A photovoltaic MPPT control system integrated with shading identification function and a shading identification method

Abstract

The present application relates to the technical field of photovoltaic power generation intelligent monitoring and operation and maintenance, in particular to a photovoltaic MPPT control system integrated with a shading identification function and a shading identification method, comprising a data sampling module, a data normalization module, a single-chip microcomputer control module, a non-periodic shading identification algorithm module, a fault protection module, a communication module and a dynamic historical baseline calculation module. The present application can realize the normalization fusion of multi-source photovoltaic data, accurately distinguish non-periodic shading from meteorological fluctuations and periodic cloud shadows, and reduce the false alarm rate.

Description

Technical Field

[0001] This invention belongs to the field of intelligent monitoring and operation and maintenance technology of photovoltaic power generation, specifically relating to a photovoltaic MPPT control system and shading identification method with integrated shading identification function. Background Technology

[0002] As a crucial component of clean energy, photovoltaic (PV) power generation's operational efficiency and reliability are highly dependent on the health of its PV modules. In real-world operating environments, PV modules are extremely susceptible to damage from factors such as tree branches and leaves, building shadows, dust accumulation, bird droppings, snow accumulation, and mutual shading between modules. Shading not only leads to localized hot spot effects and accelerated module aging but can also cause sudden power drops or even system shutdowns, severely reducing power generation and economic viability. Therefore, achieving rapid, accurate, and automated identification of PV module shading is one of the core requirements for PV operation and maintenance.

[0003] In recent years, with the development of sensing technology and image monitoring technology, photovoltaic shading recognition has gradually evolved towards multi-source data collaboration. Multi-source data has strong complementarity in physical representation, but the data units are inconsistent. How to effectively integrate multi-source data and build a highly robust shading recognition model has become a key technical challenge for the industry.

[0004] The existing technical solutions have the following drawbacks: When relying on a single data point for occlusion judgment, it is easily affected by periodic factors, such as clouds blocking the sun on cloudy days or during the transition from morning to evening, which can easily lead to misjudgment of occlusion.

[0005] Multi-source data, lacking normalization or only undergoing simple ratio calculations, suffers from inconsistent parameters, large dynamic ranges, and interference from non-shading factors such as variations in irradiance (early and late evening), temperature drift, and module degradation. Fixed thresholds are prone to misjudging shading during low-irradiance periods or underreporting due to compensation effects in partially shading situations. Existing solutions lack a normalization mapping mechanism specific to photovoltaic operating conditions, preventing effective fusion of multi-source data within the same feature space.

[0006] The lack of a historical data baseline comparison mechanism makes it difficult to distinguish between normal operating condition fluctuations and actual shading. Existing solutions mostly rely on instantaneous data and fixed reference values ​​for hard threshold judgments, without establishing a historical baseline model that dynamically evolves with region, time, season, and irradiance conditions. When the photovoltaic system is in cloudy weather or during the transition period between morning and evening, normal fluctuations in output power are often misidentified as shading events. The system cannot achieve adaptive calibration based on operating conditions through time-series trend analysis, resulting in poor robustness in identification.

[0007] The occlusion monitoring system architecture relies on the cloud, resulting in poor data real-time performance, susceptibility to network interference, and high deployment costs. Traditional solutions typically upload data to a cloud platform for centralized processing, relying on machine learning models for status diagnosis, which suffers from high communication latency, network outage failures, and high deployment costs. Furthermore, existing equipment data export largely depends on wired or gateway communication, lacking low-power, configuration-free direct connection methods on-site. Maintenance personnel must carry specialized tools for parameter reading and threshold configuration, leading to low deployment and debugging efficiency. Summary of the Invention

[0008] To address the issues of high false alarm rates and poor multi-source data fusion in existing shading identification systems and methods, this invention aims to provide a photovoltaic MPPT control system and shading identification method that integrates shading identification functionality. This system achieves normalized fusion of multi-source photovoltaic data, accurately distinguishes between non-periodic shading and meteorological fluctuations and periodic cloud shadows, and reduces the false alarm rate.

[0009] The technical solution of the present invention is as follows: A photovoltaic MPPT control system with integrated shading recognition function includes a data sampling module, a data normalization module, a microcontroller control module, and a non-periodic shading recognition algorithm module; The data sampling module collects photovoltaic data, including the photovoltaic output voltage, output current, output power, and ambient light intensity. The data normalization module receives photovoltaic data collected by the data sampling module and performs normalization preprocessing on the photovoltaic data. The parameters after normalization preprocessing are input into the non-periodic shading recognition algorithm module. The non-periodic occlusion recognition algorithm module obtains comprehensive feature parameters by weighting and fusing the parameters through preset dynamic weight coefficients, calculates the relative deviation rate based on the comprehensive feature parameters and historical comprehensive feature parameters, and determines whether non-periodic occlusion has occurred based on the relative deviation rate, rate of change and duration. The microcontroller control module runs the MPPT algorithm and the non-periodic occlusion recognition algorithm, as well as data storage, data communication processing, and fault protection processing in parallel.

[0010] Preferably, the data sampling module acquires photovoltaic parameters in real time through an isolated voltage divider circuit and acquires ambient light intensity through an ambient light sensor.

[0011] Preferably, the microcontroller control module is an embedded microcontroller (MCU) that integrates an analog-to-digital converter module, a complementary PWM output module, a floating-point arithmetic unit, and a memory.

[0012] Preferably, it also includes a fault protection module, which is used to cut off the output circuit when the embedded microcontroller MCU detects that the battery overvoltage, battery short circuit, output overload, or other protections are triggered, so as to avoid danger and user loss.

[0013] Preferably, the microcontroller control module has a memory capacity of 512KB and an SRAM capacity of 128KB to meet the needs of data storage and algorithm operation.

[0014] Preferably, it also includes a communication module and an APP installed on the user terminal. The communication module integrates a low-power Bluetooth BLE unit, which transmits photovoltaic data, non-periodic shading warning signals, and fault protection signals to the APP.

[0015] Preferably, it also includes a dynamic historical baseline calculation module for storing historical comprehensive feature parameters and constructing a dynamic historical baseline. The specific method for constructing the dynamic historical baseline is as follows: the day is divided into K 3-minute time windows. For each time window, samples with G(t) ≥ G_clear × 0.9 in the historical data are selected, and the median of the historical comprehensive feature parameters is taken as the dynamic historical baseline. During the process, data at the time of the occlusion event are removed to ensure that the historical comprehensive feature parameters are not contaminated. Here, G_clear represents the theoretical clear sky irradiance, and G(t) represents the real-time value of the current light intensity.

[0016] Preferably, the normalization preprocessing specifically includes: The normalization preprocessing employs a sliding window extremum mapping combined with a weighted calculation strategy, as detailed below: The normalization calculation formula is: x_res = (x(t)-x_min) / (x_max-x_min); Where x_max represents the maximum value of each parameter within the sampling period, x_min represents the minimum value of each parameter within the sampling period, x(t) represents the real-time sampled data of each parameter, and x_res is the normalized calculation result of each parameter; The preprocessed photovoltaic voltage parameters are expressed as: V_res = (V(t)-V_min) / (V_max-V_min); The preprocessed photovoltaic current parameters are expressed as: I_res = (I(t)-I_min) / (I_max-I_min); The preprocessed photovoltaic power parameters are expressed as: P_res = (P(t)-P_min) / (P_max-P_min); The preprocessed results of photovoltaic ambient light parameters are expressed as: G_res = (G(t)-G_min) / (G_max-G_min); After normalizing the voltage V, current I, power P, and illuminance G as described above, a weighted calculation is performed using weighting coefficients for each parameter. The calculation formula is as follows: S = V_r * V_res + I_r * I_res + P_r * P_res + G_r* G_res; Where S represents the photovoltaic comprehensive characteristic parameter, specifically the final result obtained by weighted summation after normalization preprocessing of photovoltaic data, which is the comprehensive characteristic parameter within a certain sampling period; X_r represents the weighting coefficient of each parameter; V_r represents the weight of the photovoltaic voltage value in the comprehensive characteristic parameter after normalization; I_r represents the weight of the photovoltaic current value in the comprehensive characteristic parameter after normalization; P_r represents the weight of the photovoltaic power value in the comprehensive characteristic parameter after normalization; and G_r represents the weight of the photovoltaic ambient light value in the comprehensive characteristic parameter after normalization. In addition, V_r + I_r + P_r + G_r = 1, and P_r≥0.4 and G_r≤0.2, with the weights dynamically adjusted according to the seasonal irradiance interval.

[0017] Preferably, the non-periodic occlusion recognition algorithm module specifically comprises: after normalization preprocessing, photovoltaic data outputs comprehensive feature parameters as input parameters for algorithm execution; and the following calculations are performed only when the real-time data of light intensity is greater than a first threshold: The relative deviation rate D(t) is calculated as follows: D(t) = |S(t) - H(t)| / H(t) * 100%; Where S(t) is the comprehensive feature parameter obtained after weighted calculation, and H(t) is the historical comprehensive feature parameter at the same time. The rate of change R(t) is calculated as follows: R(t) = |S(t) - S(t-1)| / S(t-1) * 100% (τ is the sampling interval); Wherein, S(t-1) is the comprehensive characteristic parameter obtained after weighted calculation at the previous time step; When D(t) > the second threshold, R(t) > the third threshold, and the duration T > the duration threshold T_min, it is determined to be a non-periodic occlusion; If only Δ(t) > the second threshold and R(t) ≤ the second threshold or T ≤ the duration threshold T_min, then it is determined to be a meteorological fluctuation or periodic cloud shadow, and no alarm is triggered; The specific definition of the calculation method for the duration T is as follows: When the threshold conditions are met in the system software design, the duration counter starts counting. For example, if the second and third thresholds are met, the timing starts. When the timing reaches the set time, it is determined to be a non-periodic occlusion. Adding the duration judgment can avoid misjudgment caused by sampling data jitter.

[0018] The first, second, and third thresholds were obtained through testing multiple sets of typical occlusion events. The duration threshold T_min was used to filter short-term cloud occlusion to ensure that the response only applies to non-periodic occlusion.

[0019] A non-periodic shading identification method is proposed, which collects photovoltaic data in real time, including the output voltage, output current, output power and ambient light intensity of the photovoltaic. The photovoltaic data is preprocessed by normalization and mapped to the range of [0,1]. The mapped photovoltaic data is weighted and fused by preset dynamic weight coefficients to obtain comprehensive feature parameters. Based on historical data, a dynamic historical baseline is constructed. When the light intensity is greater than the first threshold, the relative deviation rate and the rate of change are calculated. When the relative deviation rate exceeds the second threshold, the rate of change exceeds the third threshold, and the duration T exceeds the duration threshold T_min, it is determined that non-periodic occlusion has occurred.

[0020] Compared with the prior art, the present invention has the following beneficial effects: This invention employs a three-pronged approach of relative deviation rate, rate of change, and duration for determination. It can identify both the amplitude characteristics of occlusion and the steepness characteristics of sudden changes. Furthermore, by filtering short-term cloud shadows based on duration, it effectively distinguishes between non-persistent occlusion and meteorological fluctuations or periodic cloud shadows, resulting in high accuracy and low false alarm rate.

[0021] In an embodiment of the present invention, by designing a sliding window extreme value mapping, four parameters of different dimensions, namely voltage, current, power and light intensity, are uniformly mapped to the [0,1] interval. Then, combined with dynamic weight coefficients, they are weighted and fused into comprehensive feature parameters, which eliminates the interference caused by differences in dimensions and differences in irradiance intervals, and avoids misjudgment during low irradiance periods and missed reporting of partial shading.

[0022] In the embodiments of the present invention, a time-segmented dynamic baseline is constructed using historical clear sky data, and shading event data is removed to avoid baseline contamination. This allows the reference value to adapt to seasonal, time-segmented, and regional irradiance characteristics, significantly improving the adaptability to operating conditions compared to a fixed threshold scheme. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the hardware architecture of the non-periodic occlusion recognition system described in this invention. Figure 2This is a schematic diagram of the overall workflow of the non-periodic occlusion recognition method described in this invention. Detailed Implementation

[0024] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0025] See Figure 1 The non-periodic occlusion recognition system of the present invention includes a data sampling module, a data normalization module, a microcontroller control module, and a non-periodic occlusion recognition algorithm module. At the hardware level, the core is a microcontroller (MCU) control module, with peripheral connections including a data sampling module, a DC-DC module, an output control module, a battery, a Bluetooth communication module, and a fault protection module. The output terminals of the photovoltaic modules are connected to the output control module and the DC-DC module, respectively. The DC-DC module is electrically connected to the MCU control module. The data sampling module collects electrical parameters and ambient light parameters from the photovoltaic and battery sides and transmits them to the MCU control module. The MCU control module outputs PWM signals to control the DC-DC module and the output control module. The Bluetooth communication module and the fault protection module are both electrically connected to the MCU control module.

[0026] The data normalization module and the aperiodic occlusion recognition algorithm module are integrated into the microcontroller control module (MCU).

[0027] The overall process of system operation is as follows: Figure 2 As shown: After the program is powered on and initialized, it sequentially executes data sampling, Bluetooth communication processing, MPPT algorithm, fault protection processing, and historical power generation calculation. Then, it executes the non-periodic occlusion identification strategy to complete one operating cycle.

[0028] The following provides a detailed explanation of each module and the recognition process: The data sampling module is responsible for raw data acquisition and consists of two parts: electrical quantity acquisition and environmental parameter acquisition. Electrical quantity acquisition uses an isolated voltage divider circuit and a current sampling circuit to obtain the output voltage and current of the photovoltaic module in real time, and the output power is calculated by the microcontroller control module. Environmental parameters are obtained by collecting real-time ambient light intensity through an ambient light sensor. All sampled data is transmitted to the data normalization module at fixed intervals.

[0029] The data normalization module receives the raw sampled data and uses a sliding window extreme value mapping strategy to perform normalization preprocessing on each parameter, mapping parameters with different dimensions and different dynamic ranges to the [0,1] interval, laying the foundation for subsequent multi-source data fusion.

[0030] x_res = (x(t)-x_min) / (x_max-x_min); Where x_max represents the maximum value of each parameter within the sampling period, x_min represents the minimum value of each parameter within the sampling period, x(t) represents the real-time sampled data of each parameter, and x_res is the normalized calculation result of each parameter; The preprocessed photovoltaic voltage parameters are expressed as: V_res = (V(t)-V_min) / (V_max-V_min); The preprocessed photovoltaic current parameters are expressed as: I_res = (I(t)-I_min) / (I_max-I_min); The preprocessed photovoltaic power parameters are expressed as: P_res = (P(t)-P_min) / (P_max-P_min); The preprocessed results of photovoltaic ambient light parameters are expressed as: G_res = (G(t)-G_min) / (G_max-G_min); After normalizing the voltage V, current I, power P, and illuminance G as described above, a weighted calculation is performed using weighting coefficients for each parameter. The calculation formula is as follows: S = V_r * V_res + I_r * I_res + P_r * P_res + G_r* G_res; Where S represents the photovoltaic comprehensive characteristic parameter, specifically the final result obtained by weighted summation after normalization preprocessing of photovoltaic data, which is the comprehensive characteristic parameter within a certain sampling period; X_r represents the weighting coefficient of each parameter; V_r represents the weight of the photovoltaic voltage value in the comprehensive characteristic parameter after normalization; I_r represents the weight of the photovoltaic current value in the comprehensive characteristic parameter after normalization; P_r represents the weight of the photovoltaic power value in the comprehensive characteristic parameter after normalization; and G_r represents the weight of the photovoltaic ambient light value in the comprehensive characteristic parameter after normalization. In addition, V_r + I_r + P_r + G_r = 1, and P_r≥0.4 and G_r≤0.2. The weighting coefficients can be dynamically adjusted according to the irradiation range of different seasons to further improve the adaptability under different working conditions.

[0031] The dynamic historical baseline calculation module is used to construct a reference baseline that changes dynamically over time, providing a benchmark for occlusion determination. The specific construction method is as follows: Divide the day into K time windows, each with a duration of 3 minutes; for each time window, select those that meet the requirements from historical data. Samples with G(t) ≥ G_clear × 0.9 in historical data were selected and compared with theoretical clear-sky irradiance, and samples with large differences from theoretical clear-sky irradiance were removed. The median of the historical comprehensive characteristic parameters is taken as the dynamic historical baseline. During the process, data at the time of occlusion events are removed to ensure that the historical comprehensive characteristic parameters are not contaminated. Among them, G_clear represents the theoretical clear sky irradiance, and G(t) represents the real-time value of the current irradiance.

[0032] During the baseline update process, data from periods identified as occlusion events are automatically removed to prevent occlusion data from contaminating the baseline reference value and to ensure that the baseline always reflects the characteristic level under normal clear sky conditions.

[0033] The non-periodic occlusion recognition algorithm module takes comprehensive feature parameters as input, combines them with a dynamic historical baseline, and achieves occlusion recognition through a three-condition joint judgment. The specific process is as follows: The first step is pre-verification of light intensity: First, it is determined whether the real-time light intensity is greater than the first threshold. If it is not satisfied, it means that the current period is a low-irradiance period (such as night or dawn and dusk), and the data does not have the value of occlusion judgment. Invalid data is directly deleted and will not be included in subsequent calculations.

[0034] The second step is to calculate the characteristic quantities: If the light intensity meets the conditions, then calculate the relative deviation rate and the rate of change respectively. Relative deviation rate D(t): Characterizes the deviation of the current comprehensive characteristic parameter from the historical baseline, and the formula is: The relative deviation rate D(t) is calculated as follows: D(t) = |S(t) - H(t)| / H(t) * 100%; Where S(t) is the comprehensive feature parameter obtained after weighted calculation, and H(t) is the historical comprehensive feature parameter at the same time. Rate of change R(t): Characterizes the degree of instantaneous change in the comprehensive characteristic parameters, and the formula is: The rate of change R(t) is calculated as follows: R(t) = |S(t) - S(t-1)| / S(t-1) * 100%; Wherein, S(t-1) is the comprehensive characteristic parameter obtained after weighted calculation at the previous time step; The third step is to determine the occlusion type: simultaneously verify the following three conditions: 1. The relative deviation rate D(t) > the second threshold; 2. The rate of change R(t) > the third threshold; 3. The duration T that satisfies both of the above conditions is greater than the duration threshold T_min.

[0035] When all three conditions are met simultaneously, it is determined that non-periodic occlusion has occurred, and an alarm signal can be triggered. If only the relative deviation rate condition is met, but the rate of change does not meet the standard or the duration is insufficient, it is determined to be a meteorological fluctuation or periodic cloud shadow, and the obstruction alarm will not be triggered.

[0036] The first threshold, second threshold, third threshold, and duration threshold T_min mentioned above can all be calibrated using test data from multiple sets of typical occlusion events. Among them, the duration threshold T_min is mainly used to filter short-lived cloud occlusions, ensuring that the system only responds to continuous non-periodic occlusion events.

[0037] The microcontroller control module adopts an embedded microcontroller MCU, which integrates an analog-to-digital conversion module, a complementary PWM output module, a floating-point arithmetic unit, and a memory. The program memory capacity is 512KB and the SRAM capacity is 128KB, which can simultaneously support the operation requirements of MPPT algorithm, non-periodic occlusion recognition algorithm, data storage, communication processing, and fault protection.

[0038] The MCU adopts a parallel operation mechanism, and the MPPT algorithm runs independently. It controls the DC-DC module by adjusting the PWM output to make the photovoltaic module work at the maximum power point. The non-periodic shading identification algorithm runs synchronously without interfering with the normal control logic of MPPT.

[0039] The fault protection module uses an MCU to monitor the battery and output circuit status in real time. When abnormal conditions such as battery overvoltage, battery short circuit, or output overload are detected, the MCU triggers the protection logic and controls the output control module to cut off the output circuit, thereby avoiding equipment damage and safety risks and protecting the user's property.

[0040] The communication module integrates a low-power Bluetooth BLE unit, which connects to the microcontroller control module. This allows for the wireless transmission of real-time photovoltaic operation data, non-periodic shading warning signals, and fault protection signals to the user's mobile app. Maintenance personnel can directly view the operating status and receive alarm notifications via the mobile app, eliminating the need for dedicated debugging equipment and significantly improving the convenience of on-site maintenance and debugging.

[0041] A non-periodic occlusion recognition method, corresponding to the above system, comprises the following steps: Step 1: Real-time acquisition of photovoltaic module output voltage, output current, output power, and ambient light intensity; Step 2: Perform sliding window extreme value normalization on the collected parameters and map them to the [0,1] interval; Step 3: Weight the normalized parameters using dynamic weighting coefficients to obtain the comprehensive photovoltaic characteristic parameters; Step 4: Based on historical clear-sky data, construct dynamic historical baselines for different time periods in 3-minute time windows, and remove data from periods of occlusion to avoid baseline contamination; Step 5: Determine if the real-time light intensity is greater than the first threshold. If not, clear invalid data; otherwise, proceed to the next step. Step 6: Calculate the relative deviation rate of the current integrated characteristic parameters relative to the historical baseline, and the rate of change of the integrated characteristic parameters relative to the previous time step; Step 7: If the relative deviation rate exceeds the second threshold, the rate of change exceeds the third threshold, and the duration of the state exceeds the duration threshold, then it is determined that non-periodic shading has occurred; otherwise, it is determined that it is meteorological fluctuation or periodic cloud shadow.

[0042] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A photovoltaic MPPT control system with integrated shading recognition function, characterized in that, It includes a data sampling module, a data normalization module, a microcontroller control module, and a non-periodic occlusion recognition algorithm module; The data sampling module collects photovoltaic data, including the photovoltaic output voltage, output current, output power, and ambient light intensity. The data normalization module receives photovoltaic data collected by the data sampling module and performs normalization preprocessing on the photovoltaic data. The parameters after normalization preprocessing are input into the non-periodic shading recognition algorithm module. The non-periodic occlusion recognition algorithm module obtains comprehensive feature parameters by weighting and fusing the parameters through preset dynamic weight coefficients, calculates the relative deviation rate based on the comprehensive feature parameters and historical comprehensive feature parameters, and determines whether non-periodic occlusion has occurred based on the relative deviation rate, rate of change and duration. The microcontroller control module runs the MPPT algorithm and the non-periodic occlusion recognition algorithm, as well as data storage, data communication processing, and fault protection processing in parallel.

2. The photovoltaic MPPT control system with integrated shading recognition function according to claim 1, characterized in that, The data sampling module acquires photovoltaic parameters in real time through an isolated voltage divider circuit and obtains ambient light intensity through an ambient light sensor.

3. A photovoltaic MPPT control system with integrated shading recognition function according to claim 1, characterized in that, The microcontroller control module adopts an embedded microcontroller (MCU), which integrates an analog-to-digital converter module, a complementary PWM output module, a floating-point arithmetic unit, and a memory.

4. A photovoltaic MPPT control system with integrated shading recognition function according to claim 3, characterized in that, It also includes a fault protection module, which is used to cut off the output circuit when the embedded microcontroller MCU detects that the battery overvoltage, battery short circuit, output overload, or other protections are triggered, in order to avoid danger and user loss.

5. A photovoltaic MPPT control system with integrated shading recognition function according to claim 3, characterized in that, The microcontroller control module has a memory capacity of 512KB and an SRAM capacity of 128KB to meet the needs of data storage and algorithm operation.

6. A photovoltaic MPPT control system with integrated shading recognition function according to claim 1, characterized in that, It also includes a communication module and an APP installed on the user terminal. The communication module integrates a low-power Bluetooth BLE unit, which transmits photovoltaic data, non-periodic shading warning signals, and fault protection signals to the APP.

7. A photovoltaic MPPT control system with integrated shading recognition function according to claim 1, characterized in that, It also includes a dynamic historical baseline calculation module, which is used to store historical comprehensive feature parameters and construct a dynamic historical baseline. The specific method for constructing the dynamic historical baseline is as follows: the day is divided into K 3-minute time windows. For each time window, samples with G(t) ≥ G_clear × 0.9 in the historical data are selected, and the median of the historical comprehensive feature parameters is taken as the dynamic historical baseline. During the process, data when occlusion events occur are removed to ensure that the historical comprehensive feature parameters are not contaminated. Here, G_clear represents the theoretical clear sky irradiance, and G(t) represents the real-time value of the current light intensity.

8. A photovoltaic MPPT control system with integrated shading recognition function according to claim 1, characterized in that, The normalization preprocessing specifically includes: The normalization preprocessing employs a sliding window extremum mapping combined with a weighted calculation strategy, as detailed below: The normalization calculation formula is: x_res = (x(t)-x_min) / (x_max-x_min); Where x_max represents the maximum value of each parameter within the sampling period, x_min represents the minimum value of each parameter within the sampling period, x(t) represents the real-time sampled data of each parameter, and x_res is the normalized calculation result of each parameter; The preprocessed photovoltaic voltage parameters are expressed as: V_res = (V(t)-V_min) / (V_max-V_min); The preprocessed photovoltaic current parameters are expressed as: I_res = (I(t)-I_min) / (I_max-I_min); The preprocessed photovoltaic power parameters are expressed as: P_res = (P(t)-P_min) / (P_max-P_min); The preprocessed results of photovoltaic ambient light parameters are expressed as: G_res = (G(t)-G_min) / (G_max-G_min); After normalizing the voltage V, current I, power P, and illuminance G as described above, a weighted calculation is performed using weighting coefficients for each parameter. The calculation formula is as follows: S = V_r * V_res + I_r * I_res + P_r * P_res + G_r* G_res; Where S represents the photovoltaic comprehensive characteristic parameter, specifically the final result obtained by weighted summation after normalization preprocessing of photovoltaic data, which is the comprehensive characteristic parameter within a certain sampling period; X_r represents the weighting coefficient of each parameter; V_r represents the weight of the photovoltaic voltage value in the comprehensive characteristic parameter after normalization; I_r represents the weight of the photovoltaic current value in the comprehensive characteristic parameter after normalization; P_r represents the weight of the photovoltaic power value in the comprehensive characteristic parameter after normalization; and G_r represents the weight of the photovoltaic ambient light value in the comprehensive characteristic parameter after normalization. In addition, V_r + I_r + P_r + G_r = 1, and P_r≥0.4 and G_r≤0.2, with the weights dynamically adjusted according to the seasonal irradiance interval.

9. A photovoltaic MPPT control system with integrated shading recognition function according to claim 1, characterized in that, The non-periodic occlusion recognition algorithm module specifically works as follows: After normalization preprocessing, the photovoltaic data outputs comprehensive feature parameters as input parameters for algorithm execution. The following calculations are performed only when the real-time light intensity data is greater than a first threshold: The relative deviation rate D(t) is calculated as follows: D(t) = |S(t) - H(t)| / H(t) * 100%; Where S(t) is the comprehensive feature parameter obtained after weighted calculation, and H(t) is the historical comprehensive feature parameter at the same time. The rate of change R(t) is calculated as follows: R(t) = |S(t) - S(t-1)| / S(t-1) * 100%; Wherein, S(t-1) is the comprehensive characteristic parameter obtained after weighted calculation at the previous time step; When D(t) > the second threshold, R(t) > the third threshold, and the duration T > the duration threshold T_min, it is determined to be a non-periodic occlusion; If only Δ(t) > the second threshold and R(t) ≤ the second threshold or T ≤ the duration threshold T_min, then it is determined to be a meteorological fluctuation or periodic cloud shadow, and no alarm is triggered; The first, second, and third thresholds were obtained through testing multiple sets of typical occlusion events. The duration threshold T_min was used to filter short-term cloud occlusion to ensure that the response only applies to non-periodic occlusion.

10. A non-periodic occlusion recognition method, characterized in that, Photovoltaic data is collected in real time, including the output voltage, output current, output power and ambient light intensity of photovoltaics. The photovoltaic data is normalized and preprocessed, and then mapped to the range of [0,1]. The mapped photovoltaic data is weighted and fused by preset dynamic weight coefficients to obtain comprehensive feature parameters. Based on historical data, a dynamic historical baseline is constructed. When the light intensity is greater than the first threshold, the relative deviation rate and the rate of change are calculated. When the relative deviation rate exceeds the second threshold, the rate of change exceeds the third threshold, and the duration T exceeds the duration threshold T_min, it is determined that non-periodic occlusion has occurred.

Images