CN105549677B - Photovoltaic maximum power tracking method based on simulated annealing method - Google Patents

Abstract

The invention discloses a photovoltaic maximum power tracking method based on a simulated annealing method. By means of the method, a global maximum power point can be reliably tracked by excluding local maximum power points. As an irradiance on-line monitoring mechanism is added, changes of irradiance can be recognized through the algorithm; GMPPT is restarted, and the new GMPP is rapidly tracked. Due to neighborhood adaptivity and adjustability, the tracking speed is greatly increased, power oscillation in the tracking process is reduced, and power loss is reduced. Due to finer algorithm ending conditions, GMPP tracking is more accurate, and the efficiency of a photovoltaic system is improved.

Description

A kind of photovoltaic maximum power tracking method based on simulated annealing

【Technical field】

The invention belongs to solar energy power generating research field, particularly to a kind of photovoltaic system based on simulated annealing The optimized algorithm of maximal power tracing.

【Background technology】

In recent years, due to a large amount of, the quick use of fossil fuel, environmental problem and global energy crisis are increasingly serious.Make For main clean reproducible energy, solar electrical energy generation has obtained universal concern and has paid attention to and have a wide range of applications： Energy total amount is huge, cleanliness without any pollution and geographically unrestricted, and meanwhile, solar power system safety and stability is it is easy to reality Apply and safeguard.However, in actual photovoltaic system applied environment, the blocking of surrounding objects shade, the stain of solar panels, tool Body is installed and can be had a strong impact on photovoltaic system electricity yield towards not parity problem.The problems referred to above are commonly known as the mismatch of photovoltaic system Problem, this can lead to the overall output static characteristic curve of photovoltaic array to present " multiple maximum power point (Multiple Maximum Power Point, MMPP) " phenomenon so that traditional centralized maximal power tracing (Maximum Power Point Tracking, MPPT) control algolithm be difficult to differentiate between local maximum power point (Local Maximum Power Point, LMPP) and global maximum power point (Global Maximum Power Point, GMPP), reduce the output work of photovoltaic system Rate, has a strong impact on the efficiency of its solar energy conversion.

In order to solve this problem, based on fuzzy control (Fuzzy Logic Control), particle cluster algorithm A series of new global maximum power point such as (Particle Swarm Optimization) follows the tracks of (Global Maximum Power Point Tracking, GMPPT) algorithm is suggested, substantially increases the probability tracking GMPP, improve photovoltaic system The efficiency of system.However, above-mentioned algorithm often has larger power loss or needs single for each when following the trail of GMPP Photovoltaic system is optimized.By contrast, voltage initial value is not relied on based on the GMPPT algorithm of simulated annealing it is not necessary to right Each single photovoltaic system is optimized, therefore embodies its exclusive advantage and using value.But existing simulated annealing Method can only follow the trail of the maximum of the power voltage curve of static state it is impossible to tackle the situation of irradiation level change；Algorithm stop condition is Temperature drop, to certain threshold value, follows the trail of precision not high；Lack effective adjustment of neighborhood, lead to tracking speed slow.

【Content of the invention】

In order to overcome defect on photovoltaic system GMPPT for the present day analog annealing method, the present invention proposes one kind can be fast Speed accurately follows the trail of the improvement GMPPT algorithm based on simulated annealing of photovoltaic system GMPP.

The present invention employs the following technical solutions：

A kind of photovoltaic maximum power tracking method based on simulated annealing, comprises the following steps：

(1) randomly select an initial voltage between 0 to open-circuit voltage as quiescent potential V_i, measurement is in this work Power output P under point voltage_i, as reference power；

(2) with carrying out voltage disturbance in neighborhood, obtain new operating voltage V_k, measure this operating voltage V_kUnder defeated Go out power P_k；

(3) judge that step (2) carries out power output P after disturbance_kReference power P with step (1)_iRelation, if P_k More than P_i, then accept the new reference point in this disturbance point position, meanwhile, if P_kMaximum power value P more than current record_max, then more New maximum power value；If P_kLess than P_i, then calculate acceptance probability with reference to Metropolios, judged whether this disturbance point with this As new reference point；

(4) if the number of times of disturbance has reached the disturbance times N setting in advance in a certain neighborhood_S, then adjustment neighborhood is big Little, when the number of times adjusting neighborhood reaches the adjustment number of times setting in advance, current environment is lowered the temperature, and current disturbance point is made For current maximum power point；

(5) judge whether to track peak power.

Preferably, the V in described step (2)_kCalculate according to below equation：

V_k=V_i+r*step

Wherein, r is the random number between [- 1,1], and step is current Size of Neighborhood.

Preferably, the reference Metropolios in step (3) calculates acceptance probability, calculates according to below equation：

$P_r=\exp \[\frac{P_k-P_i}{T_k}\]$

Wherein, P_rFor acceptance probability, T_kFor Current Temperatures.

Preferably, in step (3), calculate acceptance probability with reference to Metropolios, judge whether to make this disturbance point with this Method for new reference point is：If acceptance probability P_rMore than 0.1, then accept disturbance point, otherwise refuse.

Preferably, step (4) adjusts Size of Neighborhood according to below equation：

$\left.\begin{array}{cc}newstep=step*(1+2*\frac{No\_accept/N_s-0.6}{0.4}),& ifNo\_accept>0.6*N_s\\ mewstep=\frac{step}{1+2*\frac{0.4-No\_accept/N_s}{0.4}},& ifNo\_accept<0.4*N_s\\ newstep=step,& otherwise\end{array}\right\}$

Wherein, new step is the Size of Neighborhood after adjustment, and No_accept is the number of times accepting disturbance point, N_SIt is given Disturbance number of times in neighborhood, step is current neighborhood size.

Preferably, in step (4), current environment cooling is carried out according to below equation：T_k+1=α T_k, wherein, T_kFor current Temperature, T_k+1For the temperature after cooling, α is temperature lapse rate.

Preferably, judge whether that the method tracking peak power is：

Δ P=| P_i-P_max|, ifAnd continuous N power output all record peak power near, then recognize For tracking GMPP.

Preferably, further include step (6)：After tracking GMPP, the disturbance to voltage for the stopping, output voltage, and Real-time monitoring power output, judges whether the change of irradiation level, if there is the change of irradiation level, then restarts GMPPT, otherwise output voltage.

Preferably, if below equation is set up then it is assumed that irradiation level changes：

$\frac{\mathrm{\&Delta;}P}{P_{max}}>\mathrm{\&delta;}.$

Compared with prior art, the present invention at least has the advantages that：The invention discloses one kind is moved back based on simulation The improvement global maximum power point track algorithm of pyrogenic process.When the P-V characteristic of photovoltaic array present because local shades block many During peak phenomenon, can quickly and accurately track the global maximum power point of photovoltaic array；Self adaptation due to neighborhood can Adjust, tracing process can avoid fluctuating widely of power；This algorithm can detect that irradiation level changes simultaneously, and follows the trail of in time To new GMPP.

【Brief description】

Fig. 1 is the flow chart based on the GMPPT algorithm of simulated annealing for the present invention.

Fig. 2 (a), (b) are the P-V characteristic curve in embodiment.

Fig. 3 (a), (b) are GMPPT process schematic in embodiment.

【Specific embodiment】

Proposed by the present invention based on the GMPPT algorithm of simulated annealing be melted by analogue crystals after decrease temperature crystalline mistake Journey, constructs a similar cooling mechanism, carries out the disturbance of certain number of times voltage at a temperature of each, sentences with reference to Metropolis Accept according to decision or refuse new operating point, and accept, according to actual, the size that ratio shared by number of times adjusts neighborhood, thus reaching To not only quickly but also accurately following the trail of effect, this algorithm can detect that irradiation level changes simultaneously, and tracks new GMPP in time.Its stream Journey figure is as shown in Figure 1.

With reference to embodiment and accompanying drawing, the present invention will be further described in detail.The flow chart of GMPPT in embodiment As shown in Figure 1, comprise the following steps：

1st, parameter setting

Setting initial temperature T₀, temperature lapse rate α, Size of Neighborhood step, end mark flag, the disturbance in given neighborhood Times N_S, the neighborhood adjustment times N under given temperature_T.

2nd, initialization voltage

Arrive open-circuit voltage V 0_ocBetween randomly select an initial voltage V₀As initial quiescent potential.

3rd, measure power

According to initialized voltage V₀Measure corresponding power output P₀.

4th, GMPPT process

4.1 carry out voltage disturbance in neighborhood at random, obtain new operating voltage V_k.Specifically producing method is：

V_k=V_i+r*step (1)

Wherein, r is the random number between [- 1,1], and step is current Size of Neighborhood, for V_iFor, as i=1, V_i =V₀.

4.2 measure V_kCorresponding power output P_k.And disturbance times N o_perturb is added 1.

If 4.3 P_kMore than reference power P_i, for P_iFor, as i=1, P_i=P₀, then accepting disturbance point is new ginseng Examination point, meanwhile, times N o_accept accepting new explanation is added 1：

P_i=P_k,V_i=V_k(2)

If P_kMaximum power value P more than current record_max, update maximum power value：

P_max=P_k,V_max=V_k(3)

If 4.4 P_kLess than reference power P_i, for P_iFor, as i=1, P_i=P₀, then with reference to following Metropolios formula calculates acceptance probability, judges whether that accepting disturbance point is new reference point with this：

$P_r=\exp \&lsqb;\frac{P_k-P_i}{T_k}\&rsqb;---\left(4\right)$

If P_rMore than 0.1, then accept disturbance point, otherwise refuse.If received, times N o_ of new explanation will be accepted Accept adds one.

If 4.5 in a certain neighborhood disturbance number of times reach N_SSecondary, adjust Size of Neighborhood in the following manner：

$\left.\begin{array}{cc}newstep=step*(1+2*\frac{No\_accept/N_s-0.6}{0.4}),& ifNo\_accept>0.6*N_s\\ mewstep=\frac{step}{1+2*\frac{0.4-No\_accept/N_s}{0.4}},& ifNo\_accept<0.4*N_s\\ newstep=step,& otherwise\end{array}\right\}---\left(5\right)$

Reinitialize No_perturb, No_accept, and No_step plus 1.

If 4.6 neighborhood adjustment number of times reach N_TSecondary, lowered the temperature in the following manner：

T_k+1=α T_k(6)

Reinitialize No_step, reference point is set to current maximum power point：

V_i=V_max(7)

5th, judge whether to track GMPP：

Δ P=| P_i-P_max| (8)

If

$\frac{\mathrm{\&Delta;}P}{P_{max}}<\mathrm{\&delta;}---\left(9\right)$

Then make

Flag=flag+1 (10)

When flag is more than M, illustrate that continuous N power output, all near the peak power of record, now thinks and chases after Track, to GMPP, enters 6, otherwise returns to 4.

6th, export GMPP, wait algorithm to restart

After tracking GMPP, algorithm stops the disturbance to voltage, exports V_i=V_mpp, and real-time monitoring power output P_i, Judge whether irradiation level changes according to the following formula：

If

$\frac{\mathrm{\&Delta;}P}{P_{max}}>\mathrm{\&delta;}---\left(11\right)$

Illustrate that power output has larger fluctuation, there is the change of irradiation level, then restart GMPPT, return to 1.Otherwise, continue Continuous output V_i=V_mpp.

Three photovoltaic module series connection are adopted, the irradiation level of photovoltaic module 1 was undergone mutation when 0.4 second in the present embodiment.Its Corresponding P-V characteristic is Fig. 2 (a) (0.4 second before) and Fig. 2 (b) (0.4 second afterwards).As can be seen that because local shades hide Gear, leads to three pieces of photovoltaic module P-V characteristics inconsistent, therefore assumes 3 peak values in 0-0.4 second total P-V characteristic.By Fig. 3 (a) Rapidity and the reliability of algorithm, it can be seen that power output quickly reaches GMPP, are described.After 0.4 second, irradiation level occurs to dash forward Become, algorithm restarts tracking immediately, and quickly finds new GMPP again, illustrates that algorithm can tackle the change of irradiation level well Change.By Fig. 3 (b) as can be seen that due to the Automatic adjusument of neighborhood, scope range of the fluctuation of voltage is less, restrained effectively larger Power disturbance, decreases power loss and improves tracking speed.

Claims (8)

1. a kind of photovoltaic maximum power tracking method based on simulated annealing it is characterised in that：Comprise the following steps：

(1) randomly select an initial voltage between 0 to open-circuit voltage as quiescent potential V_i, measurement is in this operating point electricity Power output P of pressure_i, as reference power；

(2) carry out voltage disturbance at random in neighborhood, obtain new operating voltage V_k, measure this operating voltage V_kUnder output work Rate P_k；

(3) judge that step (2) carries out power output P after disturbance_kReference power P with step (1)_iRelation, if P_kIt is more than P_i, then accept the new reference point in this disturbance point position, meanwhile, if P_kMaximum power value P more than current record_max, then update High-power value；If P_kLess than P_i, then with reference to Metropolios calculate acceptance probability, using this judge whether by this disturbance point as New reference point；

(4) if the number of times of disturbance has reached the disturbance times N setting in advance in a certain neighborhood_S, then adjusted according to below equation Size of Neighborhood, when adjust neighborhood number of times reach the adjustment number of times setting in advance when, to current environment lower the temperature, and by current High-power point is as current reference point；

$\left.\begin{array}{cc}\mathrm{new\; step}=\mathrm{step}*(1+2*\frac{\mathrm{No}\_\mathrm{accept}/N_s-06}{0.4}),& \mathrm{if\; No}\_\mathrm{accept}>0.6*N_s\\ \mathrm{new\; step}=\frac{\mathrm{step}}{1+12*\frac{0.4-\mathrm{No}\_\mathrm{accept}/N_s}{0.4}},& \mathrm{if\; No}\_\mathrm{accept}<0.4*N_s\\ \mathrm{new\; step}=\mathrm{step},& \mathrm{otherwise}\end{array}\right\}$

Wherein, new step is the Size of Neighborhood after adjustment, and No_accept is the number of times accepting disturbance point, N_SFor in given neighborhood Disturbance number of times, step be current neighborhood size；

(5) judge whether to track peak power.

2. a kind of photovoltaic maximum power tracking method based on simulated annealing according to claim 1 it is characterised in that： V in described step (2)_kCalculate according to below equation：

V_k=V_i+r*step

Wherein, r is the random number between [- 1,1], and step is current Size of Neighborhood.

3. a kind of photovoltaic maximum power tracking method based on simulated annealing according to claim 1 it is characterised in that： Reference Metropolios in step (3) calculates acceptance probability, calculates according to below equation：

$P_k=\exp \&lsqb;\frac{P_k-P_i}{T_k}\&rsqb;$

Wherein, P_rFor acceptance probability, T_kFor Current Temperatures.

4. a kind of photovoltaic maximum power tracking method based on simulated annealing according to claim 1 or 3, its feature exists In：In step (3), calculate acceptance probability with reference to Metropolios, judged whether this disturbance point as new reference point using this Method be：If acceptance probability P_rMore than 0.1, then accept disturbance point, otherwise refuse.

5. a kind of photovoltaic maximum power tracking method based on simulated annealing according to claim 1 it is characterised in that： In step (4), current environment cooling is carried out according to below equation：T_k+1=α T_k, wherein, T_kFor Current Temperatures, T_k+1For cooling Temperature afterwards, α is temperature lapse rate.

6. a kind of photovoltaic maximum power tracking method based on simulated annealing according to claim 1 it is characterised in that： Judge whether that the method tracking peak power is：

And continuous N power output all near the peak power of record then it is assumed that Track GMPP.

7. a kind of photovoltaic maximum power tracking method based on simulated annealing according to claim 1 it is characterised in that： Further include step (6)：After tracking GMPP, stop the disturbance to voltage, output voltage, and real-time monitoring output work Rate, judges whether the change of irradiation level, if there is the change of irradiation level, then restarts GMPPT, otherwise exports electricity Pressure.

8. a kind of photovoltaic maximum power tracking method based on simulated annealing according to claim 7 it is characterised in that： If below equation is set up then it is assumed that irradiation level changes：

$\frac{\mathrm{\&Delta;}P}{P_{\max }}>\mathrm{\&delta;}.$