Photovoltaic array maximum power point tracking method and system based on quantum marine predator-disturbance observation fusion algorithm

Abstract

The application discloses a photovoltaic array maximum power point tracking method and system based on a quantum marine predator-disturbance observation fusion algorithm, and specifically is as follows: on the basis of marine predator optimization, the duty cycle of a front-stage DC-DC converter is taken as a to-be-optimized control variable, candidate duty cycles are normalized and mapped into rotation angle parameters of a parameterized quantum circuit, quantum state coding, quantum gate operation and measurement sampling are performed under a quantum computing framework, and position updating of the marine predator algorithm is corrected according to measurement results, so that global optimization of quantum computing is realized; when a search solution enters a maximum power point neighborhood, a disturbance observation method is switched to for local fine tracking; and when environmental mutation or output power deviation exceeding a threshold is detected, quantum global search is restarted. The application improves the search ability for a global maximum power point under the condition of local shadow and multi-peak power curve, has fast algorithm convergence speed, strong steady-state operation performance and a wide application range.

Description

Technical Field

[0001] This invention relates to the field of new energy power generation and intelligent control technology, and in particular to a method and system for maximum power point tracking of photovoltaic arrays based on the quantum ocean predator-perturbation observation fusion algorithm. Background Technology

[0002] With the advancement of the global energy transition, photovoltaic (PV) power generation, as a clean and renewable energy source, is increasingly accounting for a larger share of the power system. The output power of a PV array is significantly nonlinear, influenced by irradiance, ambient temperature, and load conditions. To achieve efficient energy capture, PV power generation systems are typically equipped with a Maximum Power Point Tracking (MPPT) controller, ensuring the array operates near its maximum power point in real time.

[0003] Under uniform illumination, the power-voltage characteristic curve of a photovoltaic (PV) array exhibits a single-peak shape. Traditional MPPT methods, such as Perturb and Observe (P&O) and Incremental Conductance (INC), can achieve maximum power point tracking with relatively simple control logic. However, in actual operating environments, PV arrays often face non-uniform illumination conditions such as local shading, cloud cover, and building projection. In these cases, the PV characteristic curve of the array exhibits multi-peak characteristics, meaning there are multiple local maxima, and the global maximum power point may be located at any of these peaks. Traditional MPPT algorithms based on the single-peak assumption are prone to getting trapped in local optima, causing the system to operate at a suboptimal operating point for extended periods, resulting in significant energy loss.

[0004] To address the multimodal dynamic programming problem (MPPT), researchers have proposed various intelligent optimization algorithms, such as particle swarm optimization, genetic algorithms, and differential evolution. These algorithms possess strong global search capabilities and can overcome the influence of local shading to some extent. However, they generally suffer from slow convergence speed, parameter sensitivity, and large steady-state oscillations, and struggle to balance dynamic response and steady-state accuracy when the environment changes rapidly. In recent years, the Marine Predators Algorithm (MPA), as a novel metaheuristic algorithm, has demonstrated good performance on various optimization problems by simulating the Lévy walk and Brownian motion strategies of marine predators. However, the standard marine predator algorithm uses real-number encoding, which still faces challenges such as rapid decline in population diversity and premature convergence in the multimodal dynamic programming problem.

[0005] On the other hand, quantum computing, with its superposition, entanglement, and parallelism, offers new solutions to complex optimization problems. Quantum heuristic algorithms, by introducing quantum state encoding into the classical optimization framework, can enhance population diversity while maintaining low computational complexity. However, pure quantum algorithms are limited by the hardware conditions of the current NISQ stage and are difficult to deploy directly in industrial controllers. How to organically combine the ideas of quantum computing with the classical MPPT algorithm to construct a fusion framework that can both quickly locate the global optimum and accurately track it has become a current research hotspot. Existing MPPT techniques suffer from low tracking accuracy, slow convergence speed, and susceptibility to local optima under complex conditions such as local shading and rapid changes in irradiance. Therefore, there is an urgent need for an adaptive MPPT method that can balance global search capability and local fine-grained tracking capability. Summary of the Invention

[0006] The purpose of this invention is to provide a method and system for tracking the maximum power point of a photovoltaic array based on the Quantum Marine Predator Algorithm - Perturb and Observe (QMPA-P&O) that can balance global search capability and local tracking accuracy.

[0007] The technical solution to achieve the purpose of this invention is: a method for maximum power point tracking of photovoltaic arrays based on the quantum ocean predator-perturbation observation fusion algorithm, comprising the following steps:

[0008] Step 1: Collect photovoltaic array parameters and calculate the current output power; determine the DC-DC converter duty cycle as the core control variable to be optimized, and establish the mapping relationship between the physical duty cycle and the algorithm fitness value;

[0009] Step 2: Normalize the position variables of each candidate individual and map them to the quantum state parameters of the parameterized quantum circuit. Perform quantum operations and measurement sampling under the quantum computing framework, and generate the position correction amount of the candidate individual based on the measurement results.

[0010] Step 3: Based on the search mechanism of the ocean predator algorithm, combined with the position correction amount under the quantum computing framework, perform global exploration and local development update on the candidate individuals to generate a new generation of candidate individuals and update the global optimal individual;

[0011] Step 4: Evaluate whether the current control variable meets the preset neighborhood switching conditions: If not, return to step 2 and continue to execute a new round of quantum global optimization iteration based on the updated population; if it meets the conditions, determine to enter the neighborhood of the global maximum power point and enter the perturbation observation method for local optimization tracking.

[0012] Step 5: Monitor the change in photovoltaic output power in real time during local optimization. When the power change characteristics meet the preset restart conditions, exit the perturbation observation method and return to step 2 to start a new round of quantum global optimization.

[0013] A photovoltaic array maximum power point tracking system based on the quantum ocean predator-perturbation observation fusion algorithm is disclosed. This system implements the aforementioned photovoltaic array maximum power point tracking method based on the quantum ocean predator-perturbation observation fusion algorithm. The system includes a mapping relationship construction module, a position correction generation module, a global optimal individual update module, a neighborhood switching judgment module, and a restart judgment module, wherein:

[0014] The mapping relationship construction module collects photovoltaic array parameters and calculates the current output power; it determines the duty cycle of the DC-DC converter as the core control variable to be optimized and establishes a mapping relationship between the physical duty cycle and the algorithm fitness value.

[0015] The position correction generation module normalizes the position variables of each candidate individual and maps them to the quantum state parameters of the parameterized quantum circuit. It performs quantum operations and measurement sampling under the quantum computing framework and generates the position correction amount of the candidate individual based on the measurement results.

[0016] The global optimal individual update module, based on the search mechanism of the ocean predator algorithm and combined with the position correction amount under the quantum computing framework, performs global exploration and local development updates on candidate individuals, generates a new generation of candidate individuals and updates the global optimal individual;

[0017] The neighborhood switching judgment module evaluates whether the current control variable meets the preset neighborhood switching conditions: if it does not meet the conditions, it returns to the position correction quantity generation module and continues to execute a new round of quantum global optimization iteration based on the updated population; if it meets the conditions, it determines to enter the neighborhood of the global maximum power point and enters the perturbation observation method for local optimization tracking.

[0018] The restart judgment module monitors the change in photovoltaic output power in real time during local optimization. When the power change characteristics meet the preset restart conditions, the perturbation observation method is exited, and the position correction quantity generation module is returned to start a new round of quantum global optimization.

[0019] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the photovoltaic array maximum power point tracking method based on the quantum ocean predator-perturbation observation fusion algorithm.

[0020] A computer device includes a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to perform the photovoltaic array maximum power point tracking method based on the quantum ocean predator-perturbation observation fusion algorithm.

[0021] A computer program product includes computer instructions for causing a computer to execute the photovoltaic array maximum power point tracking method based on the quantum ocean predator-perturbation observation fusion algorithm.

[0022] Compared with the prior art, the significant advantages of this invention are:

[0023] (1) This invention uses the quantum ocean predator algorithm as the first-stage global optimization method. Through the coupling effect of quantum state encoding and ocean predator search mechanism, it improves the search ability for the global maximum power point under local shadow and multi-peak power curve conditions, and reduces the probability of the traditional perturbation observation method getting trapped in local optima.

[0024] (2) The present invention uses the perturbation observation method as the second-stage local tracking method. After the working point has been searched to the neighborhood of the global maximum power point in the first stage, the small step perturbation is used for fine correction, which helps to reduce the search oscillation of the single intelligent optimization algorithm near the maximum power point and improve the steady-state operation performance.

[0025] (3) The present invention establishes a two-stage switching mechanism of “first stage global optimization - second stage local tracking - restarting the first stage due to environmental changes”, which can take into account global search capability, local tracking accuracy and dynamic environment adaptability.

[0026] (4) This invention is applicable to Boost type photovoltaic power generation systems, and also to other photovoltaic conversion systems that achieve maximum power point tracking by adjusting the photovoltaic working voltage reference value or the converter duty cycle. Attached Figure Description

[0027] Figure 1 This is a flowchart illustrating the photovoltaic array maximum power point tracking method based on the quantum ocean predator-perturbation observation fusion algorithm of the present invention.

[0028] Figure 2 The topology diagram of the front-end MPPT verification platform used in this invention.

[0029] Figure 3 This is a schematic diagram of the current-voltage (IU) characteristic curve of a photovoltaic array under partial shading conditions.

[0030] Figure 4 This is a schematic diagram of the multi-peak power-voltage (PU) characteristic curves of a photovoltaic array under partial shading conditions.

[0031] Figure 5 This is a comparison diagram of the maximum power point tracking process of the method of the present invention and the prior art under the condition of partial shadow 1.

[0032] Figure 6 This is a comparison diagram of the maximum power point tracking process of the method of the present invention and the prior art under the local shadow 2 condition. Detailed Implementation

[0033] The maximum power point tracking method for photovoltaic arrays based on the quantum ocean predator-perturbation-observation fusion algorithm is a control strategy that combines the parallel optimization capabilities of quantum computing with classical optimization logic. By introducing parameterized quantum circuits into the ocean predator algorithm framework and integrating the fast local optimization capability of the perturbation-observation method, this algorithm can effectively overcome the problem that traditional algorithms are prone to getting trapped in local optima due to the multi-peak characteristics of the PV characteristic curve of the photovoltaic system under local shading conditions, thus effectively reducing energy loss.

[0034] Combination Figure 1 This invention proposes a method for tracking the maximum power point of a photovoltaic array based on a quantum ocean predator-perturbation observation fusion algorithm, comprising the following steps:

[0035] Step 1: Collect photovoltaic array parameters and calculate the current output power; determine the DC-DC converter duty cycle as the core control variable to be optimized, and establish the mapping relationship between the physical duty cycle and the algorithm fitness value;

[0036] Step 2: Normalize the position variables of each candidate individual and map them to the quantum state parameters of the parameterized quantum circuit. Perform quantum operations and measurement sampling under the quantum computing framework, and generate the position correction amount of the candidate individual based on the measurement results.

[0037] Step 3: Based on the search mechanism of the ocean predator algorithm, combined with the position correction amount under the quantum computing framework, perform global exploration and local development update on the candidate individuals to generate a new generation of candidate individuals and update the global optimal individual;

[0038] Step 4: Evaluate whether the current control variable meets the preset neighborhood switching conditions: If not, return to step 2 and continue to execute a new round of quantum global optimization iteration based on the updated population; if it meets the conditions, determine to enter the neighborhood of the global maximum power point and enter the perturbation observation method for local optimization tracking.

[0039] Step 5: Monitor the change in photovoltaic output power in real time during local optimization. When the power change characteristics meet the preset restart conditions, exit the perturbation observation method and return to step 2 to start a new round of quantum global optimization.

[0040] As a specific example, step 1 involves collecting photovoltaic array parameters, calculating the current output power, determining the DC-DC converter duty cycle as the core control variable to be optimized, and establishing a mapping relationship between the physical duty cycle and the algorithm fitness value, as detailed below:

[0041] Step 1.1: Acquire the output voltage of the photovoltaic array and output current And calculate the current output power of the photovoltaic array based on the sampling results. ;

[0042] Step 1.2: Determine the control variable to be optimized as the duty cycle of the upstream DC-DC converter. Based on the permissible operating range of the photovoltaic array, the safe operating range of the converter, and the system operating constraints, the search interval for the control variable to be optimized is determined, satisfying: ,in This is the lower bound for duty cycle search. This is the upper limit of the duty cycle search;

[0043] Step 1.3: Initialize within the defined search interval The candidate individual, the first Each candidate individual corresponds to a candidate duty cycle. , of which The initial positions of the candidate individuals are represented as follows: ,in For interval Random numbers within;

[0044] Step 1.4: Calculate the fitness value based on the output power of each candidate individual at its corresponding operating point. :

[0045] (1)

[0046] in, To evaluate the window time, the values ​​of the control variables corresponding to each candidate individual are applied to the photovoltaic system, and the candidate duty cycle is considered. After being applied to the Boost converter, Then, the corresponding output voltage and current are collected; These are the real-time values ​​of the output voltage and current of the photovoltaic array; and The duty cycle is the control variable between the current step and the previous step. The duty cycle is used for smoothing weights.

[0047] As a specific example, step 2 involves normalizing the position variables of each candidate individual and mapping them to the quantum state parameters of a parameterized quantum circuit. Quantum operations and measurement sampling are then performed within a quantum computing framework. Based on the measurement results, position correction values ​​for the candidate individuals are generated, as detailed below:

[0048] Step 2.1: Analyze the candidate individual location variables obtained in Step 1. Normalization is performed, and the result is mapped to the rotation angle parameters of the parameterized quantum circuit. ;

[0049] Step 2.2: Construct a parametric circuit based on the quantum computing framework, and... Each qubit is initialized to the ground state. ; for the first A rotation gate is applied to the qubits corresponding to each candidate individual. This causes the qubit to evolve from its ground state to:

[0050] (2)

[0051] in Representing the The quantum states of each candidate individual, in adjacent qubits and Apply a controlled NOT gate CNOT between them to For control bits, A shallow entangled topology is constructed for the target bit; when the control bit is in a superposition state, the search paths between individuals are quantum correlated through CNOT gate operation, transforming the independent search with a single duty cycle into a cooperative evolution in quantum space;

[0052] Step 2.3: Perform Pauli-Z basis projection measurements on the evolved quantum state to obtain the first... The expected value of quantum measurement for each candidate individual , This represents the Pauli-Z operator;

[0053] Step 2.4: Assign the top-ranked candidate individuals based on their current fitness to an elite set. Combine this with the Pauli-Z measurement expectation obtained in Step 2.3 to construct the quantum perturbation correction for each candidate individual, specifically expressed as:

[0054] (3)

[0055] in, For the first During the nth iteration The quantum perturbation correction for each candidate individual; The quantum scaling factor; For the first The candidate individuals in the th The absolute value of the expected value of the Pauli-Z measurement corresponding to the next iteration; For the first Fitness values ​​of each candidate individual; This represents the fitness value of the current best individual. The order of fitness from best to worst is the first... The position of an elite individual; For the first The weight coefficient corresponding to each elite individual; For the first The current position of each candidate individual.

[0056] After step 2 is completed, the candidate individual information containing the quantum measurement expectation value and the quantum perturbation correction amount is output and input into step 3 for position update based on the marine predator mechanism.

[0057] As a specific example, the search mechanism based on the marine predator algorithm described in step 3, combined with the position correction amount under the quantum computing framework, performs global exploration update and local development update on candidate individuals, generates a new generation of candidate individuals, and updates the globally optimal individual, as follows:

[0058] Step 3.1: Integrate the global exploration component with the local development component to construct a unified first-order... During the nth iteration Composite evolution operator for each candidate individual This is used to describe the characteristics of candidate individuals at different search stages:

[0059] (4)

[0060] In the formula, For the first During the nth iteration Levy global exploration component or Brownian motion exploration component of each candidate individual For the global exploration step size coefficient, This represents the step size coefficient for local development. Let the random step size be denoted by the Levy distribution. Let be a standard normally distributed random variable. For the first During the nth iteration The current location of each candidate individual. For the first The position of the globally optimal individual in the population at the next iteration. This represents the maximum number of iterations.

[0061] Step 3.2: Based on the search mechanism of the marine predator algorithm, for the first... The positions of each candidate individual are updated, and the update format for each candidate individual is as follows:

[0062] (5)

[0063] in, This represents a saturation constraint function, used to restrict variables to a certain range. Within the range; This refers to the composite evolutionary increment generated by the marine predator algorithm in step 3.1 (which includes Levy flight and Brownian motion mechanisms). This is the quantum perturbation correction amount in step 2.4;

[0064] Step 3.3: Recalculate the fitness value of each candidate individual and update the current best individual based on the fitness value. ,in For the first The current best individual after the next iteration update. For the first During the nth iteration The fitness value of each candidate individual. The total number of candidate individuals;

[0065] Step 3.4: Perform fitness preservation after updating candidate individuals; when the fitness of the updated candidate individual is better than the fitness before the update, replace the old position with the new position; when the fitness after the update is worse than the fitness before the update, retain the position before the update to reduce the damage of invalid searches to the current excellent individuals.

[0066] As a specific example, step 4 involves evaluating whether the current control variable meets the preset neighborhood switching conditions. If not, the process returns to step 2, and a new round of quantum global optimization iteration continues based on the updated population. If the conditions are met, the process determines whether to enter the neighborhood of the global maximum power point and enters the perturbation-observation method for local optimization tracking, as follows:

[0067] Step 4.1: Steps 2 to 3 are used as the first stage of the quantum ocean predator algorithm's global optimization process. At the start of the first stage, candidate individual initialization is performed. Subsequently, in each iteration, quantum state encoding, quantum measurement, quantum correction generation, position updating, and fitness recalculation are performed cyclically. The maximum number of iterations for the first stage of the quantum ocean predator algorithm is optimally selected. In terms of population size Balancing global search performance with real-time online control under certain conditions;

[0068] Step 4.2: After each iteration, determine whether the candidate individual has entered the neighborhood of the global maximum power point. If not, return to Step 2 based on the updated population to perform the next round of quantum state encoding and search update. If it has entered, and the neighborhood switching condition is met, then change the current global optimal candidate duty cycle. As a switching operating point, the second stage of local continuous tracking using the disturbance observation method is initiated;

[0069] Step 4.3: The neighborhood switching condition must meet at least one of the following:

[0070] First, the normalized mean positional deviation of the updated population relative to the current best individual is no greater than the corresponding value in the previous iteration:

[0071] (6)

[0072] Secondly, the normalized average fitness deviation of the updated population relative to the current best individual is no greater than the corresponding value in the previous iteration:

[0073] (7)

[0074] Third, the normalized variance change rate of the updated population distribution shrinks to within the evolutionary stability region:

[0075] (8)

[0076] in This represents the initial optimal individual position. For the first The current best individual after the next iteration; where, equation (6) represents the population gathering towards the current best individual in the location space; equation (7) represents the population converging towards the current best individual in the fitness space; equation (8) represents the overall dispersion of the population shrinking to the stable region;

[0077] When one of the above conditions is met, it is determined that the first stage has searched the operating point to the vicinity of the global maximum power point;

[0078] Step 4.4: The output of the first stage is the switching operating point. The switching operating point is not directly used as the final steady-state control result, but as the initial operating point for the local tracking of the disturbance observation method in the second stage.

[0079] Step 4.5: In each sampling period, acquire the output power at the current operating point. Output power at the previous sampling time , Represents the current sampling time. Represents the previous sampling time and calculates the power change. ; Calculate the change in control variables , For the control variables at the current operating point, For the control variables at the previous sampling time, the update of the control variables in the second stage satisfies:

[0080] (9)

[0081] in For symbolic functions, The perturbation step size;

[0082] Step 4.6, according to as well as The sign relation determines the perturbation direction for the next sampling period: when When, keep the current perturbation direction unchanged; when At that time, the perturbation direction of the next sampling period is changed;

[0083] As a specific example, step 5 involves real-time monitoring of the photovoltaic output power during local optimization. When the power change characteristics meet the preset restart conditions, the perturbation-observation method is exited, and the process returns to step 2 to start a new round of quantum global optimization, as detailed below:

[0084] Step 5.1: During the second-stage perturbation observation method local tracking process, continuously monitor the changes in photovoltaic output power and adjust the power based on the current sampling period. Power of the previous sampling period Calculate the rate of change of power;

[0085] Step 5.2: When any of the following conditions are met, the operating state of the photovoltaic system is determined to have changed and the global optimization process is restarted:

[0086] (a) Satisfies the restart determination function The algorithm determines that the photovoltaic system's operating status has changed and restarts the first stage of the quantum ocean predator algorithm's global optimization process (steps 2 and 3).

[0087] (10)

[0088] in Indicates the past The standard deviation of photovoltaic output power within each sampling period Indicates the past The arithmetic mean of photovoltaic output power within each sampling period; The discrete coefficients are used to achieve adaptive normalization, so that the criterion can maintain consistent detection sensitivity under both strong and weak light conditions. This is the power normalization factor, used to adjust the weight of the right term in the comprehensive criterion, and is taken as the system rated power or the reference power under the current operating conditions;

[0089] Reflecting the severity of the operating point shift during steady-state tracking. Approaching zero, when a sudden environmental change causes a drastic shift in the PV curve, Increase.

[0090] To preset the judgment threshold, when Exceed When the system is determined to be in an unsteady state, the first stage of quantum global optimization restart is triggered.

[0091] (b) The local tracking operation time of the disturbance observation method reaches the preset restart cycle. It automatically exits the local tracking phase of the perturbation observation method and determines the optimal individual;

[0092] Step 5.3: When the conditions described in Step 5.2 are met, exit the second-stage perturbation observation method local tracking process and re-enter the first-stage quantum ocean predator algorithm global optimization process consisting of Steps 2 to 3.

[0093] Step 5.4: After triggering the re-search, a memory-enhanced restart strategy is adopted: retain the global best individual information obtained from the most recent or multiple recent first-stage runs, reintroduce the current second-stage working point as one of the candidate individuals into the population, and only randomize some candidate individuals, while generating the remaining candidate individuals around the historical advantage region to improve the convergence efficiency in the re-search process.

[0094] This invention also provides a photovoltaic array maximum power point tracking system based on the quantum ocean predator-perturbation observation fusion algorithm. This system is used to implement the aforementioned photovoltaic array maximum power point tracking method based on the quantum ocean predator-perturbation observation fusion algorithm. The system includes a mapping relationship construction module, a position correction generation module, a global optimal individual update module, a neighborhood switching judgment module, and a restart judgment module, wherein:

[0095] The mapping relationship construction module collects photovoltaic array parameters and calculates the current output power; it determines the duty cycle of the DC-DC converter as the core control variable to be optimized and establishes a mapping relationship between the physical duty cycle and the algorithm fitness value.

[0096] The position correction generation module normalizes the position variables of each candidate individual and maps them to the quantum state parameters of the parameterized quantum circuit. It performs quantum operations and measurement sampling under the quantum computing framework and generates the position correction amount of the candidate individual based on the measurement results.

[0097] The global optimal individual update module, based on the search mechanism of the ocean predator algorithm and combined with the position correction amount under the quantum computing framework, performs global exploration and local development updates on candidate individuals, generates a new generation of candidate individuals and updates the global optimal individual;

[0098] The neighborhood switching judgment module evaluates whether the current control variable meets the preset neighborhood switching conditions: if it does not meet the conditions, it returns to the position correction quantity generation module and continues to execute a new round of quantum global optimization iteration based on the updated population; if it meets the conditions, it determines to enter the neighborhood of the global maximum power point and enters the perturbation observation method for local optimization tracking.

[0099] The restart judgment module monitors the change in photovoltaic output power in real time during local optimization. When the power change characteristics meet the preset restart conditions, the perturbation observation method is exited, and the position correction quantity generation module is returned to start a new round of quantum global optimization.

[0100] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the photovoltaic array maximum power point tracking method based on the quantum ocean predator-perturbation observation fusion algorithm.

[0101] The present invention also provides a computer device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the photovoltaic array maximum power point tracking method based on the quantum ocean predator-perturbation observation fusion algorithm.

[0102] The present invention also provides a computer program product, including computer instructions for causing a computer to execute the photovoltaic array maximum power point tracking method based on the quantum ocean predator-perturbation observation fusion algorithm.

[0103] Example

[0104] This embodiment constructs a local shading simulation model of a photovoltaic array and compares it with the traditional maximum power point tracking control method to verify the tracking performance of the method of the present invention under complex lighting conditions.

[0105] A front-end MPPT verification platform was built based on the Python platform. The verification platform includes a photovoltaic array, a boost converter, a DC bus, and a DC load. In this embodiment, the optimization process is implemented using the Python language, and the parameterized quantum circuit is built based on the PennyLane framework. The simulation is performed by calling the lighting.qubit device through the qml.device function.

[0106] The search interval in this method Set as It can cover the feasible control range near the maximum power point, while avoiding insufficient boost due to an excessively small duty cycle or excessive input current and increased system stress due to an excessively large duty cycle; candidate individuals Number of qubits Evaluation window time Pick Duty cycle smoothing weights Set to 0.05; Global exploration step size coefficient Local development step size coefficient Elite Individuals Set to 3; perturbation step size Set to 0.002; power normalization factor Set the value to 2216W; preset the judgment threshold. Take 0.15; for photovoltaic arrays, under standard test conditions, the ambient irradiance is... Ambient temperature The main parameters of a single photovoltaic module are: short-circuit current. Open circuit voltage Peak current Peak voltage Maximum power The photovoltaic array uses a 6-series, 2-parallel connection, meaning each series branch consists of 6 photovoltaic modules connected in series, with 2 parallel branches, resulting in a total of 12 photovoltaic modules in the array. Under standard test conditions, the array's maximum power point voltage is approximately 220.32V, the maximum power point current is approximately 10.06A, and the theoretical maximum output power is approximately 2216W. The input inductor of the Boost converter is... To limit the input current ripple of the Boost converter and ensure continuous change in photovoltaic current; the output capacitor is selected as... To suppress bus voltage ripple and improve the stability of system steady-state operation; the switching frequency is set to 20kHz, and the DC bus reference voltage is set to 400V to meet the Boost output requirements and match the load conditions of this embodiment. The DC load resistor is set to... During the simulation, the power stage integration step size is taken as follows: The MPPT control cycle is set to 2ms.

[0107] The perturbation-observation method uses a fixed-step perturbation approach, with the Boost converter duty cycle as the control variable. The initial duty cycle is set to 0.85; the perturbation step size is... The values ​​are taken as 0.01 and 0.002 respectively, and the MPPT control cycle remains consistent with the method of this invention. In each control cycle, the output voltage and output current of the photovoltaic array are collected, the current output power is calculated, and compared with the previous cycle.

[0108] To verify the maximum power point tracking performance of the method of the present invention under different operating conditions, three test environments were set up:

[0109] (1) Uniform illumination (ideal state): All 12 photovoltaic panels are in uniform illumination. Under the full sunlight.

[0110] (2) Local shadow 1 (severe occlusion):

[0111] Branch 1: 3 boards fully illuminated ( ); 1 board dropped to ; 2 boards dropped to .

[0112] Branch 2: 4 boards fully illuminated ( ); 2 boards dropped to .

[0113] (3) Local shadow 2 (slight occlusion):

[0114] Branch 1: 3 boards fully illuminated ( ); 3 boards dropped to .

[0115] Branch 2: 4 boards fully illuminated ( ); 2 boards dropped to .

[0116] Figure 1 This is a flowchart illustrating the maximum power point tracking method for photovoltaic arrays based on the quantum ocean predator-perturbation-observation fusion algorithm of the present invention. The process includes data acquisition and state feature construction at the environmental perception layer, mode determination at the decision scheduling layer, global search using the quantum ocean predator algorithm and local tracking using the perturbation-observation method at the execution layer, and a restart criterion and smooth migration mechanism. Specific steps have been described in detail in the invention description and specific embodiments. Figure 2 The diagram shows the topology of the front-end MPPT verification platform, which includes a photovoltaic array, a boost converter, a DC bus, and a DC load. Figure 3 The figure shows the IU characteristic curve of a photovoltaic array under partial shading conditions. When the array is partially shaded, the branch current mismatch occurs due to the inconsistent irradiance received by each photovoltaic module. At this time, in order to protect the shaded modules from burning out, the bypass diodes start to conduct, resulting in a significant step-like distortion in the IU curve. Figure 4The figure shows the power output (PU) characteristic curve of a photovoltaic array under partial shading conditions. This figure illustrates the multi-peak characteristics of the output power as a function of voltage when the three modules in the photovoltaic array receive different irradiance levels, including three local maxima. This demonstrates the limitations of the traditional single-peak assumption MPPT algorithm under multi-peak conditions. Figure 5 , Figure 6 The figure shows a comparison of the maximum power point tracking process between the method of this invention and the prior art under local shading conditions 1 and 2. The horizontal axis represents time, and the vertical axis represents the output power of the photovoltaic array. Under different local shading conditions, only the method of this invention can accurately track the global maximum power point in a short time with small steady-state fluctuations. In contrast, the traditional perturbation observation method is prone to getting trapped in local optima and cannot find the global maximum power point. This comparison verifies the superiority of the method of this invention.

[0117] The above are merely preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for maximum power point tracking of photovoltaic arrays based on a quantum ocean predator-perturbation observation fusion algorithm, characterized in that, Includes the following steps: Step 1: Collect photovoltaic array parameters and calculate the current output power; The duty cycle of the DC-DC converter is identified as the core control variable to be optimized, and a mapping relationship between the physical duty cycle and the algorithm fitness value is established. Step 2: Normalize the position variables of each candidate individual and map them to the quantum state parameters of the parameterized quantum circuit. Perform quantum operations and measurement sampling under the quantum computing framework, and generate the position correction amount of the candidate individual based on the measurement results. Step 3: Based on the search mechanism of the ocean predator algorithm, combined with the position correction amount under the quantum computing framework, perform global exploration and local development update on the candidate individuals to generate a new generation of candidate individuals and update the global optimal individual; Step 4: Evaluate whether the current control variable meets the preset neighborhood switching conditions: If not, return to step 2 and continue to execute a new round of quantum global optimization iteration based on the updated population; if it meets the conditions, determine to enter the neighborhood of the global maximum power point and enter the perturbation observation method for local optimization tracking. Step 5: Monitor the change in photovoltaic output power in real time during local optimization. When the power change characteristics meet the preset restart conditions, exit the perturbation observation method and return to step 2 to start a new round of quantum global optimization. Step 2 includes: Step 2.1: Analyze the candidate individual location variables obtained in Step 1. Normalization is performed, and the result is mapped to the rotation angle parameters of the parameterized quantum circuit. ; Step 2.2: Construct a parametric circuit based on the quantum computing framework, and... Each qubit is initialized to the ground state. ; for the first A rotation gate is applied to the qubits corresponding to each candidate individual. This causes the qubit to evolve from its ground state to: (2) in Representing the The quantum states of each candidate individual, in adjacent qubits and Apply a controlled NOT gate CNOT between them to For control bits, A shallow entangled topology is constructed for the target bit; when the control bit is in a superposition state, the search paths between individuals are quantum correlated through CNOT gate operation, transforming the independent search with a single duty cycle into a cooperative evolution in quantum space; Step 2.3: Perform Pauli-Z basis projection measurements on the evolved quantum state to obtain the first... The expected value of quantum measurement for each candidate individual , This represents the Pauli-Z operator; Step 2.4: Assign the top-ranked candidate individuals based on their current fitness to an elite set. Combine this with the Pauli-Z measurement expectation obtained in Step 2.3 to construct the quantum perturbation correction for each candidate individual, specifically expressed as: (3) in, For the first During the nth iteration The quantum perturbation correction for each candidate individual; The quantum scaling factor; For the first The candidate individuals in the th The absolute value of the expected value of the Pauli-Z measurement corresponding to the next iteration; For the first Fitness values ​​of each candidate individual; This represents the fitness value of the current best individual. The order of fitness from best to worst is the first... The position of an elite individual; For the first The weight coefficient corresponding to each elite individual; For the first The current location of each candidate individual; After step 2 is completed, the candidate individual information containing the quantum measurement expectation value and the quantum perturbation correction amount is output and input into step 3 for position update based on the marine predator mechanism; Step 3 includes: Step 3.1: Integrate the global exploration component with the local development component to construct a unified first-order... During the nth iteration Composite evolution operator for each candidate individual This is used to describe the characteristics of candidate individuals at different search stages: (4) In the formula, For the first During the nth iteration Levy global exploration component or Brownian motion exploration component of each candidate individual For the global exploration step size coefficient, This represents the step size coefficient for local development. Let the random step size be denoted by the Levy distribution. Let be a standard normally distributed random variable. For the first During the nth iteration The current location of each candidate individual. For the first The position of the globally optimal individual in the population at the next iteration. This represents the maximum number of iterations. Step 3.2: Based on the search mechanism of the marine predator algorithm, for the first... The positions of each candidate individual are updated, and the update format for each candidate individual is as follows: (5) in, This represents a saturation constraint function, used to restrict variables to a certain range. Within the range; The composite evolutionary increment generated by the marine predator algorithm in step 3.1 includes Levy flight and Brownian motion mechanisms; This is the quantum perturbation correction amount in step 2.4; Step 3.3: Recalculate the fitness value of each candidate individual and update the current best individual based on the fitness value. ,in For the first The current best individual after the next iteration update. For the first During the nth iteration The fitness value of each candidate individual. The total number of candidate individuals; Step 3.4: Perform fitness preservation after updating candidate individuals; when the fitness of the updated candidate individual is better than the fitness before the update, replace the old position with the new position; when the fitness after the update is worse than the fitness before the update, retain the position before the update to reduce the damage of invalid searches to the current excellent individuals.

2. The photovoltaic array maximum power point tracking method based on the quantum ocean predator-perturbation observation fusion algorithm according to claim 1, characterized in that, Step 1 includes: Step 1.1: Acquire the output voltage of the photovoltaic array and output current And calculate the current output power of the photovoltaic array based on the sampling results. ; Step 1.2: Determine the control variable to be optimized as the duty cycle of the upstream DC-DC converter. Based on the permissible operating range of the photovoltaic array, the safe operating range of the converter, and the system operating constraints, the search interval for the control variable to be optimized is determined, satisfying: ,in This is the lower bound for duty cycle search. This is the upper limit of the duty cycle search; Step 1.3: Initialize within the defined search interval The candidate individual, the first Each candidate individual corresponds to a candidate duty cycle. , of which The initial positions of the candidate individuals are represented as follows: ,in For interval Random numbers within; Step 1.4: Calculate the fitness value based on the output power of each candidate individual at its corresponding operating point. : (1) in, To evaluate the window time, the values ​​of the control variables corresponding to each candidate individual are applied to the photovoltaic system, and the candidate duty cycle is considered. After being applied to the Boost converter, Then, the corresponding output voltage and current are collected; These are the real-time values ​​of the output voltage and current of the photovoltaic array; and The duty cycle is the control variable between the current step and the previous step. The duty cycle is used for smoothing weights.

3. The photovoltaic array maximum power point tracking method based on the quantum ocean predator-perturbation observation fusion algorithm according to claim 2, characterized in that, Step 4 includes: Step 4.1: Steps 2 to 3 are used as the first stage of the quantum ocean predator algorithm's global optimization process. At the start of the first stage, candidate individual initialization is performed; subsequently, in each iteration, quantum state encoding, quantum measurement, quantum correction generation, position updating, and fitness recalculation are performed cyclically. The maximum number of iterations for the first stage of the quantum ocean predator algorithm is [value missing]. In terms of population size Balancing global search performance with real-time online control under certain conditions; Step 4.2: After each iteration, determine whether the candidate individual has entered the neighborhood of the global maximum power point. If not, return to Step 2 based on the updated population to perform the next round of quantum state encoding and search update. If it has entered, and the neighborhood switching condition is met, then change the current global optimal candidate duty cycle. As a switching operating point, the second stage of local continuous tracking using the disturbance observation method is initiated; Step 4.3: The neighborhood switching condition must meet at least one of the following: First, the normalized mean positional deviation of the updated population relative to the current best individual is no greater than the corresponding value in the previous iteration: (6) Secondly, the normalized average fitness deviation of the updated population relative to the current best individual is no greater than the corresponding value in the previous iteration: (7) Third, the normalized variance change rate of the updated population distribution shrinks to within the evolutionary stability region: (8) in This represents the initial optimal individual position. For the first The current best individual after the next iteration; where, equation (6) represents the population gathering towards the current best individual in the location space; equation (7) represents the population converging towards the current best individual in the fitness space; equation (8) represents the overall dispersion of the population shrinking to the stable region; When one of the above conditions is met, it is determined that the first stage has searched the operating point to the vicinity of the global maximum power point; Step 4.4: The output of the first stage is the switching operating point. The switching operating point is not directly used as the final steady-state control result, but as the initial operating point for the local tracking of the disturbance observation method in the second stage. Step 4.5: In each sampling period, acquire the output power at the current operating point. Output power at the previous sampling time , Represents the current sampling time. Represents the previous sampling time and calculates the power change. ; Calculate the change in control variables , For the control variables at the current operating point, For the control variables at the previous sampling time, the update of the control variables in the second stage satisfies: (9) in For symbolic functions, The perturbation step size; Step 4.6, according to as well as The sign relation determines the perturbation direction for the next sampling period: when When, keep the current perturbation direction unchanged; when At that time, the perturbation direction of the next sampling period is changed.

4. The photovoltaic array maximum power point tracking method based on the quantum ocean predator-perturbation observation fusion algorithm according to claim 3, characterized in that, Step 5 includes: Step 5.1: During the second-stage perturbation observation method local tracking process, continuously monitor the changes in photovoltaic output power and adjust the power based on the current sampling period. Power of the previous sampling period Calculate the rate of change of power; Step 5.2: When any of the following conditions are met, the operating state of the photovoltaic system is determined to have changed and the global optimization process is restarted: (a) Satisfies the restart determination function The algorithm determines that the photovoltaic system's operating status has changed and restarts the first stage of the quantum ocean predator algorithm's global optimization process (steps 2 and 3). (10) in Indicates the past The standard deviation of photovoltaic output power within each sampling period Indicates the past The arithmetic mean of photovoltaic output power within each sampling period; The discrete coefficients are used to achieve adaptive normalization, so that the criterion can maintain consistent detection sensitivity under both strong and weak light conditions. This is the power normalization factor, used to adjust the weight of the right term in the comprehensive criterion, and is taken as the system rated power or the reference power under the current operating conditions; Reflecting the severity of the operating point shift during steady-state tracking. Approaching zero, when a sudden environmental change causes a drastic shift in the PV curve, Increase; To preset the judgment threshold, when Exceed When the system is determined to be in an unsteady state, the first stage of quantum global optimization restart is triggered. (b) The local tracking operation time of the disturbance observation method reaches the preset restart cycle. It automatically exits the local tracking phase of the perturbation observation method and determines the optimal individual; Step 5.3: When the conditions described in Step 5.2 are met, exit the second-stage perturbation observation method local tracking process and re-enter the first-stage quantum ocean predator algorithm global optimization process consisting of Steps 2 to 3. Step 5.4: After triggering the re-search, a memory-enhanced restart strategy is adopted: retain the global best individual information obtained from the most recent or multiple recent first-stage runs, reintroduce the current second-stage working point as one of the candidate individuals into the population, and only randomize some candidate individuals, while generating the remaining candidate individuals around the historical advantage region to improve the convergence efficiency in the re-search process.

5. A photovoltaic array maximum power point tracking system based on a quantum ocean predator-perturbation observation fusion algorithm, characterized in that, This system is used to implement the photovoltaic array maximum power point tracking method based on the quantum ocean predator-perturbation observation fusion algorithm as described in any one of claims 1 to 4. The system includes a mapping relationship construction module, a position correction generation module, a global optimal individual update module, a neighborhood switching judgment module, and a restart judgment module, wherein: The mapping relationship construction module collects photovoltaic array parameters and calculates the current output power; it determines the duty cycle of the DC-DC converter as the core control variable to be optimized and establishes a mapping relationship between the physical duty cycle and the algorithm fitness value. The position correction generation module normalizes the position variables of each candidate individual and maps them to the quantum state parameters of the parameterized quantum circuit. It performs quantum operations and measurement sampling under the quantum computing framework and generates the position correction amount of the candidate individual based on the measurement results. The global optimal individual update module, based on the search mechanism of the ocean predator algorithm and combined with the position correction amount under the quantum computing framework, performs global exploration and local development updates on candidate individuals, generates a new generation of candidate individuals and updates the global optimal individual; The neighborhood switching judgment module evaluates whether the current control variable meets the preset neighborhood switching conditions: if it does not meet the conditions, it returns to the position correction quantity generation module and continues to execute a new round of quantum global optimization iteration based on the updated population; if it meets the conditions, it determines to enter the neighborhood of the global maximum power point and enters the perturbation observation method for local optimization tracking. The restart judgment module monitors the change in photovoltaic output power in real time during local optimization. When the power change characteristics meet the preset restart conditions, the perturbation observation method is exited, and the position correction quantity generation module is returned to start a new round of quantum global optimization.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the photovoltaic array maximum power point tracking method based on the quantum ocean predator-perturbation observation fusion algorithm as described in any one of claims 1 to 4.

7. A computer device, characterized in that, include: The system includes a memory and a processor, which are interconnected. The memory stores computer instructions, and the processor executes the computer instructions to perform the photovoltaic array maximum power point tracking method based on the quantum ocean predator-perturbation observation fusion algorithm as described in any one of claims 1 to 4.

8. A computer program product, characterized in that, Includes computer instructions for causing a computer to execute the photovoltaic array maximum power point tracking method based on the quantum ocean predator-perturbation observation fusion algorithm as described in any one of claims 1 to 4.

Images