# US 2021/0232722 A1 · Method for designing high-rigidity blade based on stochastic isogeometric analysis

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2021/0232722 A1
- Document ID: US20210232722A1
- Status: Published application
- Kind: A1
- Application number: 17/209,260
- Filed: 2021-03-23
- Published: 2021-07-29
- Earliest claimed: 2018-12-27
- Assignees: Zhejiang University
- Applicants: Zhejiang University
- Inventors: Jin CHENG; Zhenyu LIU; Minglong YANG; Jianrong TAN
- Web page: https://patentgrep.com/patent/US20210232722A1

## CPC

- G06F 30/17: Mechanical parametric or variational design
- G06F 17/00: Digital computing or data processing equipment or methods, specially adapted for specific functions (information retrieval, database structures or file system structures therefor G06F16/00)
- G06F 30/28: using fluid dynamics, e.g. using Navier-Stokes equations or computational fluid dynamics [CFD]
- G06N 3/12: using genetic models
- G06F 2111/06: Multi-objective optimisation, e.g. Pareto optimisation using simulated annealing [SA], ant colony algorithms or genetic algorithms [GA]
- G06F 2111/08: Probabilistic or stochastic CAD
- G06F 2119/06: Power analysis or power optimisation
- Y02E 10/30: Energy from the sea, e.g. using wave energy or salinity gradient
- Y02P 70/50: Manufacturing or production processes characterised by the final manufactured product
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- Y02E: REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02P: CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS

## Abstract

In a method for designing a high-rigidity blade based on stochastic isogeometric analysis, the models of stochastic fields of the material property and the external load of the blade are established based on manufacturing conditions and service environment of the blade; an optimization model of the blade is established according to high-rigidity design requirements of the blade and a constraint condition of lift-drag ratio, which is then solved. In the solution procedure, a stochastic isogeometric analysis method is used to calculate the stochastic displacement of the blade under the influence of the randomness of the material property and the external load, and the maximum lift-drag ratio of the blade airfoil is also calculated, based on which the fitness values of individuals in the current population are calculated, so that the high-rigidity design of the blade is realized in the premise of ensuring the lift-drag ratio.

## Claims

### Claim 1

A method for designing a high-rigidity blade based on stochastic isogeometric analysis, wherein the method comprises following steps:
1) parameterizing a blade airfoil, and determining design parameters and ranges of the design parameters according to design rules of an airfoil;
2) describing a material property and an external load of the blade considering spatially dependent uncertainties by using stochastic fields, providing expressions of an objective function and a constraint function for optimizing the blade airfoil according to a design requirement of the high-rigidity blade, and establishing a high-rigidity design model of the blade:
min
x
⁢
f
⁡
(
x
,
r
)
s
.
t
.
⁢
⁢
g
⁡
(
x
)
=
R
L
/
D
⁢
\
⁢
⁢
new
⁡
(
x
)
≥
R
L
/
D
⁢
⁢
\
⁢
ori
;
x
min
≤
x
≤
x
max
where x is a design vector of the blade, which comprises multiple design parameters of the blade; r={E(θ), F(θ)} is a stochastic field vector, two components E(θ) and F(θ) respectively represent stochastic fields of the material property and the external load of the blade with spatially dependent uncertainties, and θ is a set of samples in the stochastic fields; ƒ(x,r)=[μ(U(θ)),σ(U(θ))] is an objective function characterizing a rigidity of the blade, U(θ) is a stochastic normal displacement of the blade under a combined influence of a stochastic material property and a stochastic external load, μ(U(θ)) is an average value of the stochastic normal displacement of the blade, and σ(U(θ)) is a standard deviation of the stochastic normal displacement of the blade; g(x)=RL/D|new(x) is a maximum lift-drag ratio of the blade airfoil, the maximum lift-drag ratio is independent of r, RL/D|ori is a maximum lift-drag ratio of an initial blade airfoil; xmin is a lower limit value of the design vector of the blade, and xmax is an upper limit value of the design vector; and
(3) calculating the stochastic normal displacement of the blade under the combined influence of the stochastic material property and the stochastic external load by using the stochastic isogeometric analysis method, wherein the stochastic normal displacement of the blade characterizes the rigidity of the blade in the high-rigidity design model of the blade; and further calculating an optimal solution to the high-rigidity design model of the blade, to obtain an optimized blade airfoil.

### Claim 2 (depends on 1)

The method for designing the high-rigidity blade based on stochastic isogeometric analysis according to claim 1, wherein in the step (3), the optimal solution of the high-rigidity design model of the blade is solved through a genetic algorithm, by following sub-steps:
3.1) generating an initial population, and setting an evolutional generation as k=0;
3.2) calculating the stochastic normal displacement of the blade corresponding to each individual in a current population, by using the stochastic isogeometric analysis method;
3.3) calculating the maximum lift-drag ratio of the blade airfoil corresponding to each individual in the current population;
3.4) calculating values of the objective function and the constraint function in the high-rigidity design model of the blade according to the stochastic normal displacement of the blade and the maximum lift-drag ratio of the blade airfoil, so as to obtain fitness of all individuals in the current population;
3.5) sorting all individuals in the current population according to the fitness;
3.6) checking whether a maximum evolutional generation is reached; if the maximum evolutional generation is reached, an optimal solution is derived; otherwise, performing genetic operations comprising at least one of selection, crossover and mutation to generate a new population, and setting the evolutional generation as k=k+1, returning to the step 3.2); and
3.7) obtaining design parameter values of an optimal blade airfoil according to the optimal solution.

### Claim 3 (depends on 2)

The method for designing the high-rigidity blade based on stochastic isogeometric analysis according to claim 2, wherein in the step 3.2), the stochastic normal displacement of the blade considering spatially dependent stochastic uncertainties of the material and the external load of the blade is calculated by the stochastic isogeometric analysis method by following steps:
a) establishing a parameterized three-dimensional model of the blade based on T-spline, and setting a boundary condition;
b) discretizing the stochastic fields of the material property and the external load of the blade, by following sub-steps:
b.1) performing an eigen decomposition on a covariance function of the stochastic field, to obtain a second kind Fredholm integral equation;
b.2) expressing eigenfunctions of the covariance function of the stochastic field by T-spline basis functions, to obtain an expression of the covariance function of the stochastic field based on T-spline basis functions;
b.3) solving the second kind Fredholm integral equation by a Galerkin method, to obtain a series of eigenvalues of the covariance function of the stochastic field and values of undetermined coefficients in the expression of b.2);
b.4) obtaining discrete expressions of the stochastic fields of the material property and the external load of the blade by using a Karhunen-Loève expansion; discretizing each stochastic fields into a product of stochastic variables, coefficients and T-spline basis functions;
b.5) determining the number of retained items for Karhunen-Loève expansion according to a practical condition;
b.6) calculating stochastic stiffness matrices and a stiffness matrix at a mean value of the stochastic field of the material property, and calculating stochastic load vectors and a load vector at a mean value of the stochastic field of the external load; and
c) discretizing the stochastic normal displacement of the blade, by following sub-steps:
c.1) constructing a stochastic Krylov subspace of an equilibrium equation of the blade;
c.2) in order to reduce complexity of the subspace, using an inverse matrix of the stiffness matrix at the mean value of the stochastic field of the material property as a complexity reduction factor, multiplying the complexity reduction factor respectively on left and right sides of the equilibrium equation, to realize a preprocessing of the stochastic Krylov subspace;
c.3) calculating a displacement of the blade when the stochastic fields of the material property and the external load of the blade are both at the mean values;
c.4) calculating basis vectors of the stochastic Krylov subspace; and discretizing the stochastic normal displacement of the blade by using the basis vectors, to obtain a displacement expression of the blade;
c.5) determining the number of retained items of a discretization scheme for the basis vectors of the stochastic Krylov subspace according to a practical condition;
c.6) deducing an error expression of the discretization scheme for the basis vectors of the stochastic Krylov subspace;
c.7) minimizing an error by applying a Bubnov-Galerkin condition, to ensure an accuracy of the discretization scheme for the basis vectors of the stochastic Krylov subspace, calculating undetermined coefficient vectors in the expression in c.4), according to a weak or strong form of the Bubnov-Galerkin condition; and
c.8) obtaining the stochastic normal displacement of the blade.

### Claim 4 (depends on 3)

The method for designing the high-rigidity blade based on stochastic isogeometric analysis according to claim 3, wherein a calculation equation for the stochastic normal displacement U(θ) is discretized based on the basis vectors of the stochastic Krylov subspace as follows:
U
⁡
(
θ
)
=
a
1
⁡
(
U
_
+
K
0
-
1
⁢
∑
i
=
1
M
2
⁢
η
i
⁢
F
i
)
+
a
2
⁡
(
K
0
-
1
⁢
∑
i
=
1
M
1
⁢
ξ
i
⁢
K
i
⁢
U
_
+
(
K
0
-
1
)
2
⁢
∑
i
1
=
1
M
1
⁢
∑
i
2
=
1
M
2
⁢
ξ
i
1
⁢
η
i
2
⁢
K
i
2
⁢
U
_
)
+
⁢
⁢
a
3
⁡
(
(
K
0
-
1
)
2
⁢
∑
i
1
⁢
i
2
=
1
M
1
⁢
ξ
i
2
⁢
ξ
i
1
⁢
K
i
2
⁢
K
i
1
⁢
U
_
+
(
K
0
-
1
)
3
⁢
∑
i
1
⁢
i
2
=
1
M
1
⁢
∑
i
3
=
1
M
2
⁢
ξ
i
2
⁢
ξ
i
1
⁢
η
i
3
⁢
K
i
3
⁢
K
i
2
⁢
U
_
)
+
…
+
⁢
⁢
a
m
⁡
(
(
K
0
-
1
)
m
-
1
⁢
∑
i
1
,
…
,
i
m
-
1
=
1
M
1
⁢
ξ
i
1
...
⁢
⁢
ξ
i
m
-
1
⁢
K
i
m
-
1
...
⁢
⁢
K
i
2
⁢
U
_
+
⁢
⁢
(
K
0
-
1
)
m
⁢
∑
i
1
,
…
,
i
m
-
1
=
1
M
1
⁢
∑
i
m
=
1
M
2
⁢
ξ
i
1
⁢
.
.
.
⁢
ξ
i
m
-
1
⁢
η
i
m
⁢
K
i
m
⁢
...
⁢
⁢
K
i
2
⁢
U
_
)
where a1, a2, . . . am is a coefficient vector; Ū is the displacement of the blade in the case that the stochastic fields of the material property and the external load of the blade are both at the mean values; Ē is the mean value of the stochastic field of the material property of the blade; M1 is the number of retained items of the stochastic field of the material property of the blade using the Karhunen-Loève expansion and M2 is the number of retained items of the stochastic field of the external load of the blade using the Karhunen-Loève expansion; K0−1 is a complexity reduction factor; {Ki1, Ki2, . . . Kim} is stochastic stiffness matrices; Fi is a stochastic load vector; m is the number of retained items in the discretization scheme for the basis vectors of the stochastic Krylov subspace; {ξi1, ξi2, . . . ξim} and {ηi1, ηi2, . . . ηim} are respectively uncorrelated Gaussian stochastic variables.
