Maximum power point tracking MPPT (Maximum Power Point Tracking) system and method suitable for photovoltaic array under local shadow condition

Abstract

The invention discloses a maximum power point tracking (MPPT) system suitable for a photovoltaic array under a local shadow condition, the system comprises a microcontroller and a photovoltaic array, the microcontroller is connected with a PC upper computer and communicates with the PC upper computer, the microcontroller is connected with an acquisition module, the acquisition module comprises a current sensor ACS712 and a voltage sensor INA226, and the current sensor ACS712 and the voltage sensor INA226 are connected with the microcontroller. The current sensor ACS712 is connected in series in an output loop of the photovoltaic array, and the voltage sensor INA226 is connected in parallel at the output end of the photovoltaic array; the microcontroller is further connected with a driving chip EG2104, the driving chip EG2104 is connected with a Sepic converter module, the input of the Sepic converter is connected with the photovoltaic array, the output of the Sepic converter is connected with a load and a power meter in parallel, and the power meter communication module is connected with the PC upper computer to transmit power data in real time.

Description

TECHNICAL FIELD

[0001] The application is a maximum power point tracking (MPPT) system and method suitable for photovoltaic arrays under local shadow conditions, belonging to the field of photovoltaic systems. BACKGROUND

[0002] In the field of clean energy, photovoltaic power generation has become an important supplement to global power supply due to the universality and easy accessibility of solar energy. However, the power generation efficiency of photovoltaic systems is highly dependent on external environment and control technology. The shortcomings of existing technologies in complex scenarios directly lead to the technical needs of the present application, which has the following problems: I. Insufficient MPPT optimization ability under complex light, serious energy loss: Local shadow (such as building obstruction, cloud passing, module dust accumulation) is a common scenario in photovoltaic array operation, which not only leads to significant differences in light intensity among components in the array, but also causes two major problems: first, the working current of the shaded components is lower than that of the unshaded components, which bears negative voltage and changes from "power output" to "power absorption", converting electrical energy into heat energy to form hot spot phenomenon, which not only increases the system's own energy consumption, but also may cause system failure due to local overheating and component burning; second, the parallel bypass diode and series isolation diode to avoid hot spots make the power-voltage (P-U) curve of the photovoltaic array change from "single peak" to "multiple peaks" (containing multiple local maximum power points LMPP and one global maximum power point GMPP), and the current-voltage (I-U) curve changes from "single knee" to "multiple knees".

[0003] Traditional MPPT technologies (such as perturb and observe (PO), incremental conductance (INC), and constant voltage (CV)) perform stably under uniform light, but when facing a multi-peak P-U curve, they are easily locked into LMPP due to "local optimization inertia", and cannot break through the local optimum to find GMPP; even if some meta-heuristic algorithms (such as particle swarm optimization (PSO) and traditional Kepler optimization (KOA)) try to solve the multi-peak problem, they still have defects such as insufficient diversity of initialized population (such as random initialization leading to partial peak areas not covered) and slow convergence speed (still searching in invalid areas in the later iteration period), resulting in long-term deviation of photovoltaic array operation from GMPP, significant decrease in solar energy utilization rate, especially in fast light change (RCSIC, such as rapid movement of cloud layer) scenarios, the optimization lag problem is more prominent, further aggravating energy loss.

[0004] II. Poor system stability and adaptability, high risk of hardware loss: The existing MPPT system converter topology (such as Buck, Boost converter) has obvious limitations: first, the impedance matching blind area, Buck converter only supports step-down, Boost only supports step-up, when local shadow causes photovoltaic array output voltage dynamic change, it is difficult to cover the impedance matching demand of full voltage range, resulting in array and load impedance mismatch, output power fluctuation; second, poor current continuity, traditional topology input ripple is large, additional large capacity filtering components are required, increasing system size and cost, and long-term ripple impact easily shortens hardware life; third, lack of safety protection, lack of soft protection mechanism for short circuit and overcurrent, when the load suddenly changes or the component fails, it is easy to cause hardware damage.

[0005] At the same time, some MPPT technologies that rely on complex models (such as artificial neural network ANN) have high tracking accuracy, but they need to rely on special deep learning training platform, and the training parameters need to be frequently adjusted according to the changes of light and load, which faces the problem of "parameter adaptation difficulty and high maintenance cost" in actual deployment, and it is difficult to meet the engineering needs of small and medium-sized photovoltaic systems; while the design of fixed load adaptation cannot be compatible with the diversified power consumption scenes such as battery charging and inverter grid connection, which limits the universality of the technology.

[0006] Three, dynamic response and steady-state accuracy imbalance, low running reliability: In the scene of rapid light change, the dynamic response lag problem of traditional algorithm is significant: for example, the PO algorithm finds the power peak value by periodically perturbing the voltage, and the perturbation step is too large, which causes steady-state oscillation, and the step is too small, which causes slow response, when the light changes many times per second, the algorithm is always in the "catching up" state, and cannot lock the new GMPP in time; INC algorithm depends on the change of conductance to determine the optimization direction, but when the light changes suddenly, it is easy to appear "misjudgment", which causes the optimization direction to reverse, further prolonging the tracking time.

[0007] In addition, even if some algorithms can find the GMPP under static shadow, there is still a problem of steady-state oscillation: due to the lack of fine tuning mechanism, the power point finally located by the algorithm has a small deviation from the actual GMPP, which causes the photovoltaic array output power to fluctuate near the peak value, not only affecting the power supply stability, but also increasing the switching loss of the converter switch tube, reducing the reliability of long-term operation of the system, especially in grid-connected photovoltaic systems, power fluctuation may also cause power grid harmonic problems, which does not meet the grid connection specification requirements. SUMMARY

[0008] The technical problem to be solved by the present application is to provide a maximum power point tracking (MPPT) system and method suitable for a photovoltaic array under local shadow conditions, to solve the technical problems of MPPT of the photovoltaic array under complex conditions such as local shadow and rapid light change: first, to break through the difficulty that the traditional algorithm is easy to fall into a local maximum power point and cannot accurately locate a global maximum power point; second, to improve the dynamic response speed and convergence accuracy of MPPT, and to reduce steady-state oscillation; third, to avoid safety risks such as hot spot phenomenon, and to ensure stable operation of the system; fourth, to use easy-to-integrate hardware and lightweight algorithm, to enhance the technical practicability, to adapt to various loads, and finally to realize efficient capture of solar energy by the photovoltaic array and to improve power generation efficiency.

[0009] To solve the above technical problems, the present application adopts the following technical solutions: The maximum power point tracking (MPPT) system suitable for a photovoltaic array under local shadow conditions comprises a microcontroller and a photovoltaic array (PVArray), the microcontroller is connected with a PC host computer and communicates with the PC host computer, the microcontroller is connected with an acquisition module, the acquisition module comprises a current sensor ACS712 and a voltage sensor INA226, the current sensor ACS712 is connected in series in an output loop of the photovoltaic array PVArray, the voltage sensor INA226 is connected in parallel at an output end of the photovoltaic array, the current sensor ACS712 is connected with the microcontroller through an IIC interface, and the voltage sensor INA226 is connected with the microcontroller through an ADC interface. The microcontroller is further connected with a drive chip EG2104, the drive chip EG2104 is connected with a Sepic converter module, an input of the Sepic converter is connected with the photovoltaic array, a load and a power meter are connected in parallel at an output of the Sepic converter, and a communication module of the power meter is connected with the PC host computer to transmit power data in real time. The microcontroller is further connected with a power module, the power module comprises a voltage stabilizer AMS1117 and a power chip MP9486A, an input end of the power chip MP9486A is connected with external power supply or auxiliary power supply of the photovoltaic array, an output end of the power chip MP9486A is connected with the drive chip EG2104, and an output end of the voltage stabilizer AMS1117 is connected with the microcontroller.

[0010] Further, the Sepic converter module comprises a MOS tube, a G pole of the MOS tube is connected with the drive chip EG2104, a D pole of the MOS tube is connected with one end of an inductor L1 and one end of a capacitor C1, the other end of the inductor L1 is connected with one end of a capacitor Cin and the current sensor ACS712, the other end of the capacitor Cin is connected with the photovoltaic array PVArray, the other end of the capacitor C1 is connected with one end of an inductor L2 and one end of a diode D1, the other end of the inductor L2 is connected with the photovoltaic array PVArray, the other end of the diode D1 is connected with one end of a capacitor C2 and one end of a load R, and the other end of the capacitor C2 and the other end of the load R are connected with the photovoltaic array PVArray.

[0011] A maximum power point tracking method for photovoltaic array under local shadow condition, comprising the following steps: Step one, real-time acquisition of photovoltaic array parameters; the output current of the photovoltaic array is detected in real time by the current sensor ACS712, and the output voltage is synchronously acquired by the voltage sensor INA226, and the analog signals of the two types of sensors are converted into digital signals by ADC and then transmitted to the microcontroller; Step two, IKOA algorithm built-in microcontroller, under the condition of local shadow and rapid light change, the global maximum power point GMPP is accurately located by IKOA algorithm iteration optimization; Step three, Sepic converter impedance adjustment, by adjusting the duty cycle of the converter, the input impedance of the photovoltaic array is matched with the load impedance, and the photovoltaic array works at GMPP; The microcontroller calculates the optimal duty cycle of the Sepic converter according to the GMPP determined by IKOA, and generates a PWM pulse signal according to the calculated optimal duty cycle and transmits it to the drive chip EG2104; the drive chip EG2104 amplifies the signal and controls the on-off of the MOS tube; Step four, steady-state monitoring and dynamic adjustment; The power meter monitors the output power at the load end in real time, if the power deviates from the GMPP due to light fluctuation, the sensor reacquires the parameters, and the IKOA algorithm repeats the iteration optimization process to dynamically update the duty cycle.

[0012] Further, the step two comprises the following steps: Step 1, IKOA algorithm initialization; Step 2, IKOA algorithm iteration optimization.

[0013] Further, the IKOA algorithm initialization comprises the following steps: The microcontroller runs the IKOA algorithm, calculates the real-time power P=U×I based on the acquired voltage U and current I, and performs the following operations: The sun corresponds to the current optimal power point of the photovoltaic array, which is initially a random value; The planets correspond to the power points to be explored by MPPT, the position of each planet represents a set of voltage / current parameters of the photovoltaic array, and the mass of the planet is determined by the output power at the position. The higher the power, the greater the mass of the planet; Planetary motion: the optimization process corresponding to MPPT, when the planet rotates around the sun, the motion trajectory is adjusted through the gravity calculation in the algorithm, which is essentially to explore the area with higher power, and balance global exploration and local development through speed update; Step 1.1, initialize the position of the planet; N planetary positions are generated using a fractional-order Logistic chaotic mapping to achieve diversity in the initial search. The chaotic mapping formula is as follows: ,in, It is a constant. It is a fractional-order parameter. This is the population growth rate; a value of 3.25 indicates complete chaos. It is the gamma function. and These are adjacent iteration variables of the chaotic sequence. This is the current iteration variable, representing the state value generated by the chaotic mapping in the i-th iteration. The next iteration variable represents the state value generated by the chaotic mapping in the (i+1)th iteration, and all values ​​are in [0,1]. The chaotic sequence is calculated using the chaotic mapping formula: the initial value X0 of the chaotic mapping is set. Calculate X1 using X0, then calculate X2 using X1, and so on, iterating n times, where n = the population size, to obtain a chaotic sequence [X0, X1, X2, ..., Xn-1] containing n values; Step 1.2: Map the chaotic sequence to planetary positions. Each planetary position corresponds to a set of U and I parameters, i.e., a power point to be explored. Calculate the mass of each planet, i.e., the power P. Set the planet with the highest power as the initial sun, as follows: Map each X_k (k=0,1,...,N-1) in the chaotic sequence to the actual search space of the photovoltaic array. If the search space is the voltage range [U_min, U_max], then the initial voltage of the kth planet is U_k = U_min + X_k × (U_max - U_min); If the search space is the current range [I_min, I_max], then the initial current of the kth planet is I_k = I_min + X_k × (I_max - I_min); Finally, the initial positions (U_k, I_k) of N planets are obtained, and each planet corresponds to a power point to be explored (P_k=U_k×I_k), which lays the foundation for subsequent iterative optimization. Step 1.3, determine the orbital period and orbital eccentricity Initialization is performed before the IKOA algorithm begins, and its relevant definitions are as follows: ; Where rand is a random number between 0 and 1. It is a random number based on a normal distribution.

[0014] Furthermore, the iterative optimization of the IKOA algorithm includes the following steps: Step 2.1: Calculate the gravitational pull between the Sun and each planet, determine the direction of the gravitational pull on the planets, simulate the gravitational effect of the Sun on the planets in the solar system, and provide the driving force for the planets to move towards the current optimal solution. The formula for calculating the gravitational pull is as follows: ; in, Represents a random number between 0 and 1. It is a small parameter whose function is to prevent the divisor from being 0 in the formula. Represents the position of the sun and planetary positions Euclidean distance between The normalized value at time t The calculation formula is: ; It is an exponential function that gradually decreases over time, used to control the search accuracy of the algorithm, and is defined as: ,in It is the initial value of the decreasing function. It is a constant, t and These represent the current iteration number and the maximum iteration number, respectively. and Representing the mass of the sun and planetary mass The normalized value of the Sun, and the formula for calculating the mass of the Sun is: ; The formula for calculating the mass of a planet is: ; in The value represents a random number between 0 and 1. The `fit` function is used to calculate the optimal value. In MPPT control, the optimal value function calculates the product of the photovoltaic array's output current and output voltage, which is the output power of the photovoltaic array. The `worst` function is used to calculate the worst fitness value in the current iteration of the population. .

[0015] Furthermore, the IKOA algorithm iterative optimization also includes the following steps: Step 2.2: Calculation of planetary velocity, dynamic adjustment of planetary motion step size and direction, and balancing the global exploration and local development capabilities of the algorithm; The IKOA algorithm selects three distances between different planets to form the velocity component of the current planet's position, including the distance between the current planet's position and the random planet's position, the distance between different random planets, and the distance between the upper and lower limits of the search space; The planetary velocity update formula of IKOA algorithm is: ; Wherein represents the current planetary velocity, and the calculation formula thereof is divided into two parts by the numerical value of and are the positions of two random planets in the current planetary group; and are the upper limit and lower limit of the search space, represents the normalized Euclidean distance between the sun position and the planetary position during the planetary velocity update process: ; and are two complementary factors, , wherein M and L are determined by random values to determine whether a velocity component participates in the velocity update of the current planet, so as to avoid the algorithm from falling into local optimum, M is a factor for controlling the velocity component related to the distance between planets, L is a factor for controlling the velocity component related to the boundary of the search space, , is the semi-major axis of the elliptical orbit of the i-th planet: ; p, P1 and P2 are factors for determining whether a velocity component exists.

[0016] Further, the iterative optimization of the IKOA algorithm further includes the following steps: Step 2.3, planetary position update, simulating the physical process of planetary motion around the sun, adjusting the planetary position representing the power point to be explored through iteration, gradually approaching the global maximum power point GMPP of the photovoltaic array; The planetary position update formula has two, which are determined by the numerical value of random numbers r0 and r1, the first update formula is composed of the current planetary position, velocity, gravity and the distance between the planet and the sun: wherein represents the new position of the i-th planet at t+1, represents the velocity required for the i-th planet to reach the position at t+1, represents the best sun position found from the start of the algorithm to the current time t; ​In the first planetary position update formula, the planetary position update strategy is divided into two parts, the IKOA algorithm divides the entire iteration process into an exploration stage and a development stage to simulate the behavior of the planet, in the exploration stage, the distance between the planet and the sun represents the exploration operator, the planet moves away from the sun to find a new optimal solution; in the development stage, the velocity and gravity represent the development operator, the planet takes the current optimal solution as the sun and searches for a new optimal solution near its position; In order to provide the planet with a good speed to escape from the local optimal position, an adaptive disturbance factor is introduced in the first planetary position update formula , The calculation formula is: Where t is the iteration number at the current time, is the maximum iteration number; The second planetary position update formula keeps the planetary position unchanged when the parameter is 1; the current planetary position, the sun position and the random planetary position jointly determine the planetary position at the next time when the parameter is 0, the second planetary position update formula is: Where h is an adaptive adjustment factor for adjusting the distance between the planet and the sun at time t, , is a random number generated based on a normal distribution, is a cyclic control parameter defined in the entire optimization process and gradually decreases from -1 to -2, , wherein is the number of cyclic control parameters.

[0017] Further, the IKOA algorithm iterative optimization further includes the following steps: Step 2.4, after the iteration is completed, the iteration result is adjusted, the GSS algorithm is executed on the sun position to determine the voltage U_GMPP and the current I_GMPP corresponding to the final GMPP; After each iteration of the IKOA algorithm, the elite strategy is selected to use the planetary position with the best fitness value as the sun position, the elite strategy calculation formula is: ; In the GSS algorithm, the planetary position and the sun position after iteration are respectively taken as the endpoints and of the interval, the two points and on the search interval have: ; .

[0018] Further, the optimal duty ratio calculation process of the Sepic converter is as follows: Define core calculation parameters: GMPP key parameters: get the global maximum power point voltage and current output by IKOA algorithm ; ; Sepic hardware parameters: determine input / output inductance L1 and L2, usually L1=L2, MOS switch frequency f s ; Load real-time parameters: collect load voltage, current through power meter , calculate load resistance ; ; Determine Sepic converter operating mode: Determine mode based on inductor current ripple ratio MPPT works in continuous conduction mode CCM, calculate ripple ratio: , , L is the Sepic, input / output inductance value of the converter, D0 is the initial estimated duty ratio, is the average inductance current, , if , then the Sepic converter operating mode is CCM, otherwise it is DCM; Voltage / current relationship in CCM mode: In CCM mode, the input voltage of the Sepic converter is related to the output voltage , which is determined by the duty ratio D. The formula derivation is as follows: When the switch is on (duty ratio D stage): inductor L1 is charged by , inductor L2 is charged by capacitor C1, at this time , ; When the switch is off (duty ratio 1-D stage): inductor L1 and inductor L2 discharge, energy is transmitted to the load through capacitor C2, at this time , ; According to the inductor volt-second balance (the integral of inductor voltage over one period is 0), equations are listed for L1 and L2 respectively, and the core formula in CCM mode is finally simplified: ; If the load is a pure resistor: from , , substitute the voltage relationship formula and rearrange to get: ; If the load is not a resistor, from , , substitute the voltage relationship formula and directly solve: ; Dynamic correction and verification: Power feedback correction: Power meter monitors the actual output power of the array ,like If the power is too low, increase D; if it is too high, decrease D, with a step size of 0.01-0.05. Ripple and Temperature Compensation: If the inductor current ripple exceeds the threshold, L or f needs to be increased. s And recalculate D. In high temperature environments (≥40℃), in order to offset the on-state voltage drop of the MOSFET, increase D by 1%-3%.

[0019] The present invention adopts the above technical solution and has the following technical effects compared with the prior art: 1. GMPP tracking is accurate and efficient under complex lighting conditions; Global optimization without omissions: The IKOA algorithm initializes planetary positions through fractional-order Logistic chaotic mapping, ensuring that the initial search covers all potential regions of the multi-peak PU curve of the photovoltaic array. Combined with an adaptive perturbation factor, it effectively avoids the problem of traditional algorithms getting stuck in the local maximum power point (LMPP) and accurately locates the global maximum power point (GMPP).

[0020] It boasts both high convergence speed and high accuracy: chaotic mapping in the initialization phase accelerates search space coverage, adaptive perturbation in the iteration phase reduces invalid exploration, and the Golden Section Search (GSS) after iteration refines the fine-tuning, making the algorithm's steady-state establishment time as short as 0.06s and the tracking efficiency reach over 99.23%, far exceeding traditional algorithms such as PO, INC, and PSO.

[0021] Strong dynamic response capability: In the face of sudden changes in light intensity, the sensor collects voltage and current parameters in real time, and the IKOA algorithm iteratively updates the duty cycle to ensure that the photovoltaic array always keeps up with the GMPP drift and maintains high-efficiency output.

[0022] 2. The system operates stably and reliably; Wide impedance matching range: Sepic converters support duty cycle adjustment in the 0-1 range, which can adapt to the dynamic changes in photovoltaic array impedance caused by local shading, avoid the impedance matching blind zone of traditional Buck / Boost converters, and ensure real-time impedance matching between the array and the load under all operating conditions.

[0023] Low output ripple and high safety: The Sepic converter has better current continuity than traditional topologies, which significantly reduces input ripple and reduces filtering requirements; the coupling capacitor provides soft short-circuit protection, and the design of bypass diodes and reverse protection diodes effectively avoids hot spot phenomenon and current reverse impact, improving the long-term operating stability of the system.

[0024] Stable power supply: The dual power supply module, which uses an AMS1117 voltage regulator and an MP9486A power chip, provides a suitable and stable operating voltage for the microcontroller and the driver chip, respectively, to avoid voltage fluctuations affecting control accuracy and ensure reliable operation of the algorithm and hardware.

[0025] 3. Excellent adaptability and practicality; Hardware integration is simple: It uses mature components such as STM32F103C8T6 microcontroller, ACS712 current sensor, and INA226 voltage sensor, with strong interface compatibility, simple circuit design, and easy engineering implementation and mass application.

[0026] Flexible load adaptability: Sepic converters support buck-boost functions and can be matched with various load types such as resistive, battery, and inverter. By dynamically adjusting the duty cycle through real-time calculation of load parameters, it can meet the power demand of different photovoltaic power generation scenarios.

[0027] Convenient monitoring and debugging: The power meter collects power data from the load end in real time and uploads it to the PC host computer, realizing visualized monitoring and data recording of MPPT tracking effect, which facilitates algorithm optimization and system fault diagnosis.

[0028] 4. Energy utilization efficiency is significantly improved; Through the precise optimization of the IKOA algorithm and the efficient impedance adjustment of the Sepic converter, the photovoltaic array can still operate stably in GMPP under complex conditions such as local shading and rapid changes in illumination, maximizing the capture of solar energy and converting it into electrical energy. This significantly reduces the energy loss caused by optimization deviation or impedance mismatch in traditional algorithms, thereby improving the overall energy efficiency of the photovoltaic power generation system. Attached Figure Description

[0029] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0030] Figure 1 This is a circuit diagram of the maximum power point tracking (MPPT) system for photovoltaic arrays under partial shading conditions in this invention. Figure 2 This is a schematic diagram illustrating the application of the Golden Section Search Algorithm (GSS) in MPPT control in this invention. Detailed Implementation

[0031] Examples, such as Figure 1As shown, a maximum power point tracking (MPPT) system for photovoltaic arrays under partial shading conditions is presented. It includes a microcontroller and a photovoltaic array (PVArray). The microcontroller, model STM32F103C8T6, is connected to a PC for communication. The microcontroller is connected to a data acquisition module, which includes a current sensor ACS712 and a voltage sensor INA226. The ACS712 current sensor is connected in series in the output circuit of the PVArray, and the INA226 voltage sensor is connected in parallel at the output terminal of the PVArray. The ACS712 current sensor is connected to the microcontroller via an IIC interface, and the INA226 voltage sensor is connected to the microcontroller via an ADC interface.

[0032] The microcontroller is also connected to a driver chip EG2104, which is connected to a Sepic converter module. The input of the Sepic converter is connected to the photovoltaic array, and the output is connected in parallel to a load and a power meter. The power meter communication module is connected to a PC host computer to transmit power data in real time. The Sepic converter module includes a MOSFET. The gate of the MOSFET is connected to the driver chip EG2104, and the drain of the MOSFET is connected to one end of an inductor L1 and one end of a capacitor C1. The other end of the inductor L1 is connected to a capacitor C1. in One end is connected to the current sensor ACS712, and the capacitor C. in The other end is connected to the photovoltaic array PVArray. The other end of capacitor C1 is connected to one end of inductor L2 and one end of diode D. The other end of inductor L2 is connected to the photovoltaic array PVArray. The other end of diode D1 is connected to one end of capacitor C2 and one end of load R. The other end of capacitor C2 and the other end of load R are connected to the photovoltaic array PVArray.

[0033] Sepic converters support wide-range impedance matching. Even if local shading causes dynamic changes in array impedance, real-time matching can still be achieved through duty cycle adjustment (between 0 and 1), avoiding the impedance matching blind zone of traditional Buck / Boost converters. Sepic converters are superior to traditional Buck / Boost converters in terms of current continuity, which results in lower input ripple and reduced filtering requirements. In addition, the coupling capacitors of Sepic converters provide soft short-circuit protection, improving system stability.

[0034] The microcontroller is also connected to a power module, which includes a voltage regulator AMS1117 and a power chip MP9486A. The input of the power chip MP9486A is connected to an external power supply or a photovoltaic array auxiliary power supply, and the output of the power chip MP9486A is connected to a driver chip EG2104. The output of the voltage regulator AMS1117 is connected to the microcontroller to provide a stable operating voltage.

[0035] The ACS712 current sensor detects the output current of the photovoltaic array in real time, while the INA226 voltage sensor collects the output voltage synchronously. Both types of sensors convert analog signals into digital signals via an ADC and then transmit them to the microcontroller.

[0036] Algorithm operation and control signal generation: The microcontroller has a built-in improved Kepler optimization algorithm (IKOA) to calculate the real-time output power of the photovoltaic array based on the received voltage and current data, and determines the optimal duty cycle through algorithm iteration to generate the corresponding PWM pulse signal.

[0037] Converter drive and impedance regulation: The PWM pulse signal is transmitted to the driver chip EG2104. After the driver chip EG2104 amplifies the signal, it drives the MOSFET to switch on and off. The on and off state of the MOSFET controls the working mode of the Sepic converter (two states: switch closed / open). The input impedance of the Sepic converter is adjusted so that the impedance of the photovoltaic array is matched with the load impedance in real time, ensuring that it works at the maximum power point.

[0038] Power supply and status monitoring: The AMS1117 and MP9486 power modules provide stable and compatible operating voltages for the microcontroller and driver chip, respectively, to avoid voltage fluctuations affecting control accuracy; the power meter measures the actual output power at the load end in real time and uploads it to the PC host computer through the communication module to realize real-time monitoring and data recording of the tracking effect.

[0039] The maximum power point tracking method for photovoltaic arrays under partial shading conditions includes the following steps: Step 1: Real-time acquisition of photovoltaic array parameters; The ACS712 current sensor detects the output current of the photovoltaic array in real time, while the INA226 voltage sensor collects the output voltage synchronously. Both types of sensors convert analog signals into digital signals via an ADC and then transmit them to the microcontroller.

[0040] Step two: The microcontroller incorporates the IKOA algorithm to accurately locate the Global Maximum Power Point (GMPP) under complex conditions such as local shading and rapid changes in illumination through iterative optimization. Step 1, IKOA algorithm initialization; The microcontroller runs the IKOA algorithm to calculate the real-time power P=U×I based on the collected voltage U and current I, and performs the following operations: The current optimal power point of the solar photovoltaic array corresponds to the sun, and is initially a random value; Planets correspond to the power points to be explored in the MPPT. The position of each planet represents a set of voltage / current parameters of a photovoltaic array. Its mass is determined by the output power at that position. The higher the power, the larger the mass of the planet. Planetary motion: Corresponding to the optimization process of MPPT, when a planet revolves around the sun, the motion trajectory is adjusted through gravity calculation in the algorithm. Essentially, it is to explore regions with higher power, while balancing global exploration (finding new regions away from the sun) and local development (optimizing the current region by moving closer to the sun) through velocity updates.

[0041] Step 1.1: Initialize the planetary positions; N planetary positions are generated using a fractional-order Logistic chaotic mapping to achieve diversity in the initial search. The chaotic mapping formula is as follows: ,in, It is a constant. It is a fractional-order parameter. This is the population growth rate; a value of 3.25 indicates complete chaos. It is the gamma function. and These are adjacent iteration variables of the chaotic sequence. This is the current iteration variable, representing the state value generated by the chaotic mapping in the i-th iteration. The next iteration variable represents the state value generated by the chaotic mapping in the (i+1)th iteration, and all values ​​are in [0,1]. The chaotic sequence is calculated using the chaotic mapping formula: the initial value X0 of the chaotic mapping is set. Calculate X1 using X0, then calculate X2 using X1, and so on, iterating n times (n = population size, such as 50), to obtain a chaotic sequence [X0, X1, X2, ..., Xn-1] containing n values.

[0042] Step 1.2: Map the chaotic sequence to planetary positions. Each planetary position corresponds to a set of U and I parameters, i.e., a power point to be explored. Calculate the mass of each planet, i.e., the power P. Set the planet with the highest power as the initial sun, as follows: Map each X_k (k=0,1,...,N-1) in the chaotic sequence to the actual search space of the photovoltaic array. If the search space is the voltage range [U_min, U_max] (e.g., 0~36V), then the initial voltage of the kth planet is U_k = U_min + X_k × (U_max - U_min); If the search space is the current range [I_min, I_max] (e.g., 0~8A), then the initial current of the kth planet is I_k = I_min + X_k × (I_max - I_min); Finally, the initial positions (U_k, I_k) of N planets are obtained, and each planet corresponds to a power point to be explored (P_k=U_k×I_k), which lays the foundation for subsequent iterative optimization.

[0043] By introducing fractional-order chaos into the planetary position initialization, the algorithm's dynamic behavior during the exploration phase becomes richer, and its global optimization capability is enhanced.

[0044] Step 1.3, determine the orbital period and orbital eccentricity Initialization is performed before the IKOA algorithm begins, and its relevant definitions are as follows: ; Where rand is a random number between 0 and 1. It is a random number based on a normal distribution.

[0045] Step 2 involves iterative optimization using the IKOA algorithm. By simulating the physical process of planets orbiting the sun, the position of the planet representing the power point to be explored is dynamically adjusted. In the early stages of iteration, chaotic initialization is used to cover the multi-peak power region of the photovoltaic array to avoid missing the Global Maximum Power Point (GMPP). During iteration, gravity calculation and velocity adjustment are used to balance global exploration and local development to prevent getting stuck in the Local Maximum Power Point (LMPP). Finally, by combining an elite strategy with fine-tuning, the GMPP is accurately located, providing the optimal control basis for subsequent Sepic converter impedance adjustment, ensuring that the photovoltaic array always works at maximum power and maximizing solar energy utilization. Step 2.1 calculates the gravitational pull between the Sun and each planet, determines the direction of the gravitational pull on the planets, and simulates the gravitational effect of the Sun on the planets in the solar system. This provides the driving force for the planets (representing the power points to be explored) to move towards the current optimal solution (the Sun, corresponding to the highest power point found): by quantifying the gravitational strength between the Sun and the planets (inversely proportional to the distance between them and positively correlated with the mass), the planets are guided to gather in high-power regions, while avoiding premature convergence to the local maximum power point (LMPP). When the gravity is weak in the early stage of iteration, the planets can explore more regions; when the gravity is stronger in the later stage, the focus is on development near the current optimal solution, ultimately laying the driving foundation for accurately locating the global maximum power point (GMPP) of the photovoltaic array.

[0046] The formula for calculating gravity is as follows: ; in, Represents a random number between 0 and 1. It is a small parameter whose function is to prevent the divisor from being 0 in the formula. Represents the position of the sun and planetary positions Euclidean distance between The normalized value at time t The calculation formula is: ; It is an exponential function that gradually decreases over time, used to control the search accuracy of the algorithm, and is defined as: ,in It is the initial value of the decreasing function. It is a constant, t and These represent the current iteration number and the maximum iteration number, respectively. and Representing the mass of the sun and planetary mass The normalized value of the Sun, and the formula for calculating the mass of the Sun is: ; The formula for calculating the mass of a planet is: ; in The value represents a random number between 0 and 1. The `fit` function is used to calculate the optimal value. In MPPT control, the optimal value function calculates the product of the photovoltaic array's output current and output voltage, which is the output power of the photovoltaic array. The `worst` function is used to calculate the worst fitness value in the current iteration of the population. .

[0047] Step 2.2, planetary velocity calculation, dynamically adjust the step size and direction of the planet (representing the power point to be explored), balance the global exploration and local development capabilities of the algorithm, ensure that under complex lighting conditions such as local shadows, it can efficiently cover the multi-peak power area of ​​the photovoltaic array and accurately approach the global maximum power point (GMPP), while avoiding low optimization efficiency or getting trapped in local optima due to improper step size, and provide key motion parameter support for subsequent planetary position updates (to achieve MPPT tracking); The velocity of a planet is determined by its distance from the sun. As a planet moves away from the sun, its velocity gradually decreases; as it moves closer to the sun, its velocity gradually increases. The update step size of the planet's velocity directly affects the algorithm's convergence accuracy and speed: too small a step size leads to a decrease in the diversity of planetary positions; too large a step size can cause the algorithm to skip optimal solutions or oscillate during development. Therefore, to ensure that the planet's velocity matches its position, the IKOA algorithm selects three distances between different planets to constitute the velocity component of the current planet's position: the distance between the current planet's position and random planetary positions, the distance between different random planets, and the distance between the upper and lower bounds of the search space.

[0048] The planetary velocity update formula of the IKOA algorithm is: ; in Represents the current velocity of the planet, and its calculation formula is... The numerical value is divided into two parts. and It represents the positions of two random planets in the current planetary cluster; and These are the upper and lower limits of the search space. Represents the position of the sun during planetary velocity updates. and planetary positions Normalized Euclidean distance between them: ; and They are two complementary factors. , M and L are randomly selected to determine whether a certain velocity component participates in the current planet's velocity update, thus avoiding the algorithm getting trapped in local optima. M is a factor that controls the velocity components related to the distance between planets. L is a factor that controls the velocity components related to the search space boundary. , Here is the semi-major axis of the elliptical orbit of the i-th planet: p, P1, and P2 are all factors used to determine the existence of velocity components, and their values ​​are determined as shown in Table 1. Table 1. Conditions for the values ​​of different factors: r2, r3, r4, and r5 are random numbers ranging from 0 to 1; F is used to adjust the direction of planetary velocity updates, and its value conditions are also given in Table 1.

[0049] Step 2.3, planetary position update, simulates the physical process of planets moving around the sun, and gradually approaches the global maximum power point (GMPP) of the photovoltaic array by iteratively adjusting the position of the planets representing the power points to be explored, ultimately solving the multi-peak optimization and accurate convergence problem of maximum power point tracking (MPPT) under complex illumination (especially local shading) conditions; There are two formulas for updating planetary positions, the use of which is determined by the values ​​of random numbers r0 and r1. The first update formula is composed of the current planetary position, velocity, gravity, and the distance between the planet and the sun: ,in, This represents the new position of the i-th planet at time t+1. This represents the velocity required for the i-th planet to reach its position at time t+1. This represents the optimal position of the sun found from the start of the algorithm up to the current time t.

[0050] In the first planetary position update formula, the planetary position update strategy is divided into two parts. As the planet revolves around the sun, it first approaches the sun and then moves away from it within a certain period of time. The IKOA algorithm divides the entire iterative process into an exploration phase and a development phase to simulate this behavior of the planet. In the exploration phase, the distance between the planet and the sun represents the exploration operator, and the planet moves away from the sun to find a new optimal solution. In the development phase, velocity and gravity represent the development operator, and the planet uses the sun as the current optimal solution and searches for a new optimal solution in its vicinity. By combining the exploration and development phases, the IKOA algorithm can effectively explore the entire search area and develop new solutions based on the current optimal solution, thus avoiding the algorithm from getting trapped in local optima.

[0051] In the KOA algorithm, the direction update parameter F, used to help the algorithm escape local optima, has a fixed probability of 0.5 throughout the optimization process. This means that the probability of F changing the planet's velocity is equal throughout the entire iteration cycle. However, after the algorithm gets trapped in a local optimum, this parameter cannot effectively provide the planet with the velocity to escape its local optimal position. Therefore, an adaptive perturbation factor is introduced into the first planet's position update formula. To further optimize the IKOA algorithm and avoid getting trapped in local optima, The calculation formula is: Where t is the iteration number at the current moment, To maximize the number of iterations, the adaptive perturbation factor is based on the fundamental principle that a planet needs sufficient speed to escape the ever-increasing gravity as it approaches the sun. This increases the possibility of the algorithm escaping local optima and prevents the algorithm from getting stuck in LMPP during MPPT applications.

[0052] The formula for updating the position of the second planet in the parameters When the value is 1, the planet's position remains unchanged; parameter When the value is 0, the current planetary position, the sun's position, and the random planetary position together determine the planetary position at the next moment. The formula for updating the second planetary position is: , where h is an adaptive adjustment factor used to adjust the distance between the planet and the sun at time t. , These are random numbers generated based on a normal distribution. It is a loop control parameter defined to gradually decrease from -1 to -2 throughout the optimization process. In the formula It is the number of loop control parameters.

[0053] Step 2.4: After the iteration is complete, adjust the iteration results and execute the GSS algorithm on the sun's position to determine the voltage corresponding to the final GMPP. and current The optimization of the IKOA algorithm improves the accuracy and stability of MPPT control: by introducing the Golden Section Search (GSS) to fine-tune the optimal solution (sun position, corresponding to the initially found power point) after iteration, it compensates for the insufficient optimization accuracy that may exist in the iteration stage and reduces steady-state oscillations; at the same time, the elite strategy is combined to retain better solutions, ensuring that the final location of the global maximum power point (GMPP) of the photovoltaic array is more accurate, and adapting to the low oscillation and high accuracy requirements of MPPT under complex lighting conditions such as local shading.

[0054] After each iteration, the IKOA algorithm selects the planetary position with the best humidity value as the sun position, ensuring that subsequent iterations are always guided by the optimal candidate solution in the population. This provides a precise optimization target for maximum power point tracking (MPPT) of the photovoltaic array. The formula for calculating the elite strategy is as follows: .

[0055] After the entire iteration phase, the IKOA algorithm introduces the Golden Section Search (GSS) algorithm to fine-tune the position of the sun after iteration, further improving the accuracy of the optimization results. GSS is suitable for optimization problems with a unimodal objective function. Although the MPPT problem is a multimodal function, after the algorithm iteration, the distance between the planets responsible for developing the search space around the sun and the sun after iteration is extremely short. Within the interval formed by these planets, the problem to be optimized can be regarded as a unimodal function. The sun after iteration and the planet closest to it are selected as the endpoints of the search interval using the Golden Section Search algorithm to find possible better solutions between the planet and the position of the sun after iteration.

[0056] In the GSS algorithm, the planetary position and the sun's position after iteration are used as the endpoints of the interval. and For the search interval The two points above and have: ; .

[0057] Figure 2 This diagram illustrates the application of the Golden Section Search Algorithm (GSS) in MPPT control, visually demonstrating how GSS fine-tunes the initial optimal solution (solar position) after IKOA iteration to improve the tracking accuracy of the maximum power point (MPP) and reduce steady-state oscillations. The horizontal axis represents the output voltage (U) of the photovoltaic array, and the vertical axis represents the output power (P). The optimal solution lies in the interval If the above is true, then the optimal solution lies in the interval [0, 1]. The search process will continue on the subinterval containing the optimal solution, and the new two-point selection method will still follow... Calculation formula and The calculation formula continues until the termination condition is met. Here, the termination condition is set to shrink the interval length to 5% of its original value. After the GSS algorithm ends, the elite strategy is used again: if a better solution than the iteratively updated sun is obtained after applying the GSS algorithm, the optimal solution is updated; otherwise, the iteratively updated sun is retained as the optimal solution.

[0058] Compared to other algorithms, the IKOA algorithm converges and iterates faster, and establishes a steady state in a shorter time. The fractional-order Logistic chaotic mapping in the initialization phase (quickly covering the search space) and the adaptive perturbation factor in the iteration phase (reducing invalid exploration) jointly accelerate the approach to the optimal solution. In terms of convergence accuracy, the optimal solution is better and more stable. The Golden Section Search (GSS) algorithm introduced after the three-way iteration of IKOA finely tunes the initial optimal solution, eliminating minor deviations in the iteration phase. Simultaneously, the continuous optimization of the elite strategy ensures the optimality of the final solution and reduces steady-state oscillations.

[0059] Step 3: Sepic converter impedance adjustment. By adjusting the converter duty cycle, the input impedance of the photovoltaic array is matched with the load impedance, and the photovoltaic array operates in GMPP mode. The microcontroller calculates the optimal duty cycle of the Sepic converter based on the GMPP determined by IKOA. The microcontroller generates a PWM pulse signal based on the calculated optimal duty cycle and transmits it to the driver chip EG2104. The driver chip EG2104 amplifies the signal and controls the switching on and off of the MOSFET.

[0060] The optimal duty cycle calculation process for the Sepic converter is as follows: Define the core calculation parameters: Key GMPP parameter: Obtain the global maximum power point voltage output by the IKOA algorithm. and current ; Sepic hardware parameters: Determine the input / output inductors L1 and L2, typically L1=L2, and the MOSFET switching frequency f. s ; Real-time load parameters: Load voltage is collected via a power meter. Current Calculate the load resistance .

[0061] Determine the SEPIC converter's operating mode: Based on inductor current ripple rate Determine the mode; MPPT typically operates in continuous conduction mode (CCM). Calculate the ripple rate: ,in L represents Sepic, the input / output inductance value of the converter, and D0 is the initial estimated duty cycle. The average current of the inductor. ,like If the Sepic converter operates in CCM mode, then it operates in DCM mode; otherwise, it operates in DCM mode.

[0062] Voltage / current relationship in CCM mode: In CCM mode, the input voltage of the Sepic converter With output voltage The relationship is determined by the duty cycle D, and the formula is derived as follows: When the switching transistor is turned on (duty cycle D stage): Inductor L1 is... During charging, inductor L2 is charged by capacitor C1. , ; When the switching transistor is turned off (duty cycle 1-D stage): inductors L1 and L2 discharge, and energy is transferred to the load through capacitor C2. , ; Based on the inductor volt-second balance (the integral sum of the inductor voltages over one cycle is 0), equations are set up for L1 and L2 respectively, and finally simplified to obtain the core formula in CCM mode: ; If the load is purely resistive: from have to Substituting into the voltage relationship formula and simplifying, we get: ; If the load is non-resistive, by have to Substitute the voltage relationship formula and solve directly: .

[0063] Dynamic correction and verification: Power feedback correction: Power meter monitors the actual output power of the array ,like If the power is too low, increase D; if it is too high, decrease D, with a step size of 0.01-0.05. Ripple and Temperature Compensation: If the inductor current ripple exceeds the threshold, L or f needs to be increased. s And recalculate D. In high temperature environments (≥40℃), in order to offset the on-state voltage drop of the MOSFET, increase D by 1%-3%.

[0064] Step four: Steady-state monitoring and dynamic adjustment; The power meter monitors the output power at the load end in real time. If the power deviates from GMPP due to light fluctuations, the sensor re-acquires parameters, and the IKOA algorithm repeats the above optimization process to dynamically update the duty cycle.

[0065] The description of this invention is given for illustrative and descriptive purposes only and is not intended to be exhaustive or to limit the invention to the forms disclosed. Many modifications and variations will be apparent to those skilled in the art. The embodiments were chosen and described in order to better illustrate the principles and practical application of the invention and to enable those skilled in the art to understand the invention and design various embodiments with various modifications suitable for a particular purpose.

Claims

1. A maximum power point tracking (MPPT) system for photovoltaic arrays under partial shading conditions, characterized in that: It includes a microcontroller and a photovoltaic array (PVArray). The microcontroller is connected to a PC and communicates with the PC. The microcontroller is also connected to a data acquisition module, which includes a current sensor ACS712 and a voltage sensor INA226. The current sensor ACS712 is connected in series in the output circuit of the PVArray, and the voltage sensor INA226 is connected in parallel at the output of the PVArray. The current sensor ACS712 is connected to the microcontroller through an IIC interface, and the voltage sensor INA226 is connected to the microcontroller through an ADC interface. The microcontroller is also connected to a driver chip EG2104, which is connected to a Sepic converter module. The input of the Sepic converter is connected to the photovoltaic array, and the output is connected in parallel to a load and a power meter. The power meter communication module is connected to a PC host computer to transmit power data in real time. The microcontroller is also connected to a power module, which includes a voltage regulator AMS1117 and a power chip MP9486A. The input terminal of the power chip MP9486A is connected to an external power supply or a photovoltaic array auxiliary power supply, and the output terminal of the power chip MP9486A is connected to a driver chip EG2104. The output terminal of the voltage regulator AMS1117 is connected to the microcontroller.

2. The maximum power point tracking (MPPT) system for photovoltaic arrays under partial shading conditions as described in claim 1, characterized in that: The Sepic converter module includes a MOSFET. The gate of the MOSFET is connected to the driver chip EG2104. The drain of the MOSFET is connected to one end of inductor L1 and one end of capacitor C1. The other end of inductor L1 is connected to one end of capacitor Cin and current sensor ACS712. The other end of capacitor Cin is connected to photovoltaic array PVArray. The other end of capacitor C1 is connected to one end of inductor L2 and one end of diode D. The other end of inductor L2 is connected to photovoltaic array PVArray. The other end of diode D1 is connected to one end of capacitor C2 and one end of load R. The other end of capacitor C2 and the other end of load R are connected to photovoltaic array PVArray.

3. A maximum power point tracking method for photovoltaic arrays under partial shading conditions, characterized in that: The method, when applied to the maximum power point tracking (MPPT) system for photovoltaic arrays under partial shading conditions as described in any one of claims 1-2, includes the following steps: Step 1: Real-time acquisition of photovoltaic array parameters; the current sensor ACS712 detects the output current of the photovoltaic array in real time, and the voltage sensor INA226 collects the output voltage synchronously. The two types of sensors convert the analog signals into digital signals through ADC and then transmit them to the microcontroller. Step 2: The microcontroller has a built-in IKOA algorithm. Under complex conditions such as local shadows and rapid changes in illumination, the IKOA algorithm is used to iteratively optimize and accurately locate the global maximum power point (GMPP). Step 3: Sepic converter impedance adjustment. By adjusting the converter duty cycle, the input impedance of the photovoltaic array is matched with the load impedance, and the photovoltaic array operates in GMPP mode. The microcontroller calculates the optimal duty cycle of the Sepic converter based on the GMPP determined by IKOA. The microcontroller then generates a PWM pulse signal based on the calculated optimal duty cycle and transmits it to the driver chip EG2104. The driver chip EG2104 amplifies the signal and controls the switching on and off of the MOSFET. Step four: Steady-state monitoring and dynamic adjustment; The power meter monitors the output power at the load end in real time. If the power deviates from GMPP due to fluctuations in illumination, the sensor re-acquires parameters, and the IKOA algorithm repeats the iterative optimization process to dynamically update the duty cycle.

4. The maximum power point tracking method for photovoltaic arrays under partial shading conditions as described in claim 3, characterized in that: Step two includes the following steps: Step 1, IKOA algorithm initialization; Step 2: Iterative optimization using the IKOA algorithm.

5. The maximum power point tracking method for photovoltaic arrays under partial shading conditions as described in claim 4, characterized in that: The IKOA algorithm initialization includes the following steps: The microcontroller runs the IKOA algorithm to calculate the real-time power P=U×I based on the collected voltage U and current I, and performs the following operations: The current optimal power point of the solar photovoltaic array corresponds to the sun, and is initially a random value; Planets correspond to the power points to be explored in the MPPT. The position of each planet represents a set of voltage / current parameters of a photovoltaic array. Its mass is determined by the output power at that position. The higher the power, the larger the mass of the planet. Planetary motion: Corresponding to the optimization process of MPPT, when a planet revolves around the sun, the motion trajectory is adjusted through gravity calculation in the algorithm. In essence, it is to explore to regions with higher power, and at the same time, the global exploration and local development are balanced through velocity updates. Step 1.1: Initialize the planetary positions; N planetary positions are generated using a fractional-order Logistic chaotic mapping to achieve diversity in the initial search. The chaotic mapping formula is as follows: ,in, It is a constant. It is a fractional-order parameter. This is the population growth rate; a value of 3.25 indicates complete chaos. It is the gamma function. and These are adjacent iteration variables of the chaotic sequence. This is the current iteration variable, representing the state value generated by the chaotic mapping in the i-th iteration. The next iteration variable represents the state value generated by the chaotic mapping in the (i+1)th iteration, and all values ​​are in [0,1]. The chaotic sequence is calculated using the chaotic mapping formula: the initial value X0 of the chaotic mapping is set. Calculate X1 using X0, then calculate X2 using X1, and so on, iterating n times, where n = the population size, to obtain a chaotic sequence [X0, X1, X2, ..., Xn-1] containing n values; Step 1.2: Map the chaotic sequence to planetary positions. Each planetary position corresponds to a set of U and I parameters, i.e., a power point to be explored. Calculate the mass of each planet, i.e., the power P. Set the planet with the highest power as the initial sun, as follows: Map each X_k (k=0,1,...,N-1) in the chaotic sequence to the actual search space of the photovoltaic array. If the search space is the voltage range [U_min, U_max], then the initial voltage of the kth planet is U_k = U_min + X_k × (U_max - U_min); If the search space is the current range [I_min, I_max], then the initial current of the kth planet is I_k = I_min + X_k × (I_max - I_min); Finally, the initial positions (U_k, I_k) of N planets are obtained, and each planet corresponds to a power point to be explored (P_k=U_k×I_k), which lays the foundation for subsequent iterative optimization. Step 1.3, determine the orbital period and orbital eccentricity Initialization is performed before the IKOA algorithm begins, and the relevant definitions are as follows: ; Where rand is a random number between 0 and 1. It is a random number based on a normal distribution.

6. The maximum power point tracking method for photovoltaic arrays under partial shading conditions as described in claim 5, characterized in that: The IKOA algorithm iterative optimization includes the following steps: Step 2.1: Calculate the gravitational pull between the Sun and each planet, determine the direction of the gravitational pull on the planets, simulate the gravitational effect of the Sun on the planets in the solar system, and provide the driving force for the planets to move towards the current optimal solution. The formula for calculating the gravitational pull is as follows: ； in, Represents a random number between 0 and 1. It is a small parameter whose function is to prevent the divisor from being 0 in the formula. Represents the position of the sun and planetary positions Euclidean distance between The normalized value at time t The calculation formula is: ; It is an exponential function that gradually decreases over time, used to control the search accuracy of the algorithm, and is defined as: ,in It is the initial value of the decreasing function. It is a constant, t and These represent the current iteration number and the maximum iteration number, respectively. and Representing the mass of the sun and planetary mass The normalized value of the Sun, and the formula for calculating the mass of the Sun is: ; The formula for calculating the mass of a planet is: ; in The value represents a random number between 0 and 1. The `fit` function is used to calculate the optimal value. In MPPT control, the optimal value function calculates the product of the photovoltaic array's output current and output voltage, which is the output power of the photovoltaic array. The `worst` function is used to calculate the worst fitness value in the current iteration of the population. .

7. The maximum power point tracking method for photovoltaic arrays under partial shading conditions as described in claim 5, characterized in that: The iterative optimization of the IKOA algorithm also includes the following steps: Step 2.2: Calculation of planetary velocity, dynamic adjustment of planetary motion step size and direction, and balancing the global exploration and local development capabilities of the algorithm; The IKOA algorithm selects three distances between different planets to form the velocity component of the current planet's position, including the distance between the current planet's position and the random planet's position, the distance between different random planets, and the distance between the upper and lower limits of the search space; The planetary velocity update formula of the IKOA algorithm is: ； in Represents the current velocity of the planet, and its calculation formula is... The numerical value is divided into two parts. and It represents the positions of two random planets in the current planetary cluster; and These are the upper and lower limits of the search space. Represents the position of the sun during planetary velocity updates. and planetary positions Normalized Euclidean distance between them: ; and They are two complementary factors. , M and L are randomly selected to determine whether a certain velocity component participates in the current planet's velocity update, thus avoiding the algorithm getting trapped in local optima. M is a factor that controls the velocity components related to the distance between planets. L is a factor that controls the velocity components related to the search space boundary. , Here is the semi-major axis of the elliptical orbit of the i-th planet: p, P1, and P2 are all factors used to determine whether the velocity component exists.

8. The maximum power point tracking method for photovoltaic arrays under partial shading conditions as described in claim 7, characterized in that: The iterative optimization of the IKOA algorithm also includes the following steps: Step 2.3, planetary position update, simulate the physical process of planets moving around the sun, and gradually approach the global maximum power point (GMPP) of the photovoltaic array by iteratively adjusting the position of the planets representing the power point to be explored; There are two formulas for updating planetary positions, the use of which is determined by the values ​​of random numbers r0 and r1. The first update formula is composed of the current planetary position, velocity, gravity, and the distance between the planet and the sun: ,in, This represents the new position of the i-th planet at time t+1. This represents the velocity required for the i-th planet to reach its position at time t+1. This represents the optimal position of the sun found from the start of the algorithm to the current time t; In the first planetary position update formula, the planetary position update strategy is divided into two parts. The IKOA algorithm divides the entire iterative process into an exploration phase and a development phase to simulate the behavior of the planet. In the exploration phase, the distance between the planet and the sun represents the exploration operator, and the planet moves away from the sun to find a new optimal solution. In the development phase, velocity and gravity represent the development operator, and the planet uses the current optimal solution as the sun and searches for a new optimal solution in its vicinity. To effectively provide the planet with the velocity needed to escape its local optimum, an adaptive perturbation factor is introduced into the first planet's position update formula. , The calculation formula is: Where t is the iteration number at the current moment, This represents the maximum number of iterations. The formula for updating the position of the second planet in the parameters When the value is 1, the planet's position remains unchanged; parameter When the value is 0, the current planetary position, the sun's position, and the random planetary position together determine the planetary position at the next moment. The formula for updating the second planetary position is: , where h is an adaptive adjustment factor used to adjust the distance between the planet and the sun at time t. , These are random numbers generated based on a normal distribution. It is a loop control parameter defined to gradually decrease from -1 to -2 throughout the optimization process. In the formula It is the number of loop control parameters.

9. The maximum power point tracking method for photovoltaic arrays under partial shading conditions as described in claim 8, characterized in that: The iterative optimization of the IKOA algorithm also includes the following steps: Step 2.4: After the iteration is completed, the iteration results are adjusted. The GSS algorithm is executed on the sun position to determine the voltage U_GMPP and current I_GMPP corresponding to the final GMPP. After each iteration, the IKOA algorithm selects the planetary position with the best maxima as the sun's position using an elite strategy. The formula for calculating the elite strategy is as follows: ; In the GSS algorithm, the planetary position and the sun's position after iteration are used as the endpoints of the interval. and For the search interval The two points above and have: ; .

10. The maximum power point tracking method for photovoltaic arrays under partial shading conditions as described in claim 3, characterized in that: The optimal duty cycle calculation process for the Sepic converter is as follows: Define the core calculation parameters: Key GMPP parameter: Obtain the global maximum power point voltage output by the IKOA algorithm. and current ; Sepic hardware parameters: Determine the input / output inductors L1 and L2, typically L1=L2, and the MOSFET switching frequency f. s ; Real-time load parameters: Load voltage is collected via a power meter. Current Calculate the load resistance ; Determine the SEPIC converter's operating mode: Based on inductor current ripple rate Determine the mode; MPPT typically operates in continuous conduction mode (CCM). Calculate the ripple rate: ,in L represents Sepic, the input / output inductance value of the converter, and D0 is the initial estimated duty cycle. The average current of the inductor. ,like If the SEPIC converter operates in CCM mode, then it operates in DCM mode; otherwise, it operates in DCM mode. Voltage / current relationship in CCM mode: In CCM mode, the input voltage of the Sepic converter With output voltage The relationship is determined by the duty cycle D, and the formula is derived as follows: When the switching transistor is turned on (duty cycle D stage): Inductor L1 is... During charging, inductor L2 is charged by capacitor C1. , ; When the switching transistor is turned off (duty cycle 1-D stage): inductors L1 and L2 discharge, and energy is transferred to the load through capacitor C2. , ; Based on the inductor volt-second balance (the integral sum of the inductor voltages over one cycle is 0), equations are set up for L1 and L2 respectively, and finally simplified to obtain the core formula in CCM mode: ; If the load is purely resistive: from have to Substituting into the voltage relationship formula and simplifying, we get: ; If the load is non-resistive, by have to Substitute the voltage relationship formula and solve directly: ; Dynamic correction and verification: Power feedback correction: Power meter monitors the actual output power of the array ,like If the power is too low, increase D; if it is too high, decrease D, with a step size of 0.01-0.

05. Ripple and Temperature Compensation: If the inductor current ripple exceeds the threshold, L or f needs to be increased. s And recalculate D. In high temperature environments (≥40℃), in order to offset the on-state voltage drop of the MOSFET, increase D by 1%-3%.

Images