An ics-ip&o-based photovoltaic maximum power tracking intelligent algorithm

Abstract

The application discloses an ICS-IP&O-based photovoltaic maximum power tracking intelligent algorithm, Lévy The flight step coefficient is adaptively adjusted, the particle swarm algorithm thought is introduced, and the algorithm is combined in the preferential random walk part, so that the algorithm has strong global search randomness in the initial stage, and the possibility of falling into local optimization is reduced; in the later stage of the algorithm, local convergence is quickly carried out through the variable step size disturbance observation method, and the oscillation in output is reduced. Meanwhile, the algorithm restart condition is increased, the restart re-optimization is quickly carried out when the external illumination changes too much, and the accuracy is ensured.

Description

Technical Field

[0001] This invention belongs to the field of photovoltaic maximum power point tracking, and specifically relates to a photovoltaic maximum power point tracking intelligent algorithm based on ICS-IP&O. Background Technology

[0002] In photovoltaic (PV) power generation, changes in ambient temperature and sunlight can significantly impact the power output of the PV system. Localized shading caused by tree obstruction, cloud cover, and dust accumulation on PV module surfaces leads to uneven sunlight distribution. Maximum power point tracking (MPP) technology aims to position the PV array at its maximum power point (MPP) as close as possible to changes in ambient temperature and sunlight intensity. MPP technology can greatly reduce power loss. Existing methods such as constant voltage, short-circuit current, perturbation and observation, and admittance increment methods perform well under consistent lighting conditions, but become prone to local power extrema when localized shading occurs due to the multi-peaked output curve. Therefore, a smart MPP algorithm based on the Adaptive Cuckoo Algorithm (ICS) and Variable Step Size Perturbation and Observation (IP&O) is needed to address these issues. Summary of the Invention

[0003] The technical problem to be solved by this invention is to provide a photovoltaic maximum power point tracking intelligent algorithm based on ICS-IP&O. This method solves the problem that when photovoltaic modules are partially shaded due to uneven illumination, the power output curve of the existing technology exhibits a multi-peak phenomenon. When the difference between the peaks is very small, it is easy to get stuck in the local power extreme point during optimization, resulting in a loss of output power. This method has the characteristics of being able to quickly and accurately perform optimization when photovoltaic modules are partially shaded due to uneven illumination.

[0004] To achieve the above-mentioned technical effects, the technical solution adopted by the present invention is as follows:

[0005] A photovoltaic maximum power point tracking intelligent algorithm based on ICS-IP&O includes the following steps:

[0006] S1, initialize the parameters in ICS; calculate the output power value corresponding to the initialization position;

[0007] S2, perform Lévy flight according to the formula and update the output power; compare the updated power value and record the maximum power value;

[0008] S3, calculate the handover probability p according to the adaptive formula. a And generate a random number, and determine if r is true. i ≥p a Is it true or false?

[0009] If the inequality is true, proceed to step S2 and record the optimal power value at this point.

[0010] If the inequality does not hold, perform a random walk update.

[0011] S4, calculate whether the best and worst positions in this iteration satisfy the switching condition:

[0012] If satisfied, proceed to step S5;

[0013] If the condition is not met, proceed to step S2;

[0014] S5, end the ICS algorithm and switch to the IP&O algorithm for local optimization, and update the position using the position update formula;

[0015] S6, determine if the restart conditions are met:

[0016] If the conditions are met, the algorithm is restarted, and the process jumps to step S1.

[0017] If the conditions are not met, proceed to step S5 to continue local optimization of IP&O.

[0018] S7, output the current optimal power value, and end the algorithm.

[0019] Preferably, the specific method of step S1 is as follows:

[0020] S101, initialize the relevant parameters of the ICS algorithm, including: voltage position U, duty cycle D, power P, optimal voltage U_best and optimal power P_best;

[0021] S102, calculate the output power value corresponding to the initial position.

[0022] Preferably, the specific method in step S2 is as follows:

[0023] The initial data obtained in step S1 is iteratively updated using adaptive Lévy flight, as shown in the following formula:

[0024]

[0025]

[0026] α = 1 - 0.2log T t;

[0027] In the formula, α is the step size coefficient of Lévy flight, which is an adaptive nonlinear decreasing coefficient; L(β) is the flight step size; (symbol omitted) Dot product; It is the position of the i-th bird's nest in the t-th iteration; This represents the optimal parasitic nest in the current iteration; β = 3 / 2; both u and v follow a normal distribution, u ~ N(0, σ 2 v ~ N(0,1).

[0028] Preferably, in step S3, the switching probability p is calculated. a The method for generating a random number is as follows: In the ICS algorithm, the number of bird nests available for parasitism is fixed, and the host bird will have p a The host bird has a certain probability of discovering the parasitic egg; once discovered, it will either destroy the egg or choose to rebuild its nest, abandoning the old one; the specific methods are as follows:

[0029] S301, generate a random number r for each bird's nest. i ;

[0030] S302, for each bird's nest, apply a random probability p a The value is calculated using the following formula:

[0031] p a =0.3-0.2(Tt) / T;

[0032] In the formula, T is the total number of iterations; t is the current iteration number;

[0033] S303, for each bird's nest, r i and p a Comparison: If r i ≥p a Return to step S2 to update Lévy's flight position; if r i <p a Then, the position is updated using a random walk, and the update formula is as follows:

[0034]

[0035] In the formula, It is the i-th position in the t-th iteration; It is the individual's historical best; It is the best in the group's history; It is the group's historical second-best solution. r1, r2, and r3 are three learning factors, all of which are random numbers in the interval [0,1]. Among them, r1 is the learning factor for the individual historical best solution, representing the particle's self-learning ability; r2 and r3 are the learning factors for the global historical best solution and the global historical second-best solution, respectively, representing the particle's social learning ability.

[0036] Preferably, in step S4, the ICS algorithm is used for global optimization in the early stage, and the corresponding position and power values ​​are calculated; the currently calculated best and worst positions are compared, and when the switching condition is met, the algorithm is switched, and in the later stage, the IPO algorithm is used for local optimization; the formula for the algorithm switching condition is as follows:

[0037] |x ibest -x iworst |≤0.05;

[0038] In the formula, x ibest It is the optimal position, x iworst It's the worst position.

[0039] Preferably, in step S5, the algorithm is switched to the IPO algorithm for local optimization, and the specific method is as follows:

[0040] The PV system is sampled, and the output power before and after the disturbance is compared and calculated for optimization. The IPO algorithm uses a variable step size for position update, and the update formula is as follows:

[0041]

[0042] In the formula, λ step ε is the step size; ε is the step size scaling factor; U pv I pv Let be the output voltage and current of the PV system; P(i-1) and P(i) are the power values ​​before and after the i-th disturbance, respectively; ΔP is the power increment before and after the disturbance; if ΔP is positive after the disturbance, the next disturbance will be in the same direction; if ΔP is negative after the disturbance, the next disturbance will be in the opposite direction.

[0043] Preferably, in step S6, since natural environmental factors can cause changes in lighting conditions, the PV output power will also change accordingly. The change is calculated, and when the change meets the algorithm restart condition, the algorithm is initialized and the ICS algorithm is used for optimization again. The restart calculation formula is as follows:

[0044]

[0045] In the formula, p new p represents the changed PV output power. old This represents the PV output power before the change.

[0046] The beneficial effects of this invention are as follows:

[0047] This invention adaptively adjusts the switching probability and Lévy flight step size coefficient of the cuckoo search method, incorporating particle swarm optimization (PSO) principles and integrating them into the preferred random walk component. This gives the algorithm strong global search randomness in the early stages, reducing the possibility of getting stuck in local optima. In the later stages, a variable step size perturbation observation method is used to quickly achieve local convergence, reducing output oscillations. Furthermore, a restart condition is added to the algorithm; when external lighting changes drastically, it quickly restarts and re-optimizes, ensuring accuracy. Attached Figure Description

[0048] Figure 1 This is a flowchart of the ICS-IPO algorithm.

[0049] Figure 2 This is a static simulation power curve.

[0050] Figure 3 This is a PU curve diagram under partial shading.

[0051] Figure 4 This is a simulated power curve for partial shading.

[0052] Figure 5 This is a dynamic simulation curve for ICS-IP&O. Detailed Implementation

[0053] Example 1:

[0054] like Figure 1 As shown, a photovoltaic maximum power point tracking intelligent algorithm based on ICS-IP&O includes the following steps:

[0055] S1, initialize the parameters in ICS; calculate the output power value corresponding to the initialization position;

[0056] S2, perform Lévy flight according to the formula and update the output power; compare the updated power value and record the maximum power value;

[0057] S3, calculate the handover probability p according to the adaptive formula. a And generate a random number, and determine if r is true. i ≥p a Is it true or false?

[0058] If the inequality is true, proceed to step S2 and record the optimal power value at this point.

[0059] If the inequality does not hold, perform a random walk update.

[0060] S4, calculate whether the best and worst positions in this iteration satisfy the switching condition:

[0061] If satisfied, proceed to step S5;

[0062] If the condition is not met, proceed to step S2;

[0063] S5, end the ICS algorithm and switch to the IP&O algorithm for local optimization, and update the position using the position update formula;

[0064] S6, determine if the restart conditions are met:

[0065] If the conditions are met, the algorithm is restarted, and the process jumps to step S1.

[0066] If the conditions are not met, proceed to step S5 to continue local optimization of IP&O.

[0067] S7, output the current optimal power value, and end the algorithm.

[0068] Preferably, the specific method of step S1 is as follows:

[0069] S101, initialize the relevant parameters of the ICS algorithm, including: voltage position U, duty cycle D, power P, optimal voltage U_best and optimal power P_best;

[0070] S102, calculate the output power value corresponding to the initial position.

[0071] Preferably, the specific method in step S2 is as follows:

[0072] The initial data obtained in step S1 is iteratively updated using adaptive Lévy flight, as shown in the following formula:

[0073]

[0074]

[0075] α = 1 - 0.2log T t;

[0076] In the formula, α is the step size coefficient of Lévy flight, which is an adaptive nonlinear decreasing coefficient; L(β) is the flight step size; (symbol omitted) Dot product; It is the position of the i-th bird's nest in the t-th iteration; This represents the optimal parasitic nest in the current iteration; β = 3 / 2; both u and v follow a normal distribution, u ~ N(0, σ 2 v ~ N(0,1).

[0077] Preferably, in step S3, the switching probability p is calculated. aThe method for generating a random number is as follows: In the ICS algorithm, the number of bird nests available for parasitism is fixed, and the host bird will have p a The host bird has a certain probability of discovering the parasitic egg; once discovered, it will either destroy the egg or choose to rebuild its nest, abandoning the old one; the specific methods are as follows:

[0078] S301, generate a random number r for each bird's nest. i ;

[0079] S302, for each bird's nest, apply a random probability p a The value is calculated using the following formula:

[0080] p a =0.3-0.2(Tt) / T;

[0081] In the formula, T is the total number of iterations; t is the current iteration number;

[0082] S303, for each bird's nest, r i and p a Comparison: If r i ≥p a Return to step S2 to update Lévy's flight position; if r i <p a Then, the position is updated using a random walk, and the update formula is as follows:

[0083]

[0084] In the formula, It is the i-th position in the t-th iteration; It is the individual's historical best; It is the best in the group's history; It is the group's historical second-best solution. r1, r2, and r3 are three learning factors, all of which are random numbers in the interval [0,1]. Among them, r1 is the learning factor for the individual historical best solution, representing the particle's self-learning ability; r2 and r3 are the learning factors for the global historical best solution and the global historical second-best solution, respectively, representing the particle's social learning ability.

[0085] Preferably, in step S4, the ICS algorithm is used for global optimization in the early stage, and the corresponding position and power values ​​are calculated; the currently calculated best and worst positions are compared, and when the switching condition is met, the algorithm is switched, and in the later stage, the IPO algorithm is used for local optimization; the formula for the algorithm switching condition is as follows:

[0086] |x ibest -x iworst |≤0.05;

[0087] In the formula, xibest It is the optimal position, x iworst It's the worst position.

[0088] Preferably, in step S5, the algorithm is switched to the IPO algorithm for local optimization, and the specific method is as follows:

[0089] The PV system is sampled, and the output power before and after the disturbance is compared and calculated for optimization. The IPO algorithm uses a variable step size for position update, and the update formula is as follows:

[0090]

[0091] In the formula, λ step ε is the step size; ε is the step size scaling factor; U pv I pv Let be the output voltage and current of the PV system; P(i-1) and P(i) are the power values ​​before and after the i-th disturbance, respectively; ΔP is the power increment before and after the disturbance; if ΔP is positive after the disturbance, the next disturbance will be in the same direction; if ΔP is negative after the disturbance, the next disturbance will be in the opposite direction.

[0092] Preferably, in step S6, since natural environmental factors can cause changes in lighting conditions, the PV output power will also change accordingly. The change is calculated, and when the change meets the algorithm restart condition, the algorithm is initialized and the ICS algorithm is used for optimization again. The restart calculation formula is as follows:

[0093]

[0094] In the formula, p new p represents the changed PV output power. old This represents the PV output power before the change.

[0095] Example 2:

[0096] The invention was simulated using the software Matlab and Simulink. The parameters of a single photovoltaic module are: P max =213.15W, U oc =36.3V, I sc =7.8A.

[0097] A total of three states were simulated:

[0098] State 1 is the unshaded state, with the light intensity set to 1000W / m. 2 The temperature is constant at 25°C.

[0099] State 2 is a partial shading state, with the light intensity set to [1000, 800, 600, 400, 1000] W / m².2 The temperature is 25℃;

[0100] State 3 is a dynamic partial shading state, with a simulation time of 3 seconds;

[0101] At t=0s, the temperature is 25℃, and S=1000W / m 2 Constant; at t = 1s, the temperature is 25℃, S = [1000, 800, 600, 1000]

[0102] W / m 2 At t = 3s, the temperature is 25℃, and S = [1000, 800, 200, 400] W / m 2 .

[0103] When photovoltaic modules operate under the same conditions, the power output curve has only one peak, at which point the maximum power of the system is 1064W. The simulation output of the ICS-IP&O algorithm is performed, and the PSO and CS algorithms are compared.

[0104] Simulation results of the three algorithms are as follows Figure 2 As shown in the figure, the PSO-based system finds the optimal value in about 0.22 seconds, with an optimal value of 1062.8W; the CS-based system finds the optimal value in about 0.19 seconds, with an optimal value of 1063.6W; and the ICS-IP&O-based system finds the optimal value in about 0.17 seconds, with an optimal value of 1064.1W. It can be seen that the system using this invention has a significantly faster convergence and optimization speed compared to PSO and CS, and the stability of the power output curve is also better than the other two methods.

[0105] Simulated partial shading power output curve as shown Figure 3 As shown, the maximum power of the system at this time is 567W; the simulation output of the ICS-IP&O algorithm is performed, and the PSO and CS algorithms are compared at the same time.

[0106] Simulation results of the three algorithms are as follows Figure 4 As shown in the figure, the PSO-based system finds the optimal value in about 0.8 seconds, with an optimal value of approximately 542-543W, but gets stuck in a local optimum and the output is relatively unstable. The CS-based system finds the optimal value in about 0.42 seconds, with an optimal value of 566.8W. The ICS-IP&O-based system finds the optimal value in about 0.28 seconds, with an optimal value of 566.6W. It can be seen that the optimization speed and accuracy of the system using this invention are significantly better than the other two algorithms.

[0107] The dynamic simulation lasts for 3 seconds, simulating real-time changes in light intensity in the natural environment through jumps in light intensity; at t=0, S=1000W / m 2At a constant value, the maximum power is 851.8 W; at t=1, S=[1000,800,600,1000] W / m 2 At this point, the maximum power is 577.7W; at t=3, S=[1000,800,200,400]W / m 2 At this time, the maximum power value is 351.6W.

[0108] The simulation result curves of the system based on the ICS-IPO algorithm are as follows: Figure 5 As shown in the figure, the algorithm found the maximum value of 851.2W at 0.194 seconds; at 1 second, a jump occurred, and the system re-optimized to the maximum value of 568.78W at 1.245 seconds; a second jump occurred at 2 seconds, and the algorithm restarted at 2.22 seconds, finding the maximum value of 351.6W at 2.44 seconds; it can be seen that the present invention is still applicable in dynamic environments and maintains a certain degree of stability and speed.

[0109] The above embodiments are merely preferred technical solutions of this invention and should not be considered as limitations on this invention. The embodiments and features described in this application can be arbitrarily combined without conflict. The scope of protection of this invention should be limited to the technical solutions described in the claims, including equivalent substitutions of the technical features described in the claims. That is, equivalent substitutions and improvements within this scope are also within the scope of protection of this invention.

Claims

1. A photovoltaic maximum power point tracking intelligent algorithm based on ICS-IP&O, characterized in that: Includes the following steps: S1, initialize the parameters in ICS; calculate the output power value corresponding to the initialization position; S2, according to the formula Lévy Flight, update output power; compare the updated power value, and record the maximum power value: among which, adaptive... Lévy Flight step coefficient It decreases non-linearly with the number of iterations t, and its calculation formula is as follows: In the formula, T is the maximum number of iterations; S3, calculate the handover probability according to the adaptive formula. Generate random numbers And compare: If If the result is positive, return to step S2; otherwise, perform a random walk update that incorporates the particle swarm optimization algorithm, with the update formula as follows: in It is the best in individual history. It is the best in the group's history. It is the second-best solution in the history of the group. , , There are three learning factors; S4, Calculate the optimal position for this iteration. and worst position Does the switching condition meet? If satisfied, proceed to step S5; If the condition is not met, proceed to step S2; S5, end the ICS algorithm and switch to the IP&O algorithm for local optimization. Update the position using the position update formula. The specific method is as follows: The PV system is sampled, and the output power before and after the disturbance is compared and calculated for optimization. The IPO algorithm uses a variable step size for position update, and the update formula is as follows: In the formula, Step size; Step scaling factor; , The output voltage and current of the PV system; and These represent the power values ​​before and after the i-th disturbance, respectively. This represents the power increment before and after the disturbance; if after the disturbance... If positive, the next perturbation will be in the same direction: if the perturbation is positive... If the result is negative, the next perturbation will be made in the opposite direction; S6, determine if the restart conditions are met: If the conditions are met, the algorithm is restarted, and the process jumps to step S1. If the conditions are not met, proceed to step S5 to continue local optimization of IP&O. S7, output the current optimal power value, and end the algorithm.

2. The intelligent algorithm for photovoltaic maximum power point tracking based on ICS-IP&O according to claim 1, characterized in that: The specific method for step S1 is as follows: S101, initialize the relevant parameters of the ICS algorithm, including: voltage position U, duty cycle D, power P, optimal voltage U_best and optimal power P_best; S102, calculate the output power value corresponding to the initial position.

3. The intelligent algorithm for photovoltaic maximum power point tracking based on ICS-IP&O according to claim 1, characterized in that: The specific method in step S2 is as follows: The initial data obtained in step S1 is iteratively updated using adaptive Lévy flight, as shown in the following formula: In the formula, yes Lévy The flight step size coefficient is adaptively nonlinearly decreasing; symbol Dot product; It is the position of the i-th bird's nest in the t-th iteration; This represents the optimal parasitic nest in the current iteration; ; and They all follow a normal distribution. ; .

4. The intelligent algorithm for photovoltaic maximum power point tracking based on ICS-IP&O according to claim 1, characterized in that: In step S3, the method for calculating the switching probability and generating a random number is as follows: The number of available nests for parasitism in the ICS algorithm is fixed, while the host bird will have... The host bird has a certain probability of discovering the parasitic egg; once discovered, it will either destroy the egg or choose to rebuild its nest, abandoning the old one; the specific methods are as follows: S301, generate random numbers for each bird's nest. ; S302, perform random probability analysis on each bird's nest. The value is calculated using the following formula: In the formula, T represents the total number of iterations; t is the current iteration number; S303, for each bird's nest and Compare: If Return to step S2. Lévy Flight position update, if Then, the position is updated through random walk.

5. The intelligent algorithm for photovoltaic maximum power point tracking based on ICS-IP&O according to claim 1, characterized in that: In step S4, the ICS algorithm is used for global optimization in the early stage, and the corresponding position and power values ​​are calculated. The best and worst positions calculated at the moment are compared. When the switching conditions are met, the algorithm is switched. In the later stage, the algorithm is switched to IPO algorithm for local optimization.

6. The intelligent algorithm for photovoltaic maximum power point tracking based on ICS-IP&O according to claim 1, characterized in that: In step S6, natural environmental factors can cause changes in lighting conditions, which will also lead to changes in PV output power. The change is calculated, and when it meets the algorithm restart condition, the algorithm is initialized and the ICS algorithm is used for optimization again. The restart calculation formula is as follows: In the formula, This indicates the changed PV output power. This represents the PV output power before the change.

Images