the sensor network and the database in real time; The data fusion and working condition identification unit is used for carrying out time synchronization and abnormal cleaning on the acquired data and identifying the current and historical typical operation working conditions; The twin model core unit is integrated with a parameterized finite element model and a data-driven response surface model, and dynamically corrects model parameters by fusing real-time data so as to output high-fidelity structural response prediction; The data-driven response surface model is constructed by adopting a Gaussian Process Regression (GPR), is input into the environment wind speed, the wind direction and the tower bottom load, and is output into the tower root stress, the response surface model is used as a proxy model, time-consuming finite element analysis is replaced in optimization iteration to perform quick evaluation, when an optimization solver needs to evaluate a new design scheme, the response surface model is firstly called to obtain a stress prediction value, if the scheme enters the pareto front, the complete finite element model is called to perform accurate verification, and a verification result is added into a training set to update the response surface model on line.
10. The digital twin and multi-objective optimization-based wind turbine lightweight design system according to claim 8, further comprising a virtual verification and report generation module, wherein the virtual verification and report generation module receives the lightweight design model output by the design decision and post-processing module, drives the updated digital twin model to perform long-term time course analysis, and outputs a comprehensive verification report containing fatigue life cloud graphics, dynamic response comparison and lightweight effect quantization indexes.

Description

Wind power tower crane lightweight design method and system based on digital twin and multi-objective optimization Technical Field The invention relates to the technical field of computer aided engineering and structural optimization, in particular to a light-weight design method and system of a wind power tower crane based on digital twin and multi-objective optimization. Background With the continuous increase of global clean energy demand, wind power generation is widely applied as a mature renewable energy technology, a wind power generator is used as a core bearing structure for supporting the wind power generator, the design of the wind power generator directly relates to the safety, reliability and economy of the whole wind power generator, the design method of the traditional wind power generator mainly depends on static force and fatigue analysis based on an empirical formula, and is verified and optimized by combining finite element simulation, and the method is generally based on standardized load working conditions and material characteristics, and the basic structural safety is ensured, but the following limitations often exist in the design process: Firstly, the traditional design method generally adopts single-objective optimization (such as minimizing quality or minimizing cost), and is difficult to realize effective balance among multiple objectives of light weight, structural safety, dynamic performance and the like, and excessive pursuit of light weight can lead to the aggravation of stress concentration at key parts, the shortening of fatigue life and even the induction of structural resonance risk; Secondly, the traditional simulation model is mostly based on idealized theoretical assumption and standard load spectrum, and cannot fully reflect complex, dynamic and time-varying load effects of the wind power tower crane in an actual running environment, and structural response characteristics caused by the complex, dynamic and time-varying load effects, wherein influences of factors such as actual wind conditions, environmental factors (such as temperature and humidity), structural degradation and the like are difficult to accurately embody in the traditional model, so that deviation exists between a design result and a real working condition, and potential safety hazards can be introduced or design redundancy can be caused. Disclosure of Invention The invention aims to provide a light-weight design method and system for a wind power tower crane based on digital twin and multi-objective optimization, so as to solve the problems in the background technology. In order to achieve the purpose, the invention provides the following technical scheme that the wind power tower crane lightweight design method based on digital twin and multi-objective optimization comprises the following steps: s1, constructing a full-parameterized initial three-dimensional geometric model of a wind power tower crane, and establishing a corresponding initial finite element model in a finite element simulation platform; s2, based on real-time sensor data and historical operation data of a physical entity wind turbine, constructing a multi-dimensional dynamic digital twin model of the wind turbine, wherein the digital twin model is used for mapping and predicting structural response of the wind turbine under various working conditions; S3, defining a multi-objective optimization problem on an initial finite element model based on a prediction result of the digital twin model, wherein the multi-objective optimization problem comprises minimization of the overall structure quality of the tower crane and minimization of the maximum equivalent stress of a key part, and determining design variables and constraint conditions; S4, calling a multi-objective optimization solver, carrying out multi-round iterative adjustment and evaluation on the structural parameters of the initial finite element model under the limitation of design variables and constraint conditions, wherein the multi-objective optimization solver carries out evaluation and screening of candidate schemes based on a non-dominant sorting mechanism, and after each round of iteration, defining a currently obtained candidate scheme set which is mutually not dominant as a pareto front; S5, selecting one or more preferred solutions from the pareto front edge, updating the full-parameterized initial three-dimensional geometric model according to structural parameters corresponding to the preferred solutions, and generating a three-dimensional structural model of the wind power tower crane after light weight design; In the step S2, the multi-dimensional dynamic digital twin model dynamically corrects the local stiffness matrix and the damping coefficient in the initial finite element model by performing online data assimilation on the real-time sensor data and the initial finite element model, so that the errors of the simulation output of the model and the actual m