CN106015102A - Guide vane optimized design method capable of prolonging running down time of reactor coolant pump - Google Patents

Abstract

The invention relates to a guide vane optimized design method capable of prolonging running down time of a reactor coolant pump. According to the guide vane optimized design method, firstly, a main objective function for improving running efficiency under running down by optimizing main parameters of a radial guide vane is established; then the main parameters are divided into three groups of subordinate design variables, subordinate objective functions are optimized respectively with running down efficiency, optimized solution is performed under initial constraint conditions, and the optimum points of the performance of the parameters are obtained; and the optimum points serve as centers for determination of new constraint conditions, optimized solution is performed on the main objective function under the new constraint conditions, and the final optimized result is obtained. By means of the design of the radial guide vane of the reactor coolant pump, the problems of lower efficiency and production of secondary backflow in a running down process can be improved, the running down time is prolonged, sufficient flow passes through to cool a reactor core, and the nuclear safety is enhanced. With application of step-by-step optimization design method adopting a subordinate objective constraint random direction searching method for reduction of a constraint range and a main objective exterior penalty function method for search of the optimum points, the effectiveness of the optimized result is guaranteed, and the optimization speed is increased.

Description

A kind of Optimization Design of the stator extending the core main pump coasting time

Technical field

Involved in the present invention is core main pump idle conditions analysis and stator optimization design field, is specifically related to use the modern times Optimization of Mechanical Design method carries out the method for designing of step-by-step optimization to pump stator parameter.

Background technology

Along with economic is accelerated development, the problem consuming resource and energy scarcity excessively is the most prominent.Nuclear power is as one Technology maturation, safe, economic, the energy of cleaning that can be mass-produced, will have the biggest development in the perspective long-term plan of China Space.Core main pump full name nuclear reactor coolant main circulation pump, its function is that in driving nuclear island, high radioactivity high-temperature high pressure water follows Ring, passes to the heat energy of reactor core nuclear fission steam generator and produces steam, and pushing turbine generates electricity.Core main pump is to ensure that The most critical power-equipment of nuclear plant safety and reliability service, belongs to core work level pump, is unique slewing in nuclear island, is also one One of the pressure boundary in loop.

Core main pump runs the most safely cooling reactor core and prevents the generation of nuclear power plant accident particularly important, Therefore core main pump is often called the heart of nuclear power station.When power-off occurs, owing to inertia will proceed running down, core main pump will be Occur in short time by the transition transient process of pump condition to damped condition, progressively reduce with flow simultaneously.In low discharge work Condition, core main pump fluctuation of service, stator near exit produces secondary back, causes the vibration of pump and efficiency to reduce.One is proposed Optimize core main pump stator under low flow rate condition, and then the measure optimizing its exit flow field seems the most necessary.Therefore design is optimized Time, it is desirable to core main pump running efficiency under the operating mode of power-off running down is sufficiently high, reduces flow sudden change as far as possible, thus ensures that flow fills Foot, extends the coasting time, it is ensured that reactor core safety after station blackout simultaneously.

Through retrieval, patent related to the present invention has: the AP1000 core main pump of a kind of band deviated splitter vane stator (publication number: CN203239626U), relating to the method for designing of the AP1000 core main pump of a kind of band deviated splitter vane stator, its method for designing is length Spacing with blades, deviated splitter vane number is equal, effectively prevents generation and the development of secondary back, improves efficiency simultaneously and guarantee rill Stable operation under Liang, but only it is optimized from stator blade, other running down operation problems may be brought；AP1000 core master Pump radial vane the Hydraulic Design (CN103775387A), is characterized in that the particularity of the spherical spiral case of syncaryon main pump, uses radially The method for designing of stator guide vane, reaches water conservancy diversion and the requirement of energy conversion, the most stable and safety, optimizes its collection simultaneously Close parameter and can meet the requirement of less energy loss, but mostly by Experience Design during its Hydraulic Design, lack certain tight Careful property and comprehensive.

Summary of the invention

For above core main pump stator being optimized the problem that process runs into, the present invention provides a kind of in core main pump power-off operating mode The lower method extending the coasting time and thinking, to optimize running efficiency when the design of stator reaches to improve core main pump running down, increase Flow during strong running down, it is thus achieved that longer coasting time, it is ensured that the nuclear safety of core main pump.

The technical scheme is that

The Optimization Design of a kind of stator extending the core main pump coasting time, step includes the following:

S1: first building one can be by optimizing radial vane major parameter: stator base circle diameter (BCD) D₃, vane inlet width Degree b₃, preliminary election corneriteStator number of blade Z, vane inlet laying angle α₃, stator throat opening area F, improve the master of running down efficiency Object function；

S2: then major parameter is divided into three design of components variablees, sets up three partial objectives for functions, initially retrain in difference It is utilized respectively constraint Random direction research method in condition running down efficiency partial objectives for function is solved, obtains three partial objectives for functions Advantage；

S3: finally determine that centered by the optimum point solved using partial objectives for function a little scope is as new constraint bar Part, solves major heading function optimization by exterior point penalty function method in new constraints, obtains final optimization pass result.

In step S1, described design variable is:Then major heading function is:

In formula, D₃-stator base circle diameter (BCD), unit rice；

b₃-vane inlet width, unit rice；

-preliminary election cornerite, unit degree；

The Z-stator number of blade；

α₃-vane inlet laying angle, unit degree；

F-stator throat opening area, unit square rice.

In step S2, three described design of components variablees are:

X₁=[D₃、b₃]^T,

X₃=[α₅、F]^T,

Three described partial objectives for functions are:

$f\left(X_1\right)=\frac{1}{\mathrm{\η}\left(X_1\right)}=\frac{1}{\mathrm{\η}(D_3,b_3)}\→min;$

$f\left(X_3\right)=\frac{1}{\mathrm{\η}\left(X_3\right)}=\frac{1}{\mathrm{\η}({\mathrm{\α}}_5,F)}\→min.$

In step S2, described initializing constraint solution procedure calculates with initial reference to the following method for designing of pump handbook:

1) stator base circle diameter (BCD) D₃Refer to be cut in the circle of vane inlet helix initial point, improve efficiency and do not affect operation In the case of meet,

$D_3=9.35\·(0.974+0.016\×\frac{n_s}{100})\·{\left(\frac{n_s}{100}\right)}^{-\frac{1}{2}}\·\sqrt[3]{Q/n}+8;$

In formula, D₂-impeller outer diameter, unit rice；

n_s-specific speed；

Q-flow, cubic unit metre per second (m/s)；

N-rotating speed, unit rpm；

2) vane inlet width b₃, namely axial width is satisfied,

$b_3=0.64k_{b2}{\left(\frac{n_s}{100}\right)}^{\frac{5}{6}}\·\sqrt[3]{Q/n}+5;$

In formula, k_b2-impeller outlet width correction factor, with specific speed n_sRelevant；

3) preliminary election corneriteRelevant with runner outer wall streamline, streamline is tangent at outer wall, and expanded view to be ensured is medium-sized Linear light slide-through is suitable, generallyTake

4) stator number of blade Z and throat radius R_c, throat opening area is critical size, in practice it has proved that throat's section is close to square Its effect of shape is best, rule of thumb sums up stator number of blade Z usually: 6≤Z≤8；

5) in order to improve the shape of stator, vane inlet laying angle α₃Should be greater than fluid flow angle α '₃, i.e. α₃=α '₃+ Δ α, Δ α Take 6 °, wherein, vane inlet fluid flow angle

The vane inlet peripheral compoent of velocity

In formula, v_u2-impeller outlet the peripheral compoent of velocity, unit metre per second (m/s)；

D₂-impeller outer diameter, unit rice；

D₃-stator base circle diameter (BCD), unit rice；

Vane inlet axis plane velocity

In formula, b₃-vane inlet width, unit rice；

ψ₃-excretion coefficient, wherein

δ₃-stator vane thickness, unit millimeter；

R₃-stator base radius, unit rice；

6) throat opening area F meets:

In formula, k₃-velocity coeffficient, has certain functional relationship with specific speed；

H-lift, unit rice；

After above-mentioned six calculating complete, the described initializing constraint obtained is:

Partial objectives for function f (X₁) initializing constraint be [0.9D₃,1.1D₃]、[0.9b₃,1.1b₃]；

Partial objectives for function f (X₂) initializing constraint be[6,8]；

Partial objectives for function f (X₃) initializing constraint be [0.9 α₃,1.1α₃]、[0.9F,1.1F]。

In step S2, described partial objectives for function uses constraint Random direction research method respectively under the most initial constraint, Partial objectives for function is optimized and solves, specifically include following steps:

(1) a feasible initial point X is selected⁽⁰⁾, treating excess syndrome tests step-length α⁽⁰⁾, given convergence precision ε₁、ε₂, put number immediately N, and random search direction allows the failed total degree M, general provision M=10 of tentative calculation～20 times, if exceeding this number, and walks When long acquirement is the least, then can shut down；

(2) in (-1,1) are interval, pseudo random number r is produced_i ^(j)(i=1,2 ..., n；J=1,2 ..., N), produce as the following formula N number of n ties up random unitary vector e^(j)(j=1,2 ..., N)

$e^{\left(j\right)}=\frac{1}{{\[\underset{i=1}{\overset{n}{\Σ}}{r_i}^{\left(j\right)}\]}^{\frac{1}{2}}}\left[\begin{array}{c}{r_1}^{\left(j\right)}\\ {r_2}^{\left(j\right)}\\ .\\ .\\ .\\ {r_n}^{\left(j\right)}\end{array}\right](j=1,2,\mathrm{...},N);$

(3) test step-length α is taken₀, it is calculated as follows k point immediately

X^(j)=X⁽⁰⁾+α₀e^(j)(j=1,2 ..., N)

Obviously, N number of random point is distributed in initial point X⁽⁰⁾Centered by, to test step-length α⁽⁰⁾For on the hypersphere of radius；

(4) N number of random point X is checked^(j)(j=1,2 ..., N) whether it is feasible point, remove non-feasible point, calculate remaining The feasible target function value put immediately, compares its size, selects the some X that target function value is minimum^(L)；

(5) X is compared^(L)And X⁽⁰⁾The target function value of 2, if F is (X^(L)) ＜ F (X⁽⁰⁾), then take X^(L)And X⁽⁰⁾Line side To as the feasible direction of search；If F is (X^(L))≥F(X⁽⁰⁾), then by step-length α⁽⁰⁾Reduce, go to step (1) and recalculate, until F (X^(L)) ＜ F (X⁽⁰⁾Till).If α⁽⁰⁾Narrow down to the least (such as α⁽⁰⁾≤10-⁶), still can not find an X^(L), make F (X^(L)) ＜ F(X⁽⁰⁾), then X is described⁽⁰⁾It is a local minimum point, the most replaceable initial point, go to step (1)；

In sum, the condition producing the feasible direction of search is summarised as, and works as X^(L)Point meets

$\left\{\begin{array}{c}{g}_u\left(X^{\left(L\right)}\right)\&GreaterEqual;0(u=1,2,\mathrm{...},m)\\ F\left(X^{\left(L\right)}\right)=\min \left\{F\right(X^{\left(j\right)}\left)\right(j=1,2,\mathrm{...},N\left)\right\}\\ F\left(X^{\left(L\right)}\right)<F\left(X^{\left(0\right)}\right)\end{array}\right.$

Then, the feasible direction of search is S=X₀ ^(L)-X⁽⁰⁾；

(6) from initial point X⁽⁰⁾Setting out, being iterated calculating, until searching one with step-length α along feasible direction of search method S Individual meet whole constraints, and target function value no longer decline newly put X；

(7) if the condition of convergenceIt is met, iteration ends；Its optimal solution is X^*=X, F (X^*)=F (X), otherwise makesGo to step (2).

In step S2, described partial objectives for function solves and obtains optimal solution X₁ ^*=[D₃₍₁₎ ^*、b₃₍₁₎ ^*]^T、X₃ ^*=[α₃₍₃₎ ^*、F₍₃₎ ^*]^T。

In step S3, centered by optimal solution each stator parameter, in little scope, determine described new constraints such as Under: [0.95D₃₍₁₎ ^*,1.05D₃₍₁₎ ^*]、[0.95b₃₍₁₎ ^*,1.05b₃₍₁₎ ^*]、[0.95Z₍₂₎ ^*, 1.05Z₍₂₎ ^*]、[0.95α₃₍₃₎ ^*,1.05α₃₍₃₎ ^*]、

In step S3, described solve in new constraints major heading function use be that SUMT method is applicable to multidimensional more The exterior point penalty function method of constraints, specifically includes following steps:

(1) a suitable initial point X is selected⁽⁰⁾With a r⁽⁰⁾Value, it is stipulated that convergence precision ε₁、ε₂, and determine increase coefficient C, makes k=0；

(2) the optimum minimal point X of penalty is sought with Optimization without restriction^*(r^(k)), i.e.

$min\mathrm{\&phi;}(X,r^{\left(k\right)})=F\left(X\right)+r^{\left(k\right)}\underset{u=1}{\overset{m}{\&Sigma;}}{\{min\&lsqb;0,g_u\left(X\right)\&rsqb;\}}^2+r\left(k\right)\underset{v=1}{\overset{p}{\&Sigma;}}{\&lsqb;h_v\left(X\right)\&rsqb;}^2;$

(3) the inspection condition of convergence.If meeting | | X^*(r^(k))-X^*(r^(k-1))||≤ε₁With

Two inequality, then obtain optimal solution X^*=X^*(r^(k-1)), stop repeatedly Generation；Otherwise turn next step；

(4) orderTurn to (2nd) step.

After major heading function being carried out above-mentioned optimization design under new constraints, the optimal solution solved is carried out rounding After checking, the final optimization pass result of the Optimization Design of a kind of stator extending the core main pump coasting time can be met

The beneficial effects of the present invention is:

(1) Optimization Design of a kind of stator extending the core main pump coasting time that the present invention proposes, to core main pump footpath To manufacturing and designing of stator, there is certain positive effect, it is possible to improve during running down because of stator design unreasonable efficiency on the low side and The problem producing secondary back, extends the coasting time under core main pump station blackout situation simultaneously, it is ensured that have certain during running down Flow by reactor core, enhance the nuclear safety of core main pump.

(2) this method has used partial objectives for constraint Random direction research method to reduce restriction range, and major heading exterior point is sent a letter number Method finds the method for designing of the step-by-step optimization of optimum point, not only ensure that the effectiveness of optimum results, improves optimization simultaneously Speed.

Accompanying drawing explanation

Fig. 1 is constraint Random direction research method flow chart.

Fig. 2 is exterior point penalty function method flow chart.

Detailed description of the invention

The present invention is further described with concrete technical scheme below in conjunction with the accompanying drawings.

The Optimization Design of a kind of stator extending the core main pump coasting time, it comprises structure one can be by optimizing Stator major parameter (stator base circle diameter (BCD) D₃, vane inlet width b₃, preliminary election corneriteStator number of blade Z, vane inlet are pacified Put angle α₃, stator throat opening area F) improve the major heading function of running down efficiency, detailed process is:

With design variable it is:The major heading function built is:

In formula, D₃-stator base circle diameter (BCD), unit rice；

b₃-vane inlet width, unit rice；

-preliminary election cornerite, unit degree；

The Z-stator number of blade；

α₃-vane inlet laying angle, unit degree；

F-stator throat opening area, unit square rice；

Then major variable is divided into three design of components variablees:

X₁=[D₃、b₃]^T,

X₃=[α₅、F]^T,

Three the partial objectives for functions obtained are respectively as follows:

$f\left(X_1\right)=\frac{1}{\mathrm{\&eta;}\left(X_1\right)}=\frac{1}{\mathrm{\&eta;}(D_3,b_3)}\&RightArrow;min;$

$f\left(X_3\right)=\frac{1}{\mathrm{\&eta;}\left(X_3\right)}=\frac{1}{\mathrm{\&eta;}({\mathrm{\&alpha;}}_5,F)}\&RightArrow;min.$

The initial constraint bar must being fulfilled for during determining solving-optimizing in conjunction with relevant design parameter and design experiences Part, as follows in specific implementation process:

1) stator base circle diameter (BCD) D₃Refer to be cut in the circle of vane inlet helix initial point, improve efficiency and do not affect operation In the case of meet,

$D_3=9.35\&CenterDot;(0.974+0.016\&times;\frac{n_s}{100})\&CenterDot;{\left(\frac{n_s}{100}\right)}^{-\frac{1}{2}}\&CenterDot;\sqrt[3]{Q/n}+8;$

In formula, D₂-impeller outer diameter, unit rice；

n_s-specific speed；

Q-flow, cubic unit metre per second (m/s)；

N-rotating speed, unit rpm；

2) vane inlet width b₃, namely axial width is satisfied,

$b_3=0.64k_{b2}{\left(\frac{n_s}{100}\right)}^{\frac{5}{6}}\&CenterDot;\sqrt[3]{Q/n}+5;$

In formula, k_b2-impeller outlet width correction factor, with specific speed n_sRelevant；

3) preliminary election corneriteRelevant with runner outer wall streamline, streamline is tangent at outer wall, and expanded view to be ensured is medium-sized Linear light slide-through is suitable, generallyTake

4) stator number of blade Z and throat radius R_c, throat opening area is critical size, in practice it has proved that throat's section is close to square Its effect of shape is best, and rule of thumb summing up the usual value of stator number of blade Z is 6≤Z≤8；

5) in order to improve the shape of stator, vane inlet laying angle α₃Should be greater than fluid flow angle α '₃, i.e. α₃=α '₃+ Δ α, Δ α Take 6 °, wherein, vane inlet fluid flow angleThe vane inlet peripheral compoent of velocity

In formula, v_u2-impeller outlet the peripheral compoent of velocity, unit metre per second (m/s)；

D₂-impeller outer diameter, unit rice；

D₃-stator base circle diameter (BCD), unit rice；

Vane inlet axis plane velocity

In formula, b₃-vane inlet width, unit rice；

ψ₃-excretion coefficient, wherein

δ₃-stator vane thickness, unit millimeter；

R₃-stator base radius, unit rice；

6) throat opening area F meets:

In formula, k₃-velocity coeffficient, has certain functional relationship with specific speed；

H-lift, unit rice；

After above-mentioned six calculating complete, the new initializing constraint obtained is:

Partial objectives for function f (X₁) initializing constraint be [0.9D₃,1.1D₃]、[0.9b₃,1.1b₃]；

Partial objectives for function f (X₂) initializing constraint be[6,8]；

Partial objectives for function f (X₃) initializing constraint be [0.9 α₃,1.1α₃]、[0.9F,1.1F]。

Partial objectives for function uses in initializing constraint constraint Random direction research method solve, and retrains random The direction search method it is crucial that how to determine initial point, the direction of search and step-size in search.

As shown in Figure 1 for constraint Random direction research method flow chart, its concrete steps can carry out table by procedure below Show:

(1) a feasible initial point X is selected⁽⁰⁾, treating excess syndrome tests step-length α⁽⁰⁾, given convergence precision ε₁、ε₂, put number immediately N, and random search direction allows the failed total degree M, general provision M=10 of tentative calculation～20 times, if exceeding this number, and walks When long acquirement is the least, then can shut down.

(2) in (-1,1) are interval, pseudo random number r is produced_i ^(j)(i=1,2 ..., n；J=1,2 ..., N), produce as the following formula N number of n ties up random unitary vector e^(j)(j=1,2 ..., N)

$e^{\left(j\right)}=\frac{1}{{\&lsqb;\underset{i=1}{\overset{n}{\&Sigma;}}{r_i}^{\left(j\right)}\&rsqb;}^{\frac{1}{2}}}\left[\begin{array}{c}{r_1}^{\left(j\right)}\\ {r_2}^{\left(j\right)}\\ .\\ .\\ .\\ {r_n}^{\left(j\right)}\end{array}\right](j=1,2,\mathrm{...},N);$

(3) test step-length α is taken₀, it is calculated as follows k point immediately

X^(j)=X⁽⁰⁾+α₀e^(j)(j=1,2 ..., N)

Obviously, N number of random point is distributed in initial point X⁽⁰⁾Centered by, to test step-length α⁽⁰⁾For on the hypersphere of radius.

(4) N number of random point X is checked^(j)(j=1,2 ..., N) whether it is feasible point, remove non-feasible point, calculate remaining The feasible target function value put immediately, compares its size, selects the some X that target function value is minimum^(L)。

(5) X is compared^(L)And X⁽⁰⁾The target function value of 2, if F is (X^(L)) ＜ F (X⁽⁰⁾), then take X^(L)And X⁽⁰⁾Line side To as the feasible direction of search；If F is (X^(L))≥F(X⁽⁰⁾), then by step-length α⁽⁰⁾Reduce, go to step (1) and recalculate, until F (X^(L)) ＜ F (X⁽⁰⁾Till).If α⁽⁰⁾Narrow down to the least (such as α⁽⁰⁾≤10^-6), still can not find an X^(L), make F (X^(L)) ＜ F (X⁽⁰⁾), then X is described⁽⁰⁾It is a local minimum point, the most replaceable initial point, go to step (1).

In sum, the condition producing the feasible direction of search is summarised as, and works as X^(L)Point meets

$\left\{\begin{array}{c}{g}_u\left(X^{\left(L\right)}\right)\&GreaterEqual;0(u=1,2,\mathrm{...},m)\\ F\left(X^{\left(L\right)}\right)=\min \left\{F\right(X^{\left(j\right)}\left)\right(j=1,2,\mathrm{...},N\left)\right\}\\ F\left(X^{\left(L\right)}\right)<F\left(X^{\left(0\right)}\right)\end{array}\right.$

Then, the feasible direction of search is S=X₀ ^(L)-X⁽⁰⁾。

(6) from initial point X⁽⁰⁾Setting out, being iterated calculating, until searching one with step-length α along feasible direction of search method S Individual meet whole constraints, and target function value no longer decline newly put X.

(7) if the condition of convergenceIt is met, iteration ends.Its optimal solution is X^*=X, F (X^*)=F (X), otherwise makesGo to step (2).

Partial objectives for function solves and obtains optimal solution X₁ ^*=[D₃₍₁₎ ^*、b₃₍₁₎ ^*]^T、X₃ ^*= [α₃₍₃₎ ^*、F₍₃₎ ^*]^TAfter, we are centered by each parameter of optimal solution, determine the new constraint of major heading function in little scope Condition is as follows:

Then using exterior point penalty function method to solve major heading function under new constraints, wherein exterior point penalizes letter Number methods it is crucial that the structure of exterior point penalty function, constitution step is as follows:

For g_u≤ 0, u=1,2, (X) ..., the form of m is taken as

In formula, I_iFor violating the set of constraints, i.e. I_i=u | g_u(X) ＞ 0, u=1,2 ..., m}

When feasible zone is outer, then g_u(X) ＞ 0；When in feasible zone, then g_u(X)≤0。

Being exterior point penalty function method flow chart as shown in Figure 2, its process below step represents:

(1) a suitable initial point X is selected⁽⁰⁾With a r⁽⁰⁾Value, it is stipulated that convergence precision ε₁、ε₂, and determine increase coefficient C, makes k=0.

(2) the optimum minimal point X of penalty is sought with Optimization without restriction^*(r^(k)), i.e.

$min\mathrm{\&phi;}(X,r^{\left(k\right)})=F\left(X\right)+r^{\left(k\right)}\underset{u=1}{\overset{m}{\&Sigma;}}{\{min\&lsqb;0,g_u\left(X\right)\&rsqb;\}}^2+r\left(k\right)\underset{v=1}{\overset{p}{\&Sigma;}}{\&lsqb;h_v\left(X\right)\&rsqb;}^2;$

(3) the inspection condition of convergence.If meeting | | X^*(r^(k))-X^*(r^(k-1))||≤ε₁With Two inequality, then obtain optimal solution X^*=X^*(r^(k-1)), stop iteration；Otherwise turn next step.

(4) orderTurn to (2nd) step.

After major heading function being carried out above-mentioned optimization design under new constraints, the optimal solution solved is carried out rounding After checking, the final optimization pass result of the Optimization Design of a kind of stator extending the core main pump coasting time can be met

The invention is not restricted to above-described embodiment, also comprise other embodiments and variation in the range of present inventive concept.

Claims (8)

1. the Optimization Design of the stator extending the core main pump coasting time, it is characterised in that step includes the following:

S1: first building one can be by optimizing radial vane major parameter: stator base circle diameter (BCD) D₃, vane inlet width b₃、 Preliminary election corneriteStator number of blade Z, vane inlet laying angle α₃, stator throat opening area F, improve the major heading of running down efficiency Function；

S2: then major parameter is divided into three design of components variablees, sets up three partial objectives for functions, at different initializing constraints Inside it is utilized respectively constraint Random direction research method running down efficiency partial objectives for function is solved, obtains three partial objectives for Function Optimization Point；

S3: finally determine centered by the optimum point solved using partial objectives for function a little scope as new constraints, By exterior point penalty function method, major heading function optimization is solved in new constraints, obtain final optimization pass result.

The Optimization Design of a kind of stator extending the core main pump coasting time the most according to claim 1, its feature exists In, in step S1, described design variable is:Then major heading function is:

In formula, D₃-stator base circle diameter (BCD), unit rice；

b₃-vane inlet width, unit rice；

-preliminary election cornerite, unit degree；

The Z-stator number of blade；

α₃-vane inlet laying angle, unit degree；

F-stator throat opening area, unit square rice.

The Optimization Design of a kind of stator extending the core main pump coasting time the most according to claim 1, its feature exists In, in step S2, three described design of components variablees are:

X₁=[D₃、b₃]^T,

X₃=[α₅、F]^T,

Three described partial objectives for functions are:

$f\left(X_1\right)=\frac{1}{\mathrm{\&eta;}\left(X_1\right)}=\frac{1}{\mathrm{\&eta;}(D_3,b_3)}\&RightArrow;min;$

$f\left(X_3\right)=\frac{1}{\mathrm{\&eta;}\left(X_3\right)}=\frac{1}{\mathrm{\&eta;}({\mathrm{\&alpha;}}_5,F)}\&RightArrow;min.$

The Optimization Design of a kind of stator extending the core main pump coasting time the most according to claim 1, its feature exists In, in step S2, described initializing constraint is:

Partial objectives for function f (X₁) initializing constraint be [0.9D₃,1.1D₃]、[0.9b₃,1.1b₃]；

Partial objectives for function f (X₂) initializing constraint be

Partial objectives for function f (X₃) initializing constraint be [0.9 α₃,1.1α₃]、[0.9F,1.1F]。

The Optimization Design of a kind of stator extending the core main pump coasting time the most according to claim 4, its feature exists In, in step S2, described initializing constraint solution procedure calculates with initial reference to the following method for designing of pump handbook:

1) stator base circle diameter (BCD) D₃Refer to be cut in the circle of vane inlet helix initial point, improve efficiency and do not affect the situation of operation It is lower satisfied,

$D_3=9.35\&CenterDot;(0.974+0.016\&times;\frac{n_s}{100})\&CenterDot;{\left(\frac{n_s}{100}\right)}^{-\frac{1}{2}}\&CenterDot;\sqrt[3]{Q/n}+8;$

In formula, D₂-impeller outer diameter, unit rice；

n_s-specific speed；

Q-flow, cubic unit metre per second (m/s)；

N-rotating speed, unit rpm；

2) vane inlet width b₃, namely axial width is satisfied,

$b_3=0.64k_{b2}{\left(\frac{n_s}{100}\right)}^{\frac{5}{6}}\&CenterDot;\sqrt[3]{Q/n}+5;$

In formula, k_b2-impeller outlet width correction factor, with specific speed n_sRelevant；

3) preliminary election corneriteRelevant with runner outer wall streamline, streamline is tangent at outer wall, the medium-sized linear light of expanded view to be ensured Slide-through is suitable, generallyTake

4) stator number of blade Z and throat radius R_c, throat opening area is critical size, in practice it has proved that throat's section close to square its Effect is best, rule of thumb sums up stator number of blade Z usually: 6≤Z≤8；

5) in order to improve the shape of stator, vane inlet laying angle α₃Should be greater than fluid flow angle α '₃, i.e. α₃=α '₃+ Δ α, Δ α take 6 °, Wherein, vane inlet fluid flow angle

The vane inlet peripheral compoent of velocity

In formula, v_u2-impeller outlet the peripheral compoent of velocity, unit metre per second (m/s)；

D₂-impeller outer diameter, unit rice；

D₃-stator base circle diameter (BCD), unit rice；

Vane inlet axis plane velocity

In formula, b₃-vane inlet width, unit rice；

ψ₃-excretion coefficient, wherein

δ₃-stator vane thickness, unit millimeter；

R₃-stator base radius, unit rice；

6) throat opening area F meets:

k₃-velocity coeffficient, has certain functional relationship with specific speed；

H-lift, unit rice.

The Optimization Design of a kind of stator extending the core main pump coasting time the most according to claim 1, its feature exists In, in step S2, described partial objectives for function uses constraint Random direction research method respectively under the most initial constraint, to subhead Scalar functions is optimized and solves, and specifically includes following steps:

(1) a feasible initial point X is selected⁽⁰⁾, treating excess syndrome tests step-length α⁽⁰⁾, given convergence precision ε₁、ε₂, some number N immediately, with And random search direction allows the failed total degree M, general provision M=10 of tentative calculation～20 times, if exceeding this number, and step-length takes When obtaining the least, then can shut down；

(2) in (-1,1) are interval, pseudo random number r is produced_i ^(j)(i=1,2 ..., n；J=1,2 ..., N), produce N number of n as the following formula Dimension random unitary vector e^(j)(j=1,2 ..., N)

$e^{\left(j\right)}=\frac{1}{{\&lsqb;\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{r_i}^{\left(j\right)}\&rsqb;}^{\frac{1}{2}}}\left[\begin{array}{c}{r_1}^{\left(j\right)}\\ {r_2}^{\left(j\right)}\\ \begin{array}{c}.\\ .\\ .\end{array}\\ {r_n}^{\left(j\right)}\end{array}\right],(j=1,2,...,N);$

(3) test step-length α is taken₀, it is calculated as follows k point immediately

X^(j)=X⁽⁰⁾+α₀e^(j)(j=1,2 ..., N)

Obviously, N number of random point is distributed in initial point X⁽⁰⁾Centered by, to test step-length α⁽⁰⁾For on the hypersphere of radius；

(4) N number of random point X is checked^(j)(j=1,2 ..., N) whether it is feasible point, remove non-feasible point, calculate remaining feasible The target function value put immediately, compares its size, selects the some X that target function value is minimum^(L)；

(5) X is compared^(L)And X⁽⁰⁾The target function value of 2, if F is (X^(L)) ＜ F (X⁽⁰⁾), then take X^(L)And X⁽⁰⁾Line direction make For the feasible direction of search；If F is (X^(L))≥F(X⁽⁰⁾), then by step-length α⁽⁰⁾Reduce, go to step (1) and recalculate, until F (X^(L)) ＜ F (X⁽⁰⁾Till).If α⁽⁰⁾Narrow down to the least (such as α⁽⁰⁾≤10^-6), still can not find an X^(L), make F (X^(L)) ＜ F (X⁽⁰⁾), then X is described⁽⁰⁾It is a local minimum point, the most replaceable initial point, go to step (1)；

In sum, the condition producing the feasible direction of search is summarised as, and works as X^(L)Point meets

$\left\{\begin{array}{c}{g}_u\left(X^{\left(L\right)}\right)\&GreaterEqual;0,(u=1,2,...,m)\\ F\left(X^{\left(L\right)}\right)=\min \{F\left(X^{\left(j\right)}\right),(j=1,2,...,N)\}\\ F\left(X^{\left(L\right)}\right)<F\left(X^{\left(0\right)}\right)\end{array}\right.$

Then, the feasible direction of search is S=X₀ ^(L)-X⁽⁰⁾；

(6) from initial point X⁽⁰⁾Set out, be iterated calculating with step-length α along feasible direction of search method S, until searching one completely What the most all constraints, and target function value no longer declined newly puts X；

(7) if the condition of convergenceIt is met, iteration ends；Its optimal solution is X^*=X, F (X^*) =F (X), otherwise makesGo to step (2)；

Described partial objectives for function solves and obtains optimal solution X₁ ^*=[D₃₍₁₎ ^*、b₃₍₁₎ ^*]^T、X₃ ^*= [α₃₍₃₎ ^*、F₍₃₎ ^*]^T。

The Optimization Design of a kind of stator extending the core main pump coasting time the most according to claim 1, its feature exists In, step

In rapid S3, centered by optimal solution each stator parameter, in little scope, determine that described new constraints is as follows:

The Optimization Design of a kind of stator extending the core main pump coasting time the most according to claim 1, its feature exists In, in step S3, described solve in new constraints major heading function use be that SUMT method is applicable to Multi-dimensional constraint more The exterior point penalty function method of condition, specifically includes following steps:

(1) a suitable initial point X is selected⁽⁰⁾With a r⁽⁰⁾Value, it is stipulated that convergence precision ε₁、ε₂, and determine increase coefficient c, order K=0；

(2) the optimum minimal point X of penalty is sought with Optimization without restriction^*(r^(k)), i.e.

$\min \mathrm{\&phi;}(X,r^{\left(k\right)})=F\left(X\right)+r^{\left(k\right)}\underset{u=1}{\overset{m}{\&Sigma;}}{\{min\&lsqb;0,g_u\left(X\right)\&rsqb;\}}^2+r\left(k\right)\underset{v=1}{\overset{p}{\&Sigma;}}{\&lsqb;h_v\left(X\right)\&rsqb;}^2;$

(3) the inspection condition of convergence.If meeting | | X^*(r^(k))-X^*(r^(k-1))||≤ε₁With Two inequality, then obtain optimal solution X^*=X^*(r^(k-1)), stop iteration；Otherwise turn next step；

(4) orderTurn to (2nd) step；

After major heading function being carried out above-mentioned optimization design under new constraints, the optimal solution solved is carried out rounding checking After, the final optimization pass result of the Optimization Design of a kind of stator extending the core main pump coasting time can be met