CN103324239A - Method for quickly optimizing overall maximum power point of photovoltaic array under local shadow - Google Patents

Abstract

The invention discloses a method for quickly optimizing an overall maximum power point of a photovoltaic array under local shadow. Under the local overshadowing condition, output characteristics of the photovoltaic array present multiple peak points. For searching an overall maximum value, the method comprises the steps of adopting a modified fruit fly algorithm to perform overall range searching, confirming a range where maximum power is located, then finding a maximum power point in a sub-region in a confirmed region by improving a golden section method and using the maximum power point as a working point of the whole photovoltaic array. By adopting the two steps, the overall maximum power point of the photovoltaic array can be quickly optimized.

Description

The quick optimization method of photovoltaic array global maximum power point under the local shade

Technical field

The invention belongs to optimum control and optimization method research field, specifically be applied in the photovoltaic generating system, solve the method for Maximum Power Output in the shade situation.

Background technology

The part situation ubiquity in grid-connected photovoltaic system that shelters from heat or light, its meeting can cause that " hot spot " effect causes safety problem so that the photovoltaic generating system output power descends when serious.Simultaneously, under shelter from heat or light situation in the part, because intensity of illumination and temperature is different, the output characteristics of photovoltaic array can present a plurality of peak points, make conventional maximum power point optimizing algorithm very likely can be trapped in certain local maximum power value, rather than real maximum power point, cause the mismatch of output power.

Utilize the global search of revising the fruit bat algorithm and the Local Search method of improving Fibonacci method to combine, can realize the maximum power point of photovoltaic array is carried out quick optimizing.

Summary of the invention

Goal of the invention: the output power curve of photovoltaic array presents the multimodal shape under the local shade, and conventional maximum power point optimizing algorithm can't guarantee to search the problem of global maximum power point.The present invention proposes the output power that system is total as objective function, use correction fruit bat algorithm to carry out the maximal value in-scope that gross output is found in global search; Then realize single-peaked local optimal searching by improving Fibonacci method.This algorithm can avoid being absorbed in Local Extremum in the multi-peak situation, can find the global maximum power point of photovoltaic system.

Technical scheme: for achieving the above object, the present invention is by the following technical solutions: the quick optimization method of photovoltaic array global maximum power point under a kind of local shade comprises the global search of revising the fruit bat algorithm and the Local Search method two parts that improve Fibonacci method:

(1) full search method of correction fruit bat algorithm: the power function of definition photovoltaic array is objective function, and variable is the output voltage of photovoltaic array, and the step of revising the search global search of fruit bat colony is as follows:

Step 1: random initial fruit bat colony position;

InitX_axis；InitY_axis

Step 2: given fruit bat individuality utilizes random direction and the distance of sense of smell search of food;

X _i＝X_axis+RandomValue；Y _i＝Y_axis+RandomValue

Step 3: owing to can't learn the particular location of food, therefore first by following formula, calculate the distance D ist of this point and initial point; Calculate flavor concentration decision content S, this value is the inverse of distance again;

${\mathrm{Dist}}_i=\sqrt{X_i^2+Y_i^2};$ $S_i=\frac{1}{{\mathrm{Dist}}_i}$

Step 4: flavor concentration decision content S substitution flavor concentration decision function (fitness function) is to obtain the flavor concentration Smell of this fruit bat body position _i

Smell _i＝Function(S _i)

Step 5: maximizing, by following formula, find out the highest fruit bat of flavor concentration in this fruit bat colony;

[bestSmell?bestIndex]＝max(Smell)

Step 6: keep this best flavors concentration value (fitness function value) and x, y coordinate, this moment, fruit bat colony utilized vision to fly to toward this position;

Smellbest＝bestSmell

X_axis＝X(bestIndex)

Y_axis＝Y(bestIndex)

Step 7: enter the iteration optimizing, repeated execution of steps 2-5, and judge whether the flavor concentration value is better than a front iteration flavor concentration value, if then execution in step 6;

The fruit bat algorithm is in photovoltaic array maximum power point searching process, and whole algorithm is divided into two parts: a part is that the optimizing link of fruit bat, another part are determining of maximum power point region.In the FOA algorithm, the potential solution of each optimization problem (being maximum power point) is " fruit bat " in the search volume, and all fruit bat individualities have an adaptive value that is determined by objective function.

(2) the Local Search method of improvement Fibonacci method: Fibonacci method, be called again 0.618 method, it is applicable in given interval a kind of searching method of search extreme point in [a, b].The Local Search step of improving Fibonacci method is as follows:

Step 1, at first, the functional value of known f (x1), f (x2) and f (x3), because f (x3)＞f (x1) and f (x3)＞f (x2), so must there be maximal value to be present in the interval [x1, x2];

Step 2, insert a new some x4 in this interval, then f (x4) as can be known;

Step 3, in random preliminary examination fruit bat colony position curve (curve among Fig. 1 shown in the solid line), if f (x4)＞f (x3) is arranged, then maximum of points is positioned at interval (x3, x2).Insert the x4 point this moment, and give up the x1 point, again obtains the new narrower region of search (x3, x2), and three known point x3, x4 and x2 are still arranged in this interval;

Step 4, in random preliminary examination fruit bat colony position curve (curve among Fig. 1 shown in the solid line), if f (x4)＜f (x3) is arranged, then maximum of points is positioned at interval (x1, x4).Still insert the x4 point this moment, but give up the x2 point, again obtains another new narrower region of search (x1, x4), and three known point x1, x3 and x4 are still arranged in this interval;

Step 5, repeatedly circulation step 2 can reach predefined scope to step 4 until the region of search is very little, and till when guaranteeing that the new point that inserts is approximately maximum of points.

Beneficial effect: compared with prior art, the present invention has the following advantages: the global search of the correction fruit bat algorithm that the present invention proposes and the Local Search method of the improving Fibonacci method algorithm that combines, realization is to the quick optimizing of global maximum power point, thereby guaranteed to occur under the local shade circumstance of occlusion at photovoltaic array, still can effectively find the maximum power point in the multi-peak situation.Compare with conventional photovoltaic array maximum power point optimizing algorithm, not only overcome the defective of unimodal value searching algorithm, prevent from being absorbed in the part and be worth most a little, and strengthened the to external world quick adaptive faculty of environmental change of this two-step approach.

Description of drawings

Fig. 1 is structural principle schematic diagram of the present invention;

Fig. 2 is the schematic diagram of fruit bat colony iterative search food among the present invention;

Fig. 3 is the realistic model figure of photovoltaic array plate of the present invention;

Fig. 4 is that local shade blocks down, photovoltaic array U-P family curve;

Fig. 5 is that local shade blocks down, photovoltaic array U-I family curve;

Fig. 6 is the global search schematic flow sheet based on revising the fruit bat optimization algorithm of the present invention;

Fig. 7 is the search procedure of improvement Fibonacci method of the present invention;

Fig. 8 is the Local Search schematic flow sheet based on improving Fibonacci method of the present invention.

Embodiment

Below in conjunction with the drawings and specific embodiments, further illustrate the present invention, should understand these embodiment only is used for explanation the present invention and is not used in and limits the scope of the invention, after having read the present invention, those skilled in the art all fall within the application's claims limited range to the modification of the various equivalent form of values of the present invention.

Under different illumination intensity and the temperature, the output of photovoltaic array presents different U-I and U-P family curve.The photovoltaic array that is together in series, the electric current that flows through equates; The photovoltaic array that is together in parallel, terminal voltage equates.When the part occurring and shelter from heat or light phenomenon, so that the U-P family curve of photovoltaic array integral body presents a plurality of extreme points, such as Fig. 3, Fig. 4, shown in Figure 5, this moment, traditional unimodal value optimizing algorithm then can lose efficacy.

This method first carries out global search by revising the fruit bat algorithm, determines the regional extent at maximum power point place, and second portion carries out Local Search by improving Fibonacci method in the regulation regional extent, be stabilized in the maximum power point place.

Fruit bat optimization algorithm (Fly Optimization Algorithm is called for short FOA) is a kind of new method of seeking global optimization of deducing out based on the fruit bat foraging behavior, and search procedure as shown in Figure 1.

The step of search global optimum of fruit bat colony is as follows:

(1) initialization of algorithm

Position, population quantity, iterations and the iteration step value of initial fruit bat colony are set.

(2) population is estimated

After various parameters are carried out initialization, calculate the adaptive value of the corresponding objective function of each fruit bat.Objective function is the general power of photovoltaic array output, and the expression formula of fitness function is:

${\mathrm{Smell}}_{i,j}=\underset{j=1}{\overset{m}{\mathrm{\Σ}}}\{I_j\×\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\mathrm{PVprog}(I_j,{\mathrm{Sun}}_i,T_i)\}---\left(1\right)$

$\mathrm{PVprog}(I,\mathrm{Sun},T)=1.1103\×\log \left(\frac{{3.8\×G-I+2.2\×10}^{-8}}{2.2\×{10}^{-8}}\right)-0.2844\×I---\left(2\right)$

Wherein, being the output characteristic function of photovoltaic array, is 25 ° to fixed temperature.

By calling the size of a fitness function calculated power value, this function has comprised in fact the current/voltage formula of each module and they has been added and have formed the power stage of total system.

(3) relatively determine individuality and overall fitness and extreme value

The current fitness value of more single fruit bat and the historical preferably size of adaptive value if current adaptive value is larger, are upgraded the fitness value of fruit bat so.The performance number that preferably adapts to of each fruit bat is determined rear mutually relatively to determine the best fitness value of the overall situation.

(4) position of renewal fruit bat.

(5) check termination condition, if satisfy, finish optimizing, the scope of output optimum solution; Otherwise go to (2), termination condition is that optimizing reaches maximum evolutionary generation.

Based on the optimizing flow process of revising the fruit bat optimization algorithm as shown in Figure 6.

Fibonacci method (Golden Section Search is called for short GSS) is called again 0.618 method, and it is applicable in given interval a kind of searching method of search extreme point in [a, b], and search procedure as shown in Figure 6.

Obviously, iteration has two the possible region of search (x3, x2) and (x1, x4) each time in Fig. 7, and only has one will be selected as the next region of search between the two.This just needs these two search volumes wide, otherwise in the situation that worse, if the frequency of utilization of wider search volume is high, speed of convergence will slow down so, and will increase search time.Therefore, this just needs the new some x4 that inserts to satisfy:

| x1, x4|=|x3, x2|, i.e. a+b=b+c among the upper figure.

So in iteration each time, by these three spacings that the region of search that determines all is equal proportion, then this algorithm is with a constant speed convergence.

Following formula is found the solution and can be got:

Herein

It is the gold rate.

Derive as can be known by analysis:

If 1 need to narrow down to 0.1% of former region of search length, then want 15 iteration to get final product.Therefore, the speed of convergence of Fibonacci method is very fast.

If arrange between 2 original areas rationally, can find out maximum power point with Fibonacci method.Search variables can be electric current, can also be voltage.

3, using Fibonacci method to seek maximum power point does not need to carry out differential calculation, so robustness and noise resisting ability are very strong.In real system, sensor and signal fluctuation are always inevitable in differential calculation, especially the switchgear in the power converter.

4, in addition, have certain interval between the maximum power point of the maximum power point that Fibonacci method draws in iterative process and reality, therefore, this method can be allowed relatively large noise and disturbance.The photovoltaic system of the overwhelming majority all need to use switch-mode converter with the power stage of photovoltaic array to electrical network, independent load, perhaps energy-storage travelling wave tube, transducer all contains internal noise and harmonic power in this process, and the strong robustness of Fibonacci method just is being fit to the maximum power point optimizing in the photovoltaic generating system.

Claims (1)

1. the quick optimization method of photovoltaic array global maximum power point under the local shade comprises the global search of revising the fruit bat algorithm and the Local Search method two parts that improve Fibonacci method:

(1) full search method of correction fruit bat algorithm: the power function of definition photovoltaic array is objective function, and variable is the output voltage of photovoltaic array, and the step of revising the search global search of fruit bat colony is as follows:

Step 1: random initial fruit bat colony position;

InitX_axis；InitY_axis

Step 2: given fruit bat individuality utilizes random direction and the distance of sense of smell search of food;

X _i＝X_axis+RandomValue；Y _i＝Y_axis+RandomValue

Step 3: owing to can't learn the particular location of food, therefore first by following formula, calculate the distance D ist of this point and initial point; Calculate flavor concentration decision content S, this value is the inverse of distance again;

${\mathrm{Dist}}_i=\sqrt{X_i^2+Y_i^2};$ $S_i=\frac{1}{{\mathrm{Dist}}_i}$

Step 4: flavor concentration decision content S substitution flavor concentration decision function is to obtain the flavor concentration Smell of this fruit bat body position _i

Smell _i＝Function(S _i)

Step 5: maximizing, by following formula, find out the highest fruit bat of flavor concentration in this fruit bat colony;

[bestSmellbestIndex]＝max(Smell)

Step 6: keep this best flavors concentration value and x, y coordinate, this moment, fruit bat colony utilized vision to fly to toward this position;

Smellbest＝bestSmell

X_axis＝X(bestIndex)

Y_axis＝Y(bestIndex)

Step 7: enter the iteration optimizing, repeated execution of steps 2-5, and judge whether the flavor concentration value is better than a front iteration flavor concentration value, if then execution in step 6;

(2) the Local Search method of improvement Fibonacci method: the Local Search step of improving Fibonacci method is as follows:

Step 1, at first, the functional value of known f (x1), f (x2) and f (x3), because f (x3)＞f (x1) and f (x3)＞f (x2), so must there be maximal value to be present in the interval [x1, x2];

Step 2, insert a new some x4 in this interval, then f (x4) as can be known;

Step 3, in random preliminary examination fruit bat colony position curve, if f (x4)＞f (x3) is arranged, then maximum of points is positioned at interval (x3, x2).Insert the x4 point this moment, and give up the x1 point, again obtains the new narrower region of search (x3, x2), and three known point x3, x4 and x2 are still arranged in this interval;

Step 4, in random preliminary examination fruit bat colony position curve, if f (x4)＜f (x3) is arranged, then maximum of points is positioned at interval (x1, x4).Still insert the x4 point this moment, but give up the x2 point, again obtains another new narrower region of search (x1, x4), and three known point x1, x3 and x4 are still arranged in this interval;

Step 5, repeatedly circulation step 2 can reach predefined scope to step 4 until the region of search is very little, and till when guaranteeing that the new point that inserts is approximately maximum of points.

Images