An improved hybrid global maximum power tracking control method based on particle swarm optimization and disturbance observation
By combining particle swarm optimization and perturbation observation methods, a hybrid global maximum power point tracking (MPPT) control method was developed, which solved the problem of multiple peak values under partial shading of photovoltaic panels, achieved fast and accurate maximum power point tracking, and improved the output efficiency of the photovoltaic array.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-09-27
- Publication Date
- 2026-03-24
AI Technical Summary
When photovoltaic panels are partially shaded, the traditional MPPT algorithm has difficulty effectively tracking the global maximum power point, resulting in multiple peaks in the power output curve of the photovoltaic array. Existing intelligent algorithms are complex and computationally intensive, while heuristic algorithms lack accuracy and efficiency.
By combining particle swarm optimization and perturbation observation methods, and by improving the initialization method and setting transition conditions, the number of particles is reduced, the accuracy of the algorithm is improved, and the algorithm is restarted to find the maximum power point when the environment changes.
It accelerates the algorithm's convergence speed, reduces the probability of getting stuck in local optima, improves the accuracy and efficiency of maximum power point tracking, and adapts to changes in the external environment.
Smart Images
Figure CN115599161B_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The application belongs to the field of photovoltaic power generation, and particularly relates to an improved hybrid global maximum power tracking control method based on particle swarm optimization and perturbation observation. BACKGROUND
[0002] With the increasing consumption of traditional energy, renewable energy is rapidly developing. As a clean renewable energy, solar energy is increasingly valued by countries around the world. With the strong support of governments around the world, the global solar photovoltaic industry has developed rapidly. Due to the influence of external environment (such as solar irradiance and temperature) on the output characteristics of photovoltaic panels, the maximum power point of photovoltaic panels will change with these external factors. Therefore, effective maximum power point tracking (MPPT) technology is of great significance to improve the efficiency of photovoltaic power generation systems.
[0003] Part of the photovoltaic panels in the photovoltaic array will appear local shading due to reasons such as shadow and dust, thereby changing external conditions such as light intensity, which will cause the photovoltaic array to have mismatched electrical characteristics, resulting in a multi-peak characteristic of the power output curve of the photovoltaic array. At this time, traditional MPPT algorithms, such as the perturbation and observation method (P&O), the incremental conductance method (IC), and the constant voltage tracking method (CV), are difficult to find the global maximum power point. Therefore, many intelligent algorithms are used to find the maximum power point, such as fuzzy logic control (FLC), neural network (ANN), genetic algorithm (GA), and various heuristic algorithms such as particle swarm optimization PSO, grey wolf algorithm GWO, firefly FA, improved cat swarm algorithm MCSO, moth-flame optimization MFO, cuckoo (dove) algorithm CS, gravity search algorithm GSA, and ant colony optimization ACO. Most intelligent algorithms are relatively complex, require a large amount of data set, and require a large amount of calculation, while heuristic algorithms are relatively simple, have high tracking accuracy and efficiency, and the particle swarm algorithm is a good MPPT algorithm with simple implementation, uncomplicated calculation, fast tracking speed, and high accuracy.
[0004] The convergence speed and accuracy of a single particle swarm algorithm are not obvious, and a better method is needed to improve the efficiency and accuracy of the algorithm. SUMMARY
[0005] The application aims to provide an improved hybrid global maximum power tracking control method based on particle swarm optimization and perturbation observation to solve the problem of tracking the maximum power point of photovoltaic components under the condition of multiple peaks caused by local shading.
[0006] The technical solution for achieving the application is as follows: an improved hybrid global maximum power tracking control method based on particle swarm optimization and perturbation observation, comprising the following steps:
[0007] Step 1, find the peak value where the global maximum power point is located by using a particle swarm algorithm;
[0008] Step 2, track the global maximum power point by using a perturbation observation method;
[0009] Compared with the prior art, the present application has the following advantages:
[0010] The present application uses a method combining a particle swarm algorithm and a perturbation observation method to accelerate the convergence speed of the algorithm; the present application improves the initialization method of the particle swarm algorithm and uses fewer particles, so that the probability of the algorithm falling into a local optimal point is reduced, and the accuracy of the algorithm is improved; the present application proposes a transition condition from the particle swarm algorithm to the perturbation observation method, which can reduce the iteration number of the particle swarm algorithm and faster reach the maximum power point; the present application sets multiple restart conditions for the change of the environment, so that the algorithm can find the maximum power point again in the case of environmental change. BRIEF DESCRIPTION OF DRAWINGS
[0011] Figure 1 is a boost circuit application example diagram of the present application;
[0012] Figure 2 is a P-V output characteristic curve of a photovoltaic array;
[0013] Figure 3 is a whole flow chart of the present application;
[0014] Figure 4 is a probability curve of the initialized particles;
[0015] Figure 5 is a flow chart of the perturbation observation method of the present application;
[0016] Figure 6 is a flow chart of the restart algorithm of the present application;
[0017] Fig. 7(a) is a simulation comparison diagram of the present application and the conventional particle swarm algorithm in the first case;
[0018] Fig. 7(b) is a simulation comparison diagram of the present application and the conventional particle swarm algorithm in the second case;
[0019] Figure 8 is a comparison simulation diagram of the reinitialization algorithm of the present application; DETAILED DESCRIPTION
[0020] The present application proposes an improved hybrid global maximum power tracking control method based on particle swarm optimization and perturbation observation, which comprises the following steps:
[0021] Step 1, find the peak value where the global maximum power point is located by using a particle swarm algorithm;
[0022] Step 2, tracking the global maximum power point by using the perturbation and observation method;
[0023] Further, step 1 is specifically:
[0024] (1-1) initialize the particle swarm algorithm, the particles of the particle swarm algorithm are equivalent to the duty cycle of the boost circuit, the number of particles is determined according to the number of photovoltaic panels, and the particles are distributed according to the normal distribution probability curve.
[0025] (1-2) measure the current voltage and current value;
[0026] (1-3) calculate the current power;
[0027] (1-4) judge whether the maximum power point is found, if yes, execute step 2, if not, execute the next step.
[0028] (1-5) iterate according to the particle speed position formula through the particle swarm algorithm;
[0029] (1-6) output the current duty cycle and return to (1-2).
[0030] Further, the particle swarm algorithm parameters include inertia weight, individual cognitive coefficient, social cognitive coefficient, individual extreme value, global extreme value and transition condition parameter.
[0031] Further, step 2 is specifically:
[0032] (2-1) take the global optimal particle of the particle swarm algorithm as the initial duty cycle, and give the duty cycle perturbation. According to the perturbation and observation method, change the duty cycle to track the global maximum power point.
[0033] (2-2) when the environment changes and the set condition is met, return to step 1, if not, return to (2-1).
[0034] The application will be described in detail below in combination with the drawings and specific embodiments.
[0035] Embodiment:
[0036] Taking a photovoltaic power generation system based on a boost circuit as an example, the structure diagram of the boost circuit is shown in Figure 1 , the photovoltaic panel is used as the input, four photovoltaic panels are used in series as the input in the embodiment, the inductor L, the capacitors C1 and C2, the diode D, the switch tube Q and the load R are used as the main circuit.
[0037] Figure 2 The voltage-power curve diagram of the photovoltaic array under the condition of local shading and multiple peak output characteristics. The light intensities of the four photovoltaic panels under condition 1 are 1000 W / m 2, 1000W / m 2 , 800W / m 2 , 500W / m 2 ; the light intensity of the four photovoltaic panels of condition 2 is 1000W / m 2 , 1000W / m 2 , 900W / m 2 , 700W / m 2 ; the light intensity of the four photovoltaic panels of condition 3 is 1000W / m 2 , 1000W / m 2 , 1000W / m 2 , 1000W / m 2 .
[0038] As shown in Fig. Figure 3 , it is the overall algorithm flow chart, the specific implementation steps are as follows:
[0039] Step 1, using particle swarm algorithm to find the global maximum power point of the peak value;
[0040] (1-1) initialization particle, as shown in Fig. Figure 4 , because there are four photovoltaic panels, in order to reduce the number of particles as much as possible, and in order to make each peak value initial moment as much as possible There is a particle, so use four particles, the initial position of each particle probability curve is shown in the figure, the first particle is a random number of normal distribution probability curve with 0.2 as the average and 0.2 as the standard deviation of 3, the remaining three particles are calculated in the same way. Do this can reduce the number of particles under the premise, to ensure that the particles can achieve global search, avoid the situation that all particles cannot enter the peak value where the global optimal point in the initialization stage, leading to fall into local optimal peak value;
[0041] (1-2) measure the current voltage and current value;
[0042] (1-3) calculate the current power;
[0043] (1-4) judge whether the maximum power point is found, the judgment condition is: |D i -D i-1 |<M(i=1,...,4);max{D1,...,D4}-min{D1,...,D4}<Q, in this embodiment, M is 0.04, Q is 0.06, Di is the current duty ratio of the ith particle, Di-1 is the last duty ratio of the ith particle. If the condition is met, step 2 is executed; if not, the next step is executed.
[0044] (1-5) through the particle swarm algorithm, according to the particle velocity position formula for iteration. The particle swarm algorithm is as follows:
[0045] Particle swarm optimization is a simplified model inspired by the regularity of bird flocking activities, and then using swarm intelligence to establish. Particle swarm optimization is based on the observation of animal flocking activities, using the sharing of information in the group to make the whole group movement in the problem solving space from disorder to order evolution process, so as to obtain the optimal solution.
[0046] Each particle has a speed and position Two parameters. In each iteration, the particle updates itself by tracking two "extremes". One extreme is the optimal solution found by the particle itself, called individual extremum P best ; the other extreme is the optimal solution found by the whole population at present, called global extremum G best . In finding the two optimal values, the particle updates its speed and new position according to the following formula:
[0047]
[0048]
[0049] Where ω is the inertia weight, the value of ω affects the global and local search ability, the value of ω is larger, the global search ability is stronger; The value of ω is smaller, the local search ability is stronger. C1, c2 are learning factors (acceleration coefficients), c1 is the individual cognitive coefficient, which represents the self-learning ability of the particle, c2 is the social cognitive coefficient, which represents the learning ability of the particle from the global optimal particle, c1, c2 generally take values between 0 and 2, r1, r2 are random numbers, generally take values between 0 and 1.
[0050] (1-6) output the current duty ratio to the boost circuit, return to (1-2).
[0051] Step 2, adopt perturbation observation method to track the global maximum power point;
[0052] (2-1) take the global optimal particle of the particle swarm optimization algorithm as the initial duty ratio, and give the duty ratio perturbation. According to the perturbation observation method, change the duty ratio to track the global maximum power point. The flow chart of perturbation observation method is shown in Figure 5 , as follows:
[0053] The basic idea of perturbation observation method is: first perturb the output voltage of photovoltaic module, then observe the change of photovoltaic module output power, according to the trend of power change, continuously change the direction of perturbation voltage, so that the photovoltaic cell finally works at the maximum power point.
[0054] First, measure the voltage and current, calculate the corresponding power. A small perturbation (AD) is applied to the duty cycle, calculate the corresponding power, and compare with the previous power. If the power change (AP) is positive, the perturbation direction is correct, otherwise the direction should be reversed.
[0055] (2-2) When the environment changes, and the set conditions are met, return to step 1, if not met, return to (2-1). The set conditions are as follows: Figure 6
[0056] When the rate of change of power exceeds a certain value, it can be considered that the environment has changed greatly, at this time the system should be restarted immediately; When the duty cycle continuously increases or continuously decreases and reaches a set number of times, it can be considered that the environment has changed slightly, at this time the algorithm should be restarted; When the duty cycle continuously increases or continuously decreases, but within the set number of times, it reaches a stable state, it can be considered that the environment changes very little, and the algorithm does not need to be restarted.
[0057] Figure 7(a) is a simulation comparison chart of the present application and the traditional particle swarm algorithm under condition 1, and figure 7(b) is a simulation comparison chart of the present application and the traditional particle swarm algorithm under condition 2, it can be seen that the present application has the advantage of fast convergence speed over the traditional particle swarm algorithm, and through multiple simulations, the present application is more accurate. Figure 8 is the comparison simulation chart of the present application's reinitialization algorithm, after the environment changes, the present application can better find the maximum power point according to the environment change, so that the output power is higher.
Claims
1. An improved hybrid global maximum power point tracking control method based on particle swarm optimization and perturbation observation, characterized in that, Includes the following steps: Step 1: Use the particle swarm optimization algorithm to find the peak value where the global maximum power point is located; Step 2: Use the perturbation observation method to track the global maximum power point; Step 3: After the environment changes, the particle swarm algorithm will be reinitialized. Step 1 specifically involves: (1-1) Initialize the particle swarm algorithm. The particles in the particle swarm algorithm are equivalent to the duty cycle of the boost circuit. The number of particles is determined according to the number of photovoltaic panels, and the particles are distributed according to the normal probability curve. (1-2) Measure the current voltage and current values; (1-3) Calculate the current power; (1-4) Determine whether the peak value of the maximum power point has been found. If it is found, proceed to step 2. If it is not found, proceed to the next step. (1-5) Iterate using the particle swarm algorithm based on the particle velocity-position formula; (1-6) Output the current duty cycle and return (1-2); Step 2 specifically involves: (2-1) Using the global optimal particle of the particle swarm algorithm as the initial duty cycle, and giving the duty cycle perturbation, the duty cycle is changed according to the perturbation observation method to track the global maximum power point; (2-2) When the environment changes and the set conditions are met, return to step 1; otherwise, return to (2-1).
2. The improved hybrid global maximum power point tracking control method based on particle swarm optimization and perturbation observation as described in claim 1, characterized in that, Particle swarm optimization (PSO) parameters include inertia weight, individual cognitive coefficient, social cognitive coefficient, individual extrema, global extrema, and transition condition parameters.