Photovoltaic MPPT control method based on improved mayfly algorithm

Abstract

The application discloses a photovoltaic MPPT control method based on an improved mayfly algorithm, and proposes a photovoltaic MPPT method based on the improved mayfly algorithm; according to a comparison result of a current change rate and a threshold value, a P-V curve of a photovoltaic array under a current illumination condition is divided into a single-peak curve or a multi-peak curve, and different improved mayfly search strategies are adopted for the single-peak or multi-peak curve, so that higher tracking timeliness and accuracy are achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of photovoltaic power generation, and particularly relates to a photovoltaic MPPT control method based on an improved mayfly algorithm. BACKGROUND

[0002] In the photovoltaic power generation technology, improving the power generation efficiency is a core problem, and the output characteristics of the solar cell are affected by the light intensity. The maximum power point tracking (MPPT) is one of the key technologies for improving the power generation efficiency of the photovoltaic system.

[0003] The operation condition of the photovoltaic array is more complex, and the full-time operation condition includes the uniform irradiance condition and the local shading condition. Under the uniform irradiance condition, the power-voltage characteristic (P-V) of the photovoltaic array has a single-peak characteristic, and only one maximum power point exists. The traditional methods such as the perturbation and observation method and the conductance increment method can be used to realize the MPPT control. However, under the local shading condition, multiple extreme points exist in the P-V curve of the photovoltaic array, including one global maximum power point and multiple local maximum power points. In this case, the traditional MPPT method is difficult to find the global maximum power point, and is easy to fall into one of the local maximum power points, resulting in the power generation power loss of the photovoltaic array. At present, the meta-heuristic algorithm is mainly used to realize the MPPT control of the photovoltaic array under the local shading condition, and the particle swarm optimization (PSO) algorithm is most widely used. In addition, there are the differential evolution algorithm, the firefly algorithm, the grey wolf optimization algorithm, the dragonfly algorithm, the pollination algorithm and the artificial fish swarm algorithm. Under the local shading condition, the meta-heuristic algorithm performs random optimization on the P-V curve, and although the global maximum power point can be accurately searched, the randomness of the search process is easy to cause the algorithm power consumption, which not only reduces the optimization speed, but also causes the power fluctuation. Under the uniform irradiance condition, the P-V curve of the photovoltaic array has a single-peak characteristic. For the tracking of the single-peak curve, the random search characteristic of the meta-heuristic algorithm will cause the algorithm power waste to be aggravated, and the optimization efficiency is even lower than that of the classical MPPT algorithm. Therefore, the above problems need to be improved. SUMMARY

[0004] The present application provides a photovoltaic MPPT control method based on an improved mayfly algorithm, and aims to solve the problem that the photovoltaic array is difficult to accurately track the multiple-peak global maximum power point under the local shading condition and quickly complete the single-peak maximum power point tracking under the uniform irradiance condition. The photovoltaic MPPT method based on the improved mayfly algorithm is proposed. The algorithm structure of the control method is simple and easy to realize, can effectively judge the multiple-peak-single-peak condition of the P-V curve of the photovoltaic array, and has higher tracking timeliness and accuracy in the single-peak MPPT scene and the multiple-peak MPPT scene.

[0005] To solve the above technical problems, the technical scheme adopted by the present application is:

[0006] A photovoltaic MPPT control method based on an improved mayfly algorithm, the steps are:

[0007] Step1, a mathematical model of a photovoltaic module is established, and n photovoltaic modules are connected in series to obtain a photovoltaic array, and the open-circuit voltage V oc_mod of the photovoltaic module and the open-circuit voltage V oc_array of the photovoltaic array are obtained respectively.

[0008] Step2, select n values in the interval [0, V oc_array ] as the initial positions of the mayfly population, the mayfly individual represents the reference voltage in the PI control of the photovoltaic system, and the selection formula of the initial position mayfly_n of the mayfly population is:

[0009]

[0010] Step3, the output current of the photovoltaic system under the control of each initial reference voltage is obtained, and the maximum current change rate ε is calculated as:

[0011]

[0012] Wherein, I max is the maximum output current of the photovoltaic system under the control of the initial reference voltage, and I min is the minimum output current of the photovoltaic system under the control of the initial reference voltage.

[0013] Step4, compare ε with single-peak-multiple-peak threshold ε max , to determine whether the P-V curve of the photovoltaic array under the current illumination condition is a single-peak curve or a multiple-peak curve.

[0014] Step5, according to the judgment of single-peak-multiple-peak, different improved mayfly search strategies are adopted, in the multiple-peak case, switch to the last elimination strategy to carry out global maximum power point tracking GMPPT, and in the single-peak case, switch to the bisecting search strategy to carry out local maximum power point tracking LMPPT.

[0015] Step6, update the position of the mayfly population, and the position update formula is:

[0016] X i,t+1 = X i,t + v i,t+1

[0017] Wherein, X i,t+1 and X i,t+1 are the positions of the i-th mayfly individual at t+1 and t respectively, and v i,t+1is the velocity of the i th mayfly at time t + 1, the mayfly population is divided into male and female, the mayfly individual X i,t is divided into male individual m i,t and female individual f i,t , the velocity update formula of the male individual is:

[0018]

[0019] wherein v i,t is the velocity of the male mayfly i at time t, g is the gravity coefficient, and α 1 and α 2 are positive attraction constants, respectively correcting the individual and social contribution, p best,i represents the historical best reference voltage of the i th male mayfly, g best represents the global best reference voltage in the male mayfly population, β is the visibility coefficient, r p is the Cartesian distance between m i and p best , r g is the Cartesian distance between m i and g best ;

[0020] The optimal individual in the male mayfly population will move up and down in order to avoid the algorithm falling into local optimum, and the velocity update formula of the optimal male individual is:

[0021] v i,t+1 = v i,t + d r

[0022] wherein d is the dance coefficient, and r is a random number in [-1, 1].

[0023] The female individual approaches the male individual, and the velocity update formula is:

[0024]

[0025] wherein r mf is the Cartesian distance between m i and f i , λ is the random walk coefficient, and P(·) is the output power corresponding to the mayfly;

[0026] Step 7, output the reference voltage represented by the position of each mayfly, obtain the output power of the photovoltaic system under the control of each reference voltage, and sort the male mayfly population and the female mayfly population respectively, select a pair of parents from the male and the female respectively, and generate two offspring, and the generation formula is:

[0027]

[0028] wherein m and f are the father and the mother respectively, m offspring1 and moffspring2 L is a random value in [0, 1];

[0029] Step8, calculate the corresponding maximum current rate of change in the current generation of mayfly, judge the multi-peak-single peak situation, that is, whether the mayfly optimization algorithm meets the local convergence interval in the later stage, if yes, the algorithm enters LMPPT; otherwise, continue to perform GMPPT;

[0030] Step9, judge whether the termination condition is met, if yes, stop iteration and stabilize the photovoltaic system at the reference voltage corresponding to the optimal mayfly individual; otherwise, return to Step6 and continue iteration;

[0031] Step10, monitor whether the output power of the photovoltaic system has a large mutation, when the output power change rate is greater than the set threshold, return to Step1 and restart the control algorithm.

[0032] The single diode model of the photovoltaic module is used when establishing the mathematical model of the photovoltaic module in Step1 above:

[0033]

[0034] Wherein, I L is the output current, U L is the output voltage, I ph_ref is the photo-generated current, I 0_ref is the saturation current of the photovoltaic module without light, R s is the series resistance of the photovoltaic module, R sh is the shunt resistance of the photovoltaic module, q is the charge constant (1.602×10 -19 C), A is a constant factor (A value is 1 when the positive bias voltage is large, and 2 when the positive bias voltage is small, generally 1.3), K is the Boltzmann constant (1.38×10 -23 J / K), and T is the absolute temperature value of the photovoltaic module.

[0035] The determination formula for determining whether the output power has a mutation in Step10 above is:

[0036]

[0037] Wherein, P pv (k) and P pv (k-1) are the power measurement values in the kth and k-1th sampling periods respectively, and ΔP th is the power change rate threshold value caused by rapid irradiance change.

[0038] The last elimination strategy in Step5 above includes:

[0039] In the population breeding process, the individuals with lower fitness than the set value of the population are eliminated, the population number is kept unchanged, and the population elimination mechanism is:

[0040]

[0041] Wherein, k is the iteration number, m offspring is the offspring individual, mayfly k_worst is the worst individual in the parent generation.

[0042] The bisection search strategy in the above Step5 is executed, including:

[0043] Only two individuals farthest apart are kept in the mayfly population as the two endpoints of the bisection search interval, V a,j and V b,j are the left and right endpoints of the interval respectively, and the expression of the two endpoints of the first generation bisection search interval is:

[0044]

[0045] The calculation formula of the midpoint V c,j of the jth bisection search interval is:

[0046]

[0047] The power of the photovoltaic system corresponding to the V c,j point is recorded as P1, the power of the photovoltaic system corresponding to the V c,j +ΔV point is recorded as P2, then the power difference between the V c,j and V c,j +△V points is ΔP=P2-P1, wherein ΔV is the perturbation voltage increment; when ΔP>0, the update formula of the bisection search interval of the reference voltage is:

[0048]

[0049] When ΔP<0, the update formula of the bisection search interval of the reference voltage is:

[0050]

[0051] The termination condition in the above Step9 includes whether the algorithm reaches the maximum iteration number T max ; whether the search interval of the algorithm is less than 1% V oc_array .

[0052] The application provides a photovoltaic MPPT control method based on an improved mayfly algorithm, according to the comparison result of the current change rate and a threshold value, the P-V curve of a photovoltaic array under current illumination is divided into a single-peak curve or a multi-peak curve, and different improved mayfly search strategies are adopted for the single-peak or multi-peak curve, so that the tracking timeliness and accuracy are higher. BRIEF DESCRIPTION OF DRAWINGS

[0053] The application is further described below in combination with the drawings and examples:

[0054] Figure 1 FIG. 1 is a schematic diagram of a photovoltaic MPPT control system based on a Boost circuit in an embodiment of the application;

[0055] Figure 2 FIG. 2 is a flowchart of a photovoltaic MPPT control method based on an improved mayfly algorithm in the application;

[0056] Figure 3 FIG. 3 is a P-V and I-V output characteristic curve diagram of a photovoltaic array under different illumination conditions in an embodiment of the application;

[0057] Figure 4 FIG. 4 (a) is an output power, output voltage and output current curve diagram of a photovoltaic array under UIC1 based on a particle swarm optimization algorithm for MPPT control in an embodiment of the application;

[0058] FIG. 4 (b) is an output power, output voltage and output current curve diagram of a photovoltaic array under UIC1 based on an improved mayfly algorithm for MPPT control in an embodiment of the application;

[0059] FIG. 4 (c) is an output power, output voltage and output current curve diagram of a photovoltaic array under PSC based on a particle swarm optimization algorithm for MPPT control in an embodiment of the application;

[0060] FIG. 4 (d) is an output power, output voltage and output current curve diagram of a photovoltaic array under PSC based on an improved mayfly algorithm for MPPT control in an embodiment of the application;

[0061] FIG. 4 (e) is an output power curve diagram of a photovoltaic array under UIC1, PSC and UIC2 dynamic mutation based on an improved mayfly algorithm for MPPT control in an embodiment of the application. DETAILED DESCRIPTION

[0062] The technical scheme of the application is described in detail below in combination with the drawings and examples.

[0063] A photovoltaic MPPT control method based on an improved mayfly algorithm, comprising the following steps: Step 1, establishing a mathematical model of a photovoltaic component, and obtaining a photovoltaic array by connecting n photovoltaic components in series, respectively acquiring the open circuit voltage V oc_modand the open-circuit voltage V of the photovoltaic array oc_array ;

[0064] Step 2, select n values in the interval [0, V oc_array ] as the initial positions of the mayfly population, and the mayfly individuals represent the reference voltage in the photovoltaic system PI control. The selection formula of the initial position of the mayfly population mayfly_n is:

[0065]

[0066] Step 3, obtain the output current of the photovoltaic system under the control of each initial reference voltage, and calculate the maximum current change rate ε as:

[0067]

[0068] where I max is the maximum output current of the photovoltaic system under the initial reference voltage control, and I min is the minimum output current of the photovoltaic system under the initial reference voltage control.

[0069] Step 4, compare ε with the unimodal-multimodal threshold value ε max , and determine whether the P-V curve of the photovoltaic array under the current illumination condition is a unimodal curve or a multimodal curve.

[0070] Step 5, according to the unimodal-multimodal determination, different improved mayfly search strategies are adopted. In the multimodal case, switch to the last elimination strategy for global maximum power point tracking GMPPT, and in the unimodal case, switch to the bisecting search strategy for local maximum power point tracking LMPPT.

[0071] Step 6, update the position of the mayfly population, and the position update formula is:

[0072] X i,t+1 = X i,t + v i,t+1

[0073] where X i,t+1 and X i,t+1 are the positions of the i-th mayfly individual at t+1 and t respectively, v i,t+1 is the speed of the i-th mayfly individual at t+1, and the mayfly population is divided into male and female. The mayfly individual X i,t is divided into male individual m i,t and female individual f i,t , and the speed update formula of the male individual is:

[0074]

[0075] where vi,t is the velocity of the male mayfly at time t, g is the gravity coefficient, a1 and a2 are positive attraction constants that modify the individual and social contributions, respectively, p best,i represents the historical best reference voltage of the ith male mayfly, g best represents the global best reference voltage in the male mayfly population, β is the visibility coefficient, r p is the Cartesian distance between m i and p best ; r g is the Cartesian distance between m i and g best ;

[0076] The optimal individual in the male mayfly population will move up and down in order to avoid the algorithm from falling into a local optimum, and the velocity update formula of the optimal male individual is:

[0077] v i,t+1 = v i,t + dr

[0078] where d is the dance coefficient and r is a random number in [-1, 1].

[0079] The female individual moves towards the male individual, and the velocity update formula is:

[0080]

[0081] where r mf is the Cartesian distance between m i and f i , λ is the random walk coefficient, and P(·) is the output power corresponding to the mayfly;

[0082] Step 7, output the reference voltage represented by the position of each mayfly, obtain the output power of the photovoltaic system under the control of each reference voltage, and sort the male mayfly population and the female mayfly population, respectively, select a pair of parents from the male and the female, respectively, and generate two offspring, and the generation formula is:

[0083]

[0084] where m and f are the father and the mother, respectively, m offspring1 and m offspring2 are the two offspring, and L is a random value in [0, 1];

[0085] Step 8, calculate the corresponding maximum current change rate of the mayflies in the current generation, judge the multi-peak-single-peak situation, that is, whether the mayfly algorithm meets the local convergence interval in the later stage of optimization, if it meets, the algorithm enters LMPPT; otherwise, continue to perform GMPPT;

[0086] Step 9, judging whether the termination condition is met, if yes, stopping iteration and stabilizing the photovoltaic system at the reference voltage corresponding to the optimal mayfly individual; otherwise, returning to Step 6 and continuing iteration;

[0087] Step 10, monitoring whether the output power of the photovoltaic system has a large mutation, when the output power change rate is greater than a set threshold, returning to Step 1 and restarting the control algorithm.

[0088] The single-diode model of the photovoltaic module is used in Step 1 to establish the mathematical model of the photovoltaic module as follows:

[0089]

[0090] where I L is the output current, U L is the output voltage, I ph_ref is the photo-generated current, I 0_ref is the saturation current of the photovoltaic module without light, R s is the series resistance of the photovoltaic module, R sh is the shunt resistance of the photovoltaic module, q is the charge constant (1.602×10 -19 C), A is a constant factor (A is 1 when the positive bias voltage is large, and A is 2 when the positive bias voltage is small, generally taking 1.3), K is the Boltzmann constant (1.38×10 -23 J / K), and T is the absolute temperature value of the photovoltaic module.

[0091] The determination formula for determining whether the output power has a mutation in Step 10 is as follows:

[0092]

[0093] where P pv (k) and P pv (k-1) are the power measurement values in the kth and (k-1)th sampling periods, respectively, and ΔP th is the power change rate threshold value caused by rapid irradiance change.

[0094] The last elimination strategy is executed in Step 5, including:

[0095] During the population breeding process, individuals with a fitness lower than a set value of the population are eliminated, and the population size is kept unchanged. The population elimination mechanism is as follows:

[0096]

[0097] where k is the iteration number, m offspring is the offspring individual, and mayfly k_worst is the worst individual in the parent population.

[0098] The bisection search strategy in Step 5 is executed, including:

[0099] Only two individuals farthest apart in the mayfly population are reserved as the two endpoints of the bisection search interval, V a,j and V b,j are the left and right endpoints of the interval respectively, and the expression of the two endpoints of the first generation bisection search interval is:

[0100]

[0101] The calculation formula of the midpoint V c,j of the jth bisection search interval is:

[0102]

[0103] The power of the photovoltaic system corresponding to V c,j is recorded as P1, and the power corresponding to V c,j +ΔV is recorded as P2, then the power difference between V c,j and V c,j +△V is ΔP=P2-P1, where ΔV is the perturbation voltage increment; when ΔP>0, the update formula of the bisection search interval of the reference voltage is:

[0104]

[0105] When ΔP<0, the update formula of the bisection search interval of the reference voltage is:

[0106]

[0107] The termination condition in Step 9 includes whether the algorithm reaches the maximum number of iterations T max ; whether the search interval is less than 1% V oc_array .

[0108] Embodiment:

[0109] Figure 1 The circuit structure of the photovoltaic maximum power point tracking system based on the Boost circuit in the embodiment has a photovoltaic array, an MPPT controller based on PI modulation, and a PWM generator. The photovoltaic array is composed of 4×1 photovoltaic components connected in series. The MPPT controller adjusts the output reference voltage according to the sampled photovoltaic array output power, thereby realizing maximum power point tracking.

[0110] The MPPT controller has a memory and a processor, and the memory stores a computer program executable by the processor. When the computer program is executed, the steps of the photovoltaic MPPT method based on the improved mayfly algorithm are implemented.

[0111] Considering that local shadow blocking makes the output power of the photovoltaic array unable to reach the theoretical maximum power point, because the conventional MPPT algorithm is prone to failure when tracking the maximum power point, the tracking efficiency is reduced; in order to further verify that the output characteristic curve of the photovoltaic array under different conditions presents a multi-peak phenomenon, it is necessary to establish a simulation model of the photovoltaic array, set different light conditions, and obtain the P-V and I-V characteristic curves under different light conditions, to prepare for analyzing the MPPT algorithm; Table 1 is the different environmental parameters set in the embodiment of the application, which is divided into uniform light conditions (UIC) and partial shading conditions (PSC):

[0112] Table 1 Photovoltaic array light working condition

[0113]

[0114] Figure 3 In order to obtain the P-V and I-V characteristic curves of the photovoltaic array under UIC1 and PSC in the embodiment, the situation of the photovoltaic array under shadow blocking is more clearly described; in the same embodiment, when the photovoltaic array is under local shading conditions, the I-V curve is approximately step-shaped, and the P-V curve produces multiple local peaks, so that the conventional MPPT algorithm is more prone to fall into local solution; when the photovoltaic array is under uniform light conditions, the P-V curve is a single-peak curve, and the original mayfly algorithm will lead to a too long convergence time, or even lower than the conventional MPPT algorithm.

[0115] Figure 2 The improved photovoltaic MPPT control method based on the improved mayfly algorithm described in the application comprises the following steps:

[0116] Step 1: mayfly algorithm parameter initialization, 4 values are selected in the interval [0, V oc_array ] as the initial positions of mayfly population, mayfly individuals represent the reference voltage in PI control of the photovoltaic system, and the initial positions of mayfly_1, mayfly_2, mayfly_3 and mayfly_4 are 0.7×V oc_mod , 1.5×V oc_mod , 2.3×V oc_mod and 3.1×V oc_mod , respectively;

[0117] Step 2: after outputting the reference voltage represented by the initial positions of mayflies, the output currents of the photovoltaic system under the control of each initial reference voltage are sampled, and the maximum current change rate ε in the initial mayfly is calculated:

[0118]

[0119] Wherein, I maxI min is the minimum output current of the photovoltaic system under the initial reference voltage control.

[0120] Step 3: compare the calculated ε with the unimodal-multimodal threshold ε max , when ε < ε max , the P-V curve of the photovoltaic array under the current illumination condition is a unimodal curve, and when ε > ε max , the P-V curve of the photovoltaic array under the current illumination condition is a multimodal curve.

[0121] Step 4: according to the unimodal-multimodal judgment, different improved mayfly search strategies are adopted, and in the multimodal case, the last elimination strategy is switched to, and in the unimodal case, the bisecting search strategy is switched to.

[0122] Step 5: position update is performed on the mayfly population, and the position update formula is:

[0123] X i,t+1 = X i,t + v i,t+1

[0124] where X i,t+1 and X i,t+1 are the positions of the i-th mayfly individual at t+1 and t respectively, v i,t+1 is the speed of the i-th mayfly individual at t+1, the mayfly population is divided into male and female, and the mayfly individual X i,t is divided into male individual m i,t and female individual f i,t , and the speed update formula of the male individual is:

[0125]

[0126] where v i,t is the speed of the male mayfly i at t, g is the gravity coefficient, α1 and α2 are positive attraction constants, which correct the individual and social contributions respectively, p best,i represents the historical best reference voltage of the i-th male mayfly, g best represents the global best reference voltage in the male mayfly population, β is the visibility coefficient, r p is the Cartesian distance between m i and p best , and r g is the Cartesian distance between m i and g best .

[0127] The optimal individual in the male mayfly population moves up and down to avoid the algorithm falling into local optimum, and the speed update formula of the optimal male individual is:

[0128] v i,t+1 = v i,t + dr

[0129] where d is the dance coefficient and r is a random number in [-1, 1].

[0130] The female individual approaches the male individual, and the speed update formula is:

[0131]

[0132] where r mf is the Cartesian distance between m i and f i , λ is the random walk coefficient, and P(·) is the output power corresponding to the mayfly;

[0133] Step 6: output the reference voltage represented by each mayfly position, obtain the output power of the photovoltaic system under the control of each reference voltage, and sort the male mayfly group and the female mayfly group respectively, select a pair of parents from the male and the female respectively, and generate two offspring, and the generation formula is:

[0134]

[0135] where m and f are the father and mother respectively, m offspring1 and m offspring2 are the two offspring, and L is a random value in [0, 1];

[0136] Step 7: calculate the corresponding maximum current change rate of the mayfly in this generation, judge the multi-peak-single-peak situation, that is, whether the mayfly algorithm meets the local convergence interval in the later optimization, if yes, the algorithm is based on the last elimination strategy to perform LMPPT; otherwise, continue to perform GMPPT based on the bisection search strategy;

[0137] The improved mayfly algorithm based on the last elimination strategy is improved in that after generating offspring individuals, the population size will increase, and the individual with the lowest fitness value will be eliminated to keep the size of the entire population unchanged, and the elimination mechanism is:

[0138]

[0139] where k is the iteration number of the algorithm, m offspring is the offspring individual, and mayfly k_worst is the worst individual in the parent generation;

[0140] The improved mayfly algorithm based on the bisection search strategy is improved in that when performing single-peak search, the population size is fixed to two mayflies as the two endpoints of the bisection search interval, and the optimal value is found by continuously dividing the interval, V a,j and Vb,j respectively, the expression of the two endpoints of the first generation bisection search interval is calculated as:

[0141]

[0142] The midpoint V c,j of the jth generation bisection search interval is calculated according to the search interval.

[0143]

[0144] The output V c,j corresponds to the reference voltage, and the power of the photovoltaic system is sampled, denoted as P1, and the output V c,j + ΔV corresponds to the reference voltage, and the power of the photovoltaic system is sampled, denoted as P2, and the V c,j corresponds to the reference voltage, and the power of the photovoltaic system is sampled, denoted as P1, and the output V c,j + ΔV corresponds to the power difference ΔP = P2 - P1, where ΔV is the perturbation voltage increment; when ΔP > 0, the update formula of the bisection search interval of the reference voltage is:

[0145]

[0146] When ΔP < 0, the update formula of the bisection search interval of the reference voltage is:

[0147]

[0148] Step 8: Determine whether the control algorithm meets the termination condition, if it meets, stop iteration and stabilize the photovoltaic system at the reference voltage corresponding to the optimal mayfly individual; otherwise, return to step 5 and continue iteration;

[0149] The termination condition has two, the first is to determine whether the algorithm reaches the maximum iteration number T max ; The second is to determine whether the updated bisection search interval of the algorithm is less than 1% V oc_array ;

[0150] If any of the conditions is met, the algorithm stops iteration and outputs the historical optimal value of the reference voltage of the photovoltaic system; otherwise, return to step 5;

[0151] Step 9: Continuously detect the output power in the operation of the photovoltaic system, when the output power has a large mutation, the current photovoltaic system is no longer operated at the maximum power point, and the algorithm needs to be restarted to search for a new maximum power point, and the judgment formula of the light mutation is:

[0152]

[0153] Where P pv (k) and P pv(k-1) are power measurements in the kth and k-1th sampling periods, respectively, and ΔP th is a threshold of the power change rate caused by a sharp irradiance change.

[0154] The present application takes photovoltaic system maximum power point tracking control as the research object; aiming at different output characteristics of photovoltaic array under different light conditions, a MPPT method based on improved mayfly algorithm is designed, which can effectively judge the current light condition of photovoltaic array, and sample different strategies to complete maximum power point tracking according to different light conditions. First, according to the algorithm characteristics, the control variable of the algorithm is selected as the reference voltage, and the objective function is the output power of the photovoltaic array. Secondly, the MATLAB / Simulink simulation model is built, and the circuit parameters are set; for the characteristics that the actual light condition often changes, the simulation of light intensity mutation is designed, and the termination and restart mechanism is added in the algorithm. The improved mayfly algorithm is encapsulated as a control module in the form of Matlab function, and the current and voltage are taken as input, and the reference voltage corresponding to the maximum power is taken as output. The particle swarm algorithm (PSO) is also simulated, and compared with the improved mayfly algorithm.

[0155] Figure 4 (a)~ (c) are the output characteristics curve simulation results of the photovoltaic array using different MPPT algorithms in the UIC and PSC embodiments; from Figure 4 (e) shows the output characteristics curve simulation results of the photovoltaic array using different MPPT algorithms in the UIC and PSC embodiments; from Figure 4 (a) and Figure 4 (b) are the output waveforms of the PSO algorithm and the improved mayfly algorithm under UIC, respectively, and it can be known that the PSO algorithm performs random search in the optimization interval, and the convergence time is relatively long, about 0.69s, the average value of the tracked power is 238.3W, and the optimization accuracy is 99.29%.

[0156] The improved mayfly algorithm proposed in the present application realizes the detection of shadow in 0.04s, and the convergence time is about 0.19s, which is only 27.5% of the convergence time of the PSO algorithm, the average value of the tracked power is 239.4W, and the optimization accuracy is 99.81%, which indicates that the improved mayfly algorithm has faster convergence speed and higher optimization accuracy than the PSO algorithm under UIC. Figure 4 (c) and Figure 4(d) The output waveform of PSC under PSO algorithm and improved mayfly algorithm respectively, the convergence time of PSC under PSO algorithm is about 0.66 s, the average value of the tracked power is 94.67 W, and the optimization accuracy is 99.80%; while IMA detects the shadow in 0.04 s, and the convergence time of improved mayfly algorithm is about 0.36 s, which is about 54.5% of the convergence time of PSO algorithm, the average value of the tracked power is 94.84 W, and the optimization accuracy is 99.98%; which shows that the improved mayfly algorithm has faster convergence speed and higher optimization accuracy than PSO algorithm under PSC. In addition, the search range of the improved mayfly algorithm gradually reduces, which further reduces the power oscillation range near the MPP and improves the power generation efficiency of the photovoltaic array; the output power waveform of the improved mayfly algorithm under the sudden change of light shows that the restart algorithm can effectively complete the system restart and re-track the new maximum power point. Figure 4

[0157] The simulation results show that under the standard condition and the local shadow condition, the improved mayfly algorithm has faster speed in tracking the global maximum power point and smaller power fluctuation in the tracking process than the particle swarm algorithm. Under the condition of dynamic mutation of light, the power waveform also verifies that the MPPT control method based on the improved mayfly algorithm has better dynamic stability performance and tracking effect, which effectively improves the photovoltaic power generation efficiency.​

Claims

1. A photovoltaic MPPT control method based on an improved mayfly algorithm, characterized in that the steps are as follows: for: Step 1: Establish a mathematical model of the photovoltaic module, and connect n photovoltaic modules in series to obtain a photovoltaic array. Obtain the open-circuit voltage V of each photovoltaic module. oc_mod and the open-circuit voltage V of the photovoltaic array oc_array ; Step 2: In the interval [0, V] oc_array [n values ​​are selected from the given information to represent the initial positions of the mayfly population. Each mayfly represents the reference voltage in the PI control of the photovoltaic system. The formula for selecting the initial position mayfly_i of the mayfly population is as follows:] Step 3: Obtain the output current of the photovoltaic system under each initial reference voltage control, and calculate the maximum current change rate ε: Among them, I max I represents the maximum output current of the photovoltaic system under initial reference voltage control. min This is the minimum output current of the photovoltaic system under initial reference voltage control. Step 4, compare ε with the single-peak / multi-peak threshold ε max By comparing the curves, it can be determined whether the PV curve of the photovoltaic array is a single-peak curve or a multi-peak curve under the current illumination conditions. Step 5: Based on the single-peak / multi-peak judgment, different improved mayfly search strategies are adopted. In the case of multi-peak, the last elimination strategy is switched to perform global maximum power point tracking (GMPPT), and in the case of single-peak, the bisection search strategy is switched to perform local maximum power point tracking (LMPPT). Step 6: Update the location of the mayfly population using the following formula: X i,t+1 =X i,t +v i,t+1 Among them, X i,t+1 and X i,t v represents the location of the i-th mayfly at time t+1 and time t, respectively. i,t+1 Let X be the velocity of the i-th mayfly at time t+1. The mayfly population is divided into males and females. i,t Classification of male individuals m i,t and female individuals f i,t The speed update formula for male individuals is: Among them, v i,t Let be the velocity of male mayfly i at time t, g be the gravitational coefficient, α1 and α2 be the positive attraction constants, respectively correcting for individual and social contributions, and p be the velocity of male mayfly i at time t. best,i G represents the historical best reference voltage for the i-th male mayfly. best The voltage represents the global optimal reference voltage within a male mayfly population, β is the visibility coefficient, and r p For m i,t With p best,t Cartesian distance between them, r g For m i,t With g best Cartesian distance between them; The optimal individual in the male mayfly population will move up and down to avoid the algorithm getting stuck in local optima. The velocity update formula for the optimal male individual is: v i,t+1 =v i,t +dr Where d is the dance coefficient and r is a random number in the range [-1, 1]; Female individuals move closer to male individuals, and the speed update formula is: Where, r mf For m i,t and f i,t The Cartesian distance between them, λ is the random walk coefficient, and P(·) is the output power corresponding to the mayfly; Step 7: Output the reference voltage represented by each mayfly position to obtain the output power of the photovoltaic system under the control of each reference voltage. Sort the male and female mayfly populations separately, select a parent pair from the males and females respectively, and produce two offspring. The formula for producing offspring is as follows: Where m and f are the father and mother, respectively, m offspring1 and m offspring2 There are two offspring, and L is a random value within [0,1]. Step 8: Calculate the maximum current change rate in this generation of mayflies and determine whether the multi-peak or single-peak case is satisfied in the later stage of the mayfly algorithm optimization. If it is satisfied, the algorithm enters LMPPT; otherwise, it continues to GMPPT. Step 9: Determine if the termination condition is met. If it is, stop the iteration and stabilize the photovoltaic system at the reference voltage corresponding to the optimal mayfly individual; otherwise, return to Step 6 and continue the iteration. Step 10: Monitor whether there is a significant change in the output power of the photovoltaic system. When the rate of change of output power exceeds the set threshold, return to Step 1 and restart the control algorithm.

2. The photovoltaic MPPT control method based on the improved mayfly algorithm according to claim 1, characterized in that, The criterion for determining whether the output power changes abruptly in Step 10 is as follows: Among them, P pv (k) and P pv (k-1) represent the power measurement values ​​in the kth and (k-1)th sampling periods, respectively, ΔP th It is the threshold of the power change rate caused by a sharp change in irradiance.

3. The photovoltaic MPPT control method based on the improved mayfly algorithm according to claim 2, characterized in that, Step 5 involves implementing a last-place elimination strategy, including: During population reproduction, individuals with fitness levels below the population setpoint are eliminated to maintain a constant population size. The population elimination mechanism is as follows: Where h is the number of iterations, m offspring For offspring individuals, mayfly h_worst It is the worst individual among its parents.

4. The photovoltaic MPPT control method based on the improved mayfly algorithm according to claim 3, characterized in that, Step 5 involves executing a bisection search strategy, including: In the mayfly population, only the two individuals furthest apart are retained as the two endpoints of the bisected search interval, V. a,j and V b,j Let $\begin{pmatrix}$ be the left and right endpoints of the interval, respectively. The expressions for the two endpoints of the first-generation bisection search interval are: V, the midpoint of the j-th bisection search interval c,j The formula for calculation is: V c,j The photovoltaic system power corresponding to point is denoted as P1, V c,j Let P2 be the photovoltaic system power corresponding to point +ΔV, then V c,j With V c,j The power difference corresponding to point +ΔV is ΔP = P2 - P1, where ΔV is the disturbance voltage increment; when ΔP > 0, the update formula for the bisection search interval of the reference voltage is: When ΔP < 0, the update formula for the bisection search interval of the reference voltage is:

5. The photovoltaic MPPT control method based on the improved mayfly algorithm according to claim 4, characterized in that, The termination condition in Step 9 includes: whether the algorithm has reached the maximum number of iterations T. max Is the algorithm search interval less than 1%V? oc_array .

Images