CN104657551A - Plate-fin heat exchanger core structure optimization method based on dynamic pixel granularity - Google Patents

Abstract

The invention discloses a plate-fin heat exchanger core structure optimization method based on dynamic pixel granularity. The method comprises the following steps: establishing a plate-fin heat exchanger core structure optimization design model according to a plate-fin heat exchanger core runner structure, providing dynamically-updated pixel granularity, enlarging the population search range and keeping the population diversity; and providing a pixel distance calculation model of non-head and tail particles and head and tail particles, calculating the cross and mutation operation probabilities in a self-adaption manner according to the pixel distances of the particles, and respectively adopting random cross and Gaussian mutation, so as to enhance the global population search capability, improve the local population search efficiency, prevent the algorithm from getting into local optimum and realize the purposes of wide coverage and uniform distribution of Pareto optimal solutions. According to the method, the heat exchanger core structure design efficiency can be improved, and relatively reasonable design parameters are provided. The plate-fin heat exchanger optimally designed by the method has the obvious characteristics of uniform passage load, small secondary heat transfer temperature difference, small flow resistance and high heat exchange efficiency.

Description

A kind of fin heat exchanger core structural optimization method based on dynamic pixel granularity

Technical field

Fin heat exchanger core structural optimization method of the present invention, especially relates to a kind of fin heat exchanger core structural optimization method based on dynamic pixel granularity.

Background technology

Plate type finned heat exchanger is compared to traditional heat exchangers such as tubular heat exchangers, have that heat transfer efficiency is high, temperature difference controlling is good, compact conformation, cost performance are high, scalable, reliability high, employing soldering manufactures, improve the compressive resistance of heat exchange, can be used for the heat exchange between a plurality of fluids heat exchange, medium, in fields such as air separation equipment, Aero-Space, petrochemical complex, atomic energy and national defense industry, there is widespread use.Plate type finned heat exchanger take fin as heat exchange unit, and heat transfer coefficient and heat transfer area are all better than tubular heat exchanger.When transmitting identical heat, plate type finned heat exchanger is because fin thickness is little, compact conformation, therefore its weight ratio shell-and-tube heat exchanger is light.Due to these advantages, plate type finned heat exchanger background technology of the present invention is:

Plate type finned heat exchanger complex structure is the equipment realizing the exchange heat such as condensation, liquefaction, evaporation, has little temperature difference unsteady heat transfer, secondary heat transfer, allows the distinguishing feature that resistance is little, the change of multiple flow physical property is fierce.Plate type finned heat exchanger is made up of fin, dividing plate, strip of paper used for sealing, end socket and flow deflector, and its structural core is plate bundle, comprises multiplely being put into the passage coordinating strip of paper used for sealing to form again between two dividing plates (composite plate) by fin, flow deflector.Fin is placed between composite plate, and fixes with strip of paper used for sealing, and core body is soldering in a vacuum furnace, two ends welding end socket.The core body of plate type finned heat exchanger forms by the soldering of multiple cold and hot fluid passage is stacked, traditional core structure of heat exchanger method for designing is method of trial and error, namely first heat-transfer surface, heat eliminating medium and the type of flow is selected, then repeatedly suppose that physical dimension carries out tentative calculation, until obtain the heat interchanger that meets institute's Prescribed Properties.Along with the increase of exchanger heat flow rate, existing structure method for designing is difficult to solve the problems such as heat exchanger core body passage load is uneven, heat exchange efficiency declines, structural design difficulty.

In recent years along with the widespread use of intelligent algorithm, scholar has been had intelligent algorithm to be applied in design of heat exchanger.University of Alabama of U.S. NAJAFI etc. adopt genetic algorithm to have studied the impact of plate-fin heat exchanger fin structure heat exchanger performance.University of Hamburg, Germany Federal Defence Forces ROETZEL etc. have studied fluid non-uniformity problem in the heat interchanger considering hyperbolic dispersion model, when calculating Non-Uniform Flow, and Axial Temperature Distribution in shell-and-tube heat exchanger.France process engineering laboratory, ripple city RENEAUME etc. have studied plate-fin heat exchanger optimization method, provides the continuous type formula solving plate-fin heat exchanger performance.Existing heat exchanger structure Optimization Design, exists that local optimal searching ability is strong and ability of searching optimum is poor, easily occur the shortcomings such as Premature convergence, and design accuracy is difficult to satisfy the demands.

Core body is the core of plate type finned heat exchanger and crucial heat exchanging part, accounts for the weight and volume of the heat interchanger overwhelming majority; Diabatic process mainly relies on fin to complete, and simultaneously fin the flowing of convection cell can produce resistance again, so the type of fin and size are also the principal elements affecting heat exchanger performance, therefore emphasis of the present invention is optimized design for core body and fin structure.

Summary of the invention

In order to overcome the deficiency in background technology, the object of the present invention is to provide a kind of fin heat exchanger core structural optimization method based on dynamic pixel granularity.The method, on the basis of basic particle group algorithm, dynamically updating pixel granularity by introducing, calculating crossover and mutation evolutionary operator probability adaptively, set up modified particle swarm optiziation.This algorithm application is minimum in the fin heat exchanger core Optimal Structure Designing of target with the general assembly (TW) of core body in solving, draw the optimum solution of parameter of structure design.

The step of the technical solution used in the present invention is as follows:

1) main performance requirements of plate type finned heat exchanger and the physical parameter of fluid that need optimization is determined;

2) determine core optimized variable and constraint condition thereof, namely determine the solution space of Structure Optimization Variables vector X phenotype and problem; One group of optimized variable of core body is expressed as follows:

X＝{x ₁,x ₂,x ₃,…,x _k}

In formula, x _irepresent an optimized amount in optimized variable vector, k represents optimized variable sum;

3) according to the 2nd) optimized variable that obtains of step and constraint condition thereof sets up Optimized model, determines the type of objective function and mathematical description form thereof or quantization method, namely final optimum solution; Set up Optimized model formula to be described below:

Solve: f (x ₁, x ₂, x ₃..., x _k)

Target: minf (x ₁, x ₂, x ₃..., x _k)

Constraint: g (x ₁, x ₂, x ₃..., x _k)≤0

h(x ₁,x ₂,x ₃,…,x _k)＝0

$x_i^{\min }\≤x_i\≤x_i^{\max }$

In formula, x _irepresent an optimized amount in optimized variable vector, k represents optimized variable sum, with refer to the minimum of the corresponding optimized amount in optimized variable vector and maximum possible value respectively, f () represents the objective function of optimization problem, and g () represents inequality constrain, and h () represents equality constraint.

4) after setting up Optimized model, adopt the particle in population to represent the feasible solution of optimized variable vector, population scale, iteration algebraically, pixel granularity and outside Pareto pond initiation parameter are set, and initialization is carried out to the position of all particles and speed;

5) position of Population Regeneration particle and speed;

6) calculate the fitness value of particle, judge dominance relation, upgrade inner domination and separate and non-domination solution set;

7) calculate crossover and mutation evolutionary operator probability adaptively, take random the intersection and Gaussian mutation operation respectively, the more position of new particle, upgrades inner non-domination solution set;

8) judge the dominance relation that inner non-domination solution and outside Pareto separate, and upgrade outside Pareto pond;

9) employing dynamically updates pixel granularity, calculates the location of pixels of particle in outside Pareto pond, and rejects the unnecessary particle of same location of pixels;

10) calculate the pixel distance of particle in outside Pareto pond, choosing maximum pixel distance particle is population global optimum particle;

11) whether evaluation algorithm meets end condition, if so, then terminates to calculate, obtains optimum solution, otherwise enter the 5th) step.

Described 7th), in step, calculate crossover and mutation evolutionary operator probability adaptively and refer to: the crossover and mutation evolutionary operator probability of particle carrys out Dynamic Acquisition according to the pixel distance of particle.

Described 9th) adopt in step and dynamically update pixel granularity and refer to: dynamically update according to iterations.

The beneficial effect that the present invention has is:

The present invention sets up fin heat exchanger core Optimal Structure Designing model, proposes to dynamically update pixel granularity, expands population hunting zone, keeps population diversity; Calculate crossover and mutation evolutionary operator probability adaptively, take random the intersection and Gaussian mutation respectively, strengthen ability of searching optimum, improve Local Search efficiency, avoid algorithm to be absorbed in local optimum, realize the target that Pareto optimum solution covers extensively, is evenly distributed.The present invention can improve core structure of heat exchanger design efficiency, provides more reasonably design parameter.The present invention can improve core structure of heat exchanger design efficiency, provides more reasonably design parameter.Plate type finned heat exchanger after optimal design of the present invention, has the distinguishing feature that passage load is even, secondary heat transfer temperature difference is little, resistance to flow is little, heat exchange efficiency is high.

Accompanying drawing explanation

Fig. 1 is the fin heat exchanger core structure optimization overall procedure based on dynamic pixel granularity of the present invention.

Fig. 2 is heat exchanger core body global configuration parameter schematic diagram of the present invention.

Fig. 3 is heat exchanger core body flow passage structure parameter schematic diagram of the present invention.

Fig. 4 optimizes idiographic flow based on the fin heat exchanger core of dynamic pixel granularity.

Fig. 5 is that the pixel distance of non-head and the tail particle of the present invention calculates schematic diagram.

In Fig. 5: P (k+1), P (k-1)---sort location of pixels that is rear and particle k adjacent particles;

| p _i(k+1)-p _i(k-1) |---at objective function f _ithe pixel distance of dimension; N---objective function sum.

Fig. 6 is that the pixel distance of head and the tail particle of the present invention calculates schematic diagram.

In Fig. 6: Δ P (s)---the pixel distance of the first particle after sequence; P (s+1)---with the first particle adjacent particles; Δ P (e)---the pixel distance of end particle; P (e-1)---with end particle adjacent particles; L---influence coefficient.

Embodiment

Below in conjunction with drawings and Examples, the present invention is described in further detail.

1.1 problems describe

There are two kinds of designing requirements for design of heat exchanger: a kind of is under the efficiency and resistance of satisfied setting, reduce heat interchanger physical dimension and weight as much as possible.Another is given size of heat exchanger and weight demands, makes heat exchanger efficiency high as far as possible.The present invention, selects Optimal Parameters for target with core body loss of weight, and the fin heat exchanger core structure optimization overall procedure based on dynamic pixel granularity of the present invention as shown in Figure 1.

Plate type finned heat exchanger is made up of plate bundle, end socket, adapter and bearing etc., is core body by the part of arranging plate Shu Zucheng more.The effects such as core body, as the core of heat interchanger and most important heat exchanger components, accounts for about 80% of heat interchanger general assembly (TW), and remainder all plays connection as annex, close.As shown in Figure 2, major parameter has hot side liquid length of flow L to certain heat exchanger core body general structure ₁, cold size fluid flows length L ₂, non-current direction size L ₃, fin number of plies N.As shown in Figure 3, major parameter has fin thickness δ to certain heat exchanger core body flow passage structure _f, fin pitch X, block board thickness δ _f, clapboard spacing s.When carrying out design of heat exchanger, can take core weight as objective function, be obtained the general assembly (TW) of heat interchanger by transformation of coefficient.

Diabatic process mainly relies on fin to complete, and simultaneously fin the flowing of convection cell can produce resistance again, so the type of fin and size are also the principal elements affecting heat exchanger performance, and also should as optimized variable.

Optimized variable vector X is expressed as follows:

X＝{x ₁,x ₂,x ₃,…,x _k}

X _i---represent one of core size or fin size

K---represent optimized variable project sum

1.2 set up Optimized model

Core body design optimization problem can be described below with formula:

For i＝1,2,3,…,n

Find:X＝{x ₁,x ₂,x ₃,…,x _k}

Minimize:f(x)＝f(X)

Subject to:Δp ₁≤Δp _1max

Δp ₂≤Δp _2max

η ₁≥η _min

η ₂≥η _min

g(X)≥0

$x_i^{\min }\≤x_i\≤x_i^{\max }$

Δ p ₁---the pressure drop of the hot side of display plate fin heat exchanger

Δ p _1max---the maximum allowed pressure drop of the hot side of display plate fin heat exchanger

Δ p ₂---the pressure drop of display plate fin heat exchanger cold side

Δ p _2max---the maximum allowed pressure drop of display plate fin heat exchanger cold side

η ₁---the heat exchange efficiency of the hot side of display plate fin heat exchanger

η ₂---the heat exchange efficiency of display plate fin heat exchanger cold side

η _min---the minimum permission heat transfer efficiency of display plate fin heat exchanger

G (X)---the strength check of display plate fin heat exchanger

---represent the minimum value of the respective items of optimized variable vector

---represent the maximal value of the respective items of optimized variable vector

1.3 determine fitness function

In order to embody the adaptive faculty of particle, introduce the function can measured each particle in problem, i.e. fitness function.Decided excellent, the bad degree of particle by fitness function, it embodies the survival of the fittest principle in natural evolution.For optimization problem, fitness function is exactly objective function.

F (x)---fitness value

G (x)---the fitness value under max problem

1.4 location of pixels determining particle in outside Pareto pond

Fin heat exchanger core based on dynamic pixel granularity of the present invention optimizes idiographic flow, as shown in Figure 4.For the object space S that n objective function forms _n, according to the pixel granularity G=(g in objective function dimension ₁..., g _i..., g _n), the objective function dimension of correspondence is divided into M=(m ₁..., m _i..., m _n) individual block of pixels, these block of pixels coverage goal space S _n, each block of pixels location of pixels P=(p ₁..., p _i..., p _n) mark it at object space S _nin position, wherein 0≤p _i≤ m _i, g _irepresent objective function f _ipixel granularity in dimension; m _irepresent objective function f _ithe number that the difference of the maxima and minima of dimension is divided equally.In order to determine that in outside Pareto pond, particle is at object space S _nmiddle mated block of pixels, needs the location of pixels P calculating particle.Particle is at the location of pixels p of each objective function dimension _icomputing formula as follows, and take the mode rounded up by p _iconsolidation.

$p_i=\frac{|f_i^k\left(X_i\right)-f_{i\min }^k\left(X\right)|}{|f_{i\max }^k\left(X\right)-f_{i\min }^k\left(X\right)|}*m_i=\frac{|f_i^k\left(X_i\right)-f_{i\min }^k\left(X\right)|}{g_i}$

In formula ---particle X in the outside Pareto pond that kth generation obtains _iat objective function f _ion value

---represent that in the outside Pareto pond that kth generation obtains, particle is at objective function f _ion minimum value

---represent that in the outside Pareto pond that kth generation obtains, particle is at objective function f _ion maximal value

M _i---in kth generation, is by objective function dimension f _ithe number divided equally,

G _i---kth is for objective function f _ithe pixel granularity of dimension

Block of pixels in each object space at most only belongs to particle in an outside Pareto pond, various in order to expand population hunting zone and increase population, therefore in the process of iteration, and the pixel granularity g in same dimension _idynamically update in real time.Pixel granularity g _i(t) to dynamically update formula as follows:

$g_i\left(t\right)=g_i*e^{\frac{t}{t_{\max }}-1}$

G in formula _i---ensure the pixel granularity of objective function precision

T---current iteration algebraically

T _max---greatest iteration algebraically

1.5 adaptive crossover and mutation operations

Separate for contemporary population inside domination, the present invention calculates the crossing-over rate of pairing adaptively according to the pixel distance between pairing.Adaptive crossover and mutation operation of the present invention, as shown in Figure 4.The pairing crossing-over rate that pixel distance is little is little, and the large pairing crossing-over rate of pixel distance is large.By interlace operation, strengthen diversity and the ability of searching optimum of population.Such as pairing particle X _j, X _klocation of pixels P (j), P (k), pairing crossing-over rate is P _jkc, computing formula is:

$P_{\mathrm{jkc}}=(P_{c\max }-P_{c\min })*{(1+\exp (-H_c*\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}|p_i\left(j\right)-p_i\left(k\right)|}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\left|\max \left({\mathrm{\Δp}}_i\right)\right|}))}^{-1}+P_{c\min }$

P in formula _cmax---crossing-over rate maximal value

P _cmin---crossing-over rate minimum value

H _c---crossing-over rate regulation coefficient

---particle X _jwith X _kpixel distance

---the maximum pixel distance in pairing is for the interlace operation of pairing particle, and the present invention will adopt random method of intersecting, and obtains the new particle after intersecting.

For contemporary non-domination solution, the present invention calculates the aberration rate of particle adaptively according to the pixel distance of particle.Pixel distance is larger, shows that the Local Search to this particle strengthened by needs.The large individual variation rate of pixel distance is large, and the little individual variation rate of pixel distance is little.By mutation operation, strengthen population local search ability.Such as, in inner Pareto pond particle X _jaberration rate be P _jm, computing formula is:

$P_{\mathrm{jm}}=(P_{m\max }-P_{m\min })*{(1+\exp (-H_m*\frac{\mathrm{\ΔP}\left(j\right)}{\max \left(\mathrm{\ΔP}\right)}))}^{-1}+P_{m\min }$

P in formula _mmax---aberration rate maximal value

P _mmin---aberration rate minimum value

H _m---aberration rate regulation coefficient

Δ P (j)---particle X _jpixel distance

Max (Δ P)---the maximum pixel distance of inner Pareto pond particle

For the mutation operation of particle, the present invention adopts the method for Gaussian mutation.Such as, in inner Pareto pond particle X _jthe new particle after Gaussian mutation is taked to be X' _j.

if P _jm＞P _m

X' _j＝X _j(1+0.5*N(0,1))

P in formula _m---Population Variation rate

N (0,1)---obeying expectation is 0, and variance is the Gaussian distribution of 1

1.6 determine population globally optimal solution

The pixel distance of non-head and the tail particle of the present invention calculates schematic diagram, as shown in Figure 5.The pixel distance of head and the tail particle of the present invention calculates schematic diagram, as shown in Figure 6.Location of pixels set after particle consolidation in outside Pareto pond is carried out ascending order arrangement by the target function value of Stochastic choice.Because the pixel granularity of each particle on each objective function is identical, so only the location of pixels difference calculated between adjacent particles just can determine the distance between adjacent particles.Pixel distance Δ P (k) computing formula such as between regular rear particle k and adjacent particles is as follows:

$\mathrm{\ΔP}\left(k\right)=P(k+1)-P(k-1)=\underset{i=1}{\overset{n}{\mathrm{\Σ}}}|p_i(k+1)-p_i(k-1)|$

P (k+1), P (k-1) in formula---sort location of pixels that is rear and particle k adjacent particles

| p _i(k+1)-p _i(k-1) |---at objective function f _ithe pixel distance of dimension

N---objective function sum

Above-mentioned formula does not cover the first and last particle after sequence, although particle cluster algorithm adopts the optimizing mode of random paralleling, continues to a certain extent in the optimizing of first and last particle periphery.In order to ensure more effectively to determine population globally optimal solution, preventing from causing population Pareto disaggregation coverage to reduce because giving up first and last particle, namely preventing " precocity " phenomenon, therefore needing the pixel distance reasonably calculating first and last particle.The present invention adopts following formula to calculate the pixel distance Δ P of first and last particle respectively:

$\mathrm{\ΔP}\left(s\right)=l*(P(s+1)-P\left(s\right))=l*\underset{i=1}{\overset{n}{\mathrm{\Σ}}}|p_i(s+1)-p_i\left(s\right)|$

$\mathrm{\ΔP}\left(e\right)=l*(P\left(e\right)-P(e-1))=l*\underset{i=1}{\overset{n}{\mathrm{\Σ}}}|p_i\left(e\right)-p_i(e-1)|$

Δ P (s) in formula---the pixel distance of the first particle after sequence

P (s+1)---with the first particle adjacent particles

Δ P (e)---the pixel distance of end particle

P (e-1)---with end particle adjacent particles

L---influence coefficient

At the iteration initial stage, l gets higher value, and to expand population hunting zone, increase population diversity, along with the carrying out of iteration, in order to strengthen population Local Search precision, l gets smaller value.

As shown in Figure 1, a kind of fin heat exchanger core Optimization Design that the present invention proposes, its flow process comprises: input heat exchanger main performance requirements and physical properties of fluids parameter, builds mathematical model, utilizes modified particle swarm optiziation to obtain optimum solution.

Now to optimize certain fin heat exchanger core structure optimization, exchanger heat side is two flow processs, and cold side is a flow process, heat exchanger core body general structure as shown in Figure 2, heat exchanger core body flow passage structure as shown in Figure 3, and without phase transformation in heat transfer process, implements concrete steps of the present invention as follows:

First step: the design performance requirement of tablet fin heat exchanger, refers to table 1;

The design performance requirement of table 1 heat interchanger

Second step: input cold and hot fluid physical parameter, refers to table 2;

Table 2 cold and hot fluid physical parameter

Parameter	μ/(Pa·s)	λ/(W·m -2·K -1)	c p/(kJ·kg -1·K -1)	Pr
					Hot fluid	23.8502*10 -6	3.5296*10 -6	1.0144	0.6813
Cold fluid	22.0797*10 -6	3.2376*10 -6	1.0094	0.6877

Third step: input cold and hot fluid both sides fin type, refers to table 3;

Table 3 hot and cold both sides fin type

4th step: build multiple-objection optimization mathematical model;

F (X)=f (L ₁, L ₂, s _1f, s _2f, s ₁, s ₂, δ _1f, δ _2f, δ _p)=f (x ₁, x ₂, x ₃, x ₄, x ₅, x ₆, x ₇, x ₈, x ₉), in formula, Optimal Parameters meaning and restriction range refer to table 4.

5th step: input structure Optimal Parameters constraint condition, refers to table 4;

Table 4 input structure Optimal Parameters constraint condition

6th step: initialization population;

(1) specify each correlation parameter of particle cluster algorithm, refer to table 5;

Each correlation parameter of particle cluster algorithm specified by table 5

(2) stochastic generation 200 particles, the position vector of each particle

X=[L ₁, L ₂, s _1f, s _2f, s ₁, s ₂, δ _1f, δ _2f, δ _p]=[x ₁, x ₂, x ₃, x ₄, x ₅, x ₆, x ₇, x ₈, x ₉], and

(3) the velocity vector V=[v of each particle of stochastic generation ₁, v ₂, v ₃, v ₄, v ₅, v ₆, v ₇, v ₈, v ₉], and $v_i^{\min }{\≤v}_i\≤v_i^{\max };$

(4) fitness function f is utilized to evaluate all particles;

(5) using initialization evaluation of estimate as history optimum solution P _i, find optimum solution P in total group according to particle pixel distance _g;

7th step: find optimum solution in maximum iteration time;

1) position of Population Regeneration particle and speed;

2) calculate the fitness value of particle, judge dominance relation, upgrade inner domination and separate and non-domination solution set;

3) calculate crossover and mutation evolutionary operator probability adaptively, take random the intersection and Gaussian mutation operation respectively, the more position of new particle, upgrades inner non-domination solution set;

4) judge the dominance relation that inner non-domination solution and outside Pareto separate, and upgrade outside Pareto pond;

5) employing dynamically updates pixel granularity, calculates the location of pixels of particle in outside Pareto pond, and rejects the unnecessary particle of same location of pixels;

6) calculate the pixel distance of particle in outside Pareto pond, choosing maximum pixel distance particle is population global optimum particle;

7) whether evaluation algorithm meets end condition, if so, then terminates to calculate, obtains optimum solution, otherwise enter the 5th step.

8th step: on the basis that heat exchanger structure improves, the cold and hot passage of heat exchanger carries out layout optimization, obtains the optimum solution of core structure of heat exchanger design parameter, completes the fin heat exchanger core Optimal Structure Designing based on dynamic pixel granularity.

The effect that the fin heat exchanger core optimum structure design method that the present invention proposes is applied in certain design of heat exchanger refers to table 6.

Table 6 traditional design and optimal design contrast table of the present invention

Parameter	Traditional design	Optimal design of the present invention
			Hot side liquid length of flow L 1(core body)/mm	260	255
Cold size fluid flows length L 2(core body)/mm	200	195
			Hot side spacing of fin s 1f/mm	1	0.6225
Cold side spacing of fin s 2f/mm	1.5	2
			Hot side plate distance s 1/mm	5	5.1637
Cold side distance between plates s 1/mm	7.5	6.5
			Hot side fin thickness δ 1f/mm	0.15	0.1
Cold side fin thickness δ 2f/mm	0.15	0.1
			Block board thickness δ p/mm	0.5	0.3
Heat exchange efficiency η min	0.96	0.96
			Hot side fin pressure drop Δ p 1max/kPa	8.667	9.42
Cold side fin pressure drop Δ p 2max/kPa	13.1	7.22
			Core body general assembly (TW) F/kg	3.9231	2.9716

Claims (3)

1. based on a fin heat exchanger core structural optimization method for dynamic pixel granularity, it is characterized in that, the step of the method is as follows:

1) main performance requirements of plate type finned heat exchanger and the physical parameter of fluid that need optimization is determined;

2) determine core optimized variable and constraint condition thereof, namely determine the solution space of Structure Optimization Variables vector X phenotype and problem; One group of optimized variable of core body is expressed as follows:

X＝{x ₁,x ₂,x ₃,…,x _k}

In formula, x _irepresent an optimized amount in optimized variable vector, k represents optimized variable sum;

3) according to the 2nd) optimized variable that obtains of step and constraint condition thereof sets up Optimized model, determines the type of objective function and mathematical description form thereof or quantization method, namely final optimum solution; Set up Optimized model formula to be described below:

Solve: f (x ₁, x ₂, x ₃..., x _k)

Target: minf (x ₁, x ₂, x ₃..., x _k)

Constraint: g (x ₁, x ₂, x ₃..., x _k)≤0

h(x ₁,x ₂,x ₃,…,x _k)＝0

$x_i^{\min }\≤x_i\≤x_i^{\max }$

In formula, x _irepresent an optimized amount in optimized variable vector, k represents optimized variable sum, with refer to the minimum of the corresponding optimized amount in optimized variable vector and maximum possible value respectively, f () represents the objective function of optimization problem, and g () represents inequality constrain, and h () represents equality constraint;

4) after setting up Optimized model, adopt the particle in population to represent the feasible solution of optimized variable vector, population scale, iteration algebraically, pixel granularity and outside Pareto pond initiation parameter are set, and initialization is carried out to the position of all particles and speed;

5) position of Population Regeneration particle and speed;

6) calculate the fitness value of particle, judge dominance relation, upgrade inner domination and separate and non-domination solution set;

7) calculate crossover and mutation evolutionary operator probability adaptively, take random the intersection and Gaussian mutation operation respectively, the more position of new particle, upgrades inner non-domination solution set;

8) judge the dominance relation that inner non-domination solution and outside Pareto separate, and upgrade outside Pareto pond;

9) employing dynamically updates pixel granularity, calculates the location of pixels of particle in outside Pareto pond, and rejects the unnecessary particle of same location of pixels;

10) calculate the pixel distance of particle in outside Pareto pond, choosing maximum pixel distance particle is population global optimum particle;

11) whether evaluation algorithm meets end condition, if so, then terminates to calculate, obtains optimum solution, otherwise enter the 5th) step.

2. the method realizing fin heat exchanger core structure optimization according to claim 1, it is characterized in that: the described 7th) in step, calculate crossover and mutation evolutionary operator probability adaptively and refer to: the crossover and mutation evolutionary operator probability of particle carrys out Dynamic Acquisition according to the pixel distance of particle.

3. the method realizing fin heat exchanger core structure optimization according to claim 1, is characterized in that: the described 9th) in step, and employing dynamically updates pixel granularity and refers to: dynamically update according to iterations.