Intelligent management system and method for flexible photovoltaic module

Abstract

The invention relates to the technical field of flexible photovoltaic, and discloses an intelligent management system and method for a flexible photovoltaic module, and the system comprises the flexible photovoltaic module and an integrated controller. The flexible photovoltaic module is used for converting light energy into electric energy; the integrated controller is used for collecting multi-dimensional state sensing data of the flexible photovoltaic module, and the multi-dimensional state sensing data comprises installation attitude data of the flexible photovoltaic module; obtaining corrected generated power based on the multi-dimensional state sensing data, and predicting a corrected generated power predicted value based on the corrected generated power and the multi-dimensional state sensing data by using a pre-training machine learning model and a Kalman filtering prediction algorithm; and performing maximum power point tracking on the flexible photovoltaic module based on the corrected generated power prediction value so as to adaptively adjust control parameters. According to the invention, fine management of low-power energy is realized, and the total conversion efficiency from light energy to available electric energy is obviously improved.

Description

Technical Field

[0001] This invention relates to the field of flexible photovoltaic technology, and specifically to an intelligent management system and method for flexible photovoltaic modules. Background Technology

[0002] With the development of BIPV (Building Integrated Photovoltaics) and flexible electronics technology, higher requirements are being placed on the flexibility, reliability, and intelligence of photovoltaic systems. Existing flexible photovoltaic systems, such as CN103943697A (A Flexible Photovoltaic Module), have solved the flexibility problem of the cells themselves, but their connection to the external control unit still generally uses traditional wires. The control unit is usually a separate metal box connected to the module via wiring harnesses, a method that is a major point of failure in dynamic bending environments. Another existing technology, CN104979912A (A Photovoltaic System Monitoring Device), describes a method for monitoring photovoltaic modules via wireless communication, but its monitoring target is traditional large-scale photovoltaic power plants, resulting in a complex system. It does not solve the portability problem of the physical connection between the module and the control unit, nor is it deeply integrated with flexible modules. Furthermore, it is unsuitable for low-power applications and lacks predictive intelligent energy management functions.

[0003] It is evident that existing photovoltaic systems, especially those integrated with buildings or curved surfaces, still face the following technical challenges:

[0004] 1. Low system integration and intelligence level: Power generation, management, and communication modules are scattered, resulting in complex wiring, bulky size, and an inability to achieve lightweight, integrated design. Furthermore, existing management methods are not well-suited for low-power energy management strategies and cannot intelligently optimize based on environmental changes and user needs, leading to low overall energy efficiency.

[0005] 2. Limited installation and use: Traditional photovoltaic modules and control systems are connected by welding or plugging copper wires. In flexible application scenarios (such as flexible roofs and curved vehicle bodies), long-term use and deformation can easily lead to fatigue fracture of the connection points, poor sealing, and low system reliability.

[0006] 3. Inconvenient status monitoring and operation and maintenance: Users cannot remotely and in real time monitor the working status of photovoltaic modules (such as voltage, current, power, temperature), lack effective fault early warning and power generation prediction functions, and have high operation and maintenance costs. Summary of the Invention

[0007] To address the aforementioned shortcomings in the prior art, this invention provides a flexible photovoltaic module intelligent management system and method.

[0008] To achieve the above-mentioned objectives, the technical solution adopted by this invention is as follows:

[0009] In a first aspect, the present invention proposes a flexible photovoltaic module intelligent management system, comprising:

[0010] Flexible photovoltaic modules are used to convert light energy into electrical energy;

[0011] An integrated controller is used to collect multi-dimensional state perception data of the flexible photovoltaic module, including the installation posture data of the flexible photovoltaic module; obtain the corrected power generation based on the multi-dimensional state perception data, and use a pre-trained machine learning model and Kalman filter prediction algorithm to predict the corrected power generation based on the corrected power generation and the multi-dimensional state perception data; and perform maximum power point tracking on the flexible photovoltaic module based on the corrected power generation prediction to adaptively adjust the control parameters.

[0012] Optionally, the installation posture data includes the bending angle and radius of curvature of the flexible photovoltaic module.

[0013] Optionally, the multi-dimensional state perception data also includes the output voltage and current of the flexible photovoltaic module, the input voltage and current of the integrated controller, the output voltage and current of the integrated controller, and the light intensity, ambient temperature, and backsheet temperature of the flexible photovoltaic module.

[0014] Optionally, obtaining the corrected power generation based on the multi-dimensional state-aware data includes:

[0015] A corrected model for power generation with respect to installation attitude parameters was calibrated using experimental data;

[0016] Real-time power generation is calculated based on the electrical data in the multi-dimensional state perception data.

[0017] The efficiency degradation coefficient is obtained using a modified model based on the installation posture data of the flexible photovoltaic module;

[0018] The real-time power generation is corrected using the efficiency decay coefficient to obtain the corrected power generation.

[0019] Optionally, predicting the corrected power generation value based on the corrected power generation and the multi-dimensional state-aware data using a pre-trained machine learning model and a Kalman filter prediction algorithm includes:

[0020] The state transition matrix of the Kalman filter prediction algorithm is predicted based on the multi-dimensional state-aware data using a pre-trained machine learning model.

[0021] The optimal state estimation is performed using the Kalman filter prediction algorithm based on the corrected power generation and state transition matrix to obtain the corrected power generation prediction value.

[0022] Optionally, the pre-trained machine learning model includes:

[0023] A training dataset is constructed based on historical multi-dimensional state-aware data and corresponding corrected power generation prediction values.

[0024] Local optimization and updates of machine learning models pre-trained in the cloud are performed using the training dataset.

[0025] Optionally, after the integrated controller collects multi-dimensional state perception data of the flexible photovoltaic module, it includes:

[0026] The real-time power generation is calculated based on the output voltage and current of the flexible photovoltaic module.

[0027] The expected power generation is estimated based on the light intensity and preset efficiency model of the flexible photovoltaic module.

[0028] Determine whether the ratio of the absolute value of the difference between the real-time power generation and the expected power generation to the expected power generation is less than or equal to a set threshold; if so, determine that the current multi-dimensional state perception data is valid data; otherwise, determine that the current multi-dimensional state perception data is abnormal data and maintain the current control parameters.

[0029] Optionally, the integrated controller is further configured to:

[0030] Determine whether the real-time power generation of the flexible photovoltaic module is greater than the real-time power of the load; if so, supply power to the load using the real-time power of the load and charge the energy storage module with the remaining energy; otherwise, supply power to the load using the real-time power generation and supply power to the energy storage module for the power difference.

[0031] Determine whether the corrected predicted power generation is greater than the real-time power generation; if so, increase the charging power to the energy storage module; otherwise, decrease the discharge current of the energy storage module.

[0032] Optionally, the flexible photovoltaic module includes a photovoltaic cell layer, a conductive film circuit layer, and a flexible printed circuit layer stacked together; conductive transmission lines are arranged on one surface of the conductive film circuit layer; an integrated controller is attached to one surface of the flexible printed circuit and electrically interconnected with the conductive transmission lines of the conductive film circuit layer.

[0033] Secondly, this invention proposes an intelligent management method for flexible photovoltaic modules, comprising the following steps:

[0034] Using flexible photovoltaic modules to convert light energy into electrical energy;

[0035] The flexible photovoltaic module is collected using an integrated controller. The multi-dimensional state perception data includes the installation posture data of the flexible photovoltaic module.

[0036] The corrected power generation is obtained by using the integrated controller based on the multi-dimensional state perception data, and the corrected power generation prediction value is predicted based on the corrected power generation and the multi-dimensional state perception data by using a pre-trained machine learning model and Kalman filter prediction algorithm.

[0037] The flexible photovoltaic module is subjected to maximum power point tracking based on the corrected power generation prediction using an integrated controller to adaptively adjust the control parameters.

[0038] The present invention has the following beneficial effects:

[0039] This embodiment utilizes AI dynamic prediction and intelligent control, enabling the system to proactively adjust its operating state and achieve refined management of low-power energy, significantly improving the overall conversion efficiency from light energy to usable electrical energy. The newly added mechanical sensors allow the system to automatically identify and optimize performance under different installation scenarios, expanding its application capabilities on irregular curved surfaces and providing unprecedented installation freedom and adaptive performance. Attached Figure Description

[0040] Figure 1 This is a schematic diagram of the structure of an intelligent management system for flexible photovoltaic modules according to the present invention;

[0041] Figure 2 This is a schematic diagram of the flexible photovoltaic module hierarchy in this invention;

[0042] Figure 3 This is a schematic diagram of the flexible printed circuit layer layout in this invention;

[0043] Figure 4 This is a schematic diagram of the cross-section of the flexible printed circuit layer in this invention;

[0044] Figure 5 This is a schematic diagram of the conductive film circuit layer structure in this invention;

[0045] Figure 6 This is a schematic diagram showing the connection between the flexible printed circuit layer and the conductive film circuit layer in this invention;

[0046] Figure 7 This is a schematic diagram of a flexible photovoltaic module intelligent management method according to the present invention. Detailed Implementation

[0047] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0048] like Figure 1 As shown in the figure, an embodiment of the present invention provides a flexible photovoltaic module intelligent management system, comprising:

[0049] Flexible photovoltaic modules are used to convert light energy into electrical energy;

[0050] An integrated controller is used to collect multi-dimensional state perception data of the flexible photovoltaic module, including the installation posture data of the flexible photovoltaic module; obtain the corrected power generation based on the multi-dimensional state perception data, and use a pre-trained machine learning model and Kalman filter prediction algorithm to predict the corrected power generation based on the corrected power generation and the multi-dimensional state perception data; and perform maximum power point tracking on the flexible photovoltaic module based on the corrected power generation prediction to adaptively adjust the control parameters.

[0051] In an optional embodiment of the present invention, the installation posture data includes the bending angle and radius of curvature of the flexible photovoltaic module. The multi-dimensional state perception data also includes the output voltage and current of the flexible photovoltaic module, the input voltage and current of the integrated controller, the output voltage and current of the integrated controller, and the light intensity, ambient temperature, and backsheet temperature of the flexible photovoltaic module.

[0052] In this embodiment, the integrated controller adopts a highly integrated design and is encapsulated in a small, lightweight protective shell. It can be directly pasted or embedded on the back of the component to form a compact "stick-and-play" system.

[0053] The core components of the integrated controller include a microprocessor (MCU), a power management module (including MPPT), and a multi-dimensional sensor module. The microprocessor is responsible for data acquisition, algorithm execution, status judgment, and communication control; specifically, it includes (1) calculating and detecting the input and output data of the photovoltaic module, directly controlling the power module, and extending four detection circuits from the power module; including the output voltage and current of the photovoltaic cell, the input voltage and current of the power module, and the output voltage and current of the power module; (2) organizing and sending data: it has a built-in wireless communication function and can use the software APP to view and control the working status of the photovoltaic module via Bluetooth or wireless network; (3) status monitoring and alarm: all data is uploaded to the mobile APP via wireless communication. It can perform fault diagnosis based on data analysis (such as abnormal drop in output power may indicate that the module is blocked or damaged) and send alarm information to the user. The power management module realizes the maximum power point tracking of the photovoltaic module, performs efficient DC-DC conversion, charges the energy storage battery and supplies power to the load; its key input and output parameters are monitored by the MCU, including: the output voltage of the photovoltaic module. Output current Power module output voltage Output current The multi-dimensional sensor module includes a light sensor, a temperature sensor, and a force sensor. It is used to collect data in real time, such as ambient light intensity, component temperature, and bending degree. The light sensor collects data on light intensity. (Unit: W / m²); Temperature sensor: monitors ambient temperature With component backsheet temperature Mechanical sensor: detects the bending angle of the component. and rate of change of bending .

[0054] The effectiveness of intelligent management systems is highly dependent on the quality and quantity of data. The required dataset is shown in Table 1:

[0055] Table 1. Required Dataset

[0056]

[0057] In an optional embodiment of the present invention, the integrated controller obtains the corrected power generation based on the multi-dimensional state perception data, including:

[0058] A corrected model for power generation with respect to installation attitude parameters was calibrated using experimental data;

[0059] Real-time power generation is calculated based on the electrical data in the multi-dimensional state perception data.

[0060] The efficiency degradation coefficient is obtained using a modified model based on the installation posture data of the flexible photovoltaic module;

[0061] The real-time power generation is corrected using the efficiency decay coefficient to obtain the corrected power generation.

[0062] The integrated controller uses a pre-trained machine learning model and a Kalman filter prediction algorithm to predict the corrected power generation value based on the corrected power generation and the multi-dimensional state-aware data.

[0063] The state transition matrix of the Kalman filter prediction algorithm is predicted based on the multi-dimensional state-aware data using a pre-trained machine learning model.

[0064] The optimal state estimation is performed using the Kalman filter prediction algorithm based on the corrected power generation and state transition matrix to obtain the corrected power generation prediction value.

[0065] The pre-trained machine learning model includes:

[0066] A training dataset is constructed based on historical multi-dimensional state-aware data and corresponding corrected power generation prediction values.

[0067] Local optimization and updates of machine learning models pre-trained in the cloud are performed using the training dataset.

[0068] After the integrated controller collects multi-dimensional state perception data of the flexible photovoltaic module, it includes:

[0069] The real-time power generation is calculated based on the output voltage and current of the flexible photovoltaic module.

[0070] The expected power generation is estimated based on the light intensity and preset efficiency model of the flexible photovoltaic module.

[0071] Determine whether the ratio of the absolute value of the difference between the real-time power generation and the expected power generation to the expected power generation is less than or equal to a set threshold; if so, determine that the current multi-dimensional state perception data is valid data; otherwise, determine that the current multi-dimensional state perception data is abnormal data and maintain the current control parameters.

[0072] The integrated controller is also used for:

[0073] Determine whether the real-time power generation of the flexible photovoltaic module is greater than the real-time power of the load; if so, supply power to the load using the real-time power of the load and charge the energy storage module with the remaining energy; otherwise, supply power to the load using the real-time power generation and supply power to the energy storage module for the power difference.

[0074] Determine whether the corrected predicted power generation is greater than the real-time power generation; if so, increase the charging power to the energy storage module; otherwise, decrease the discharge current of the energy storage module.

[0075] The overall workflow of this embodiment follows a closed-loop logic of "power generation-management-power consumption-monitoring", specifically including:

[0076] (1) Basic Work Process

[0077] A photovoltaic (PV) system converts solar energy into electrical energy using PV cell modules. After passing through a power module controlled by a controller integrated on the back of the PV modules, part of the electrical energy is used to power the control board itself through a step-down and rectified circuit to maintain its operation; the other part is supplied to the energy storage battery and external output interfaces. The system simultaneously executes two control paths: internally, it performs precise charging and discharging management of the energy storage battery, and downward, it provides stable power to the load side. Through a software platform, it realizes end-to-end status monitoring and remote management, ultimately forming a highly efficient and intelligent PV management system that integrates power generation, energy storage, and power consumption.

[0078] Photovoltaic conversion: Flexible photovoltaic modules generate direct current under illumination, with a theoretical maximum power... With light and temperature Related.

[0079] Power Management and Distribution: The generated power is processed by the power management module in the integrated controller. This module executes the MPPT algorithm to ensure that the components always operate near their maximum power point. Subsequently, the power is divided into two paths: one path, after being stepped down and regulated, powers the control board itself, while the other path charges the energy storage battery or directly supplies external loads.

[0080] Status monitoring and communication: The controller collects system status data in real time through sensors, processes and analyzes the data through a microprocessor, and sends the data to the software platform through a wireless communication module.

[0081] (2) Data processing and intelligent control methods

[0082] Data Processing: The controller collects real-time data on light intensity, module temperature, and module stress through sensors. This data is then combined with pre-input module efficiency, area, and reference voltage-current-light curves. The controller compares this data with the built-in basic reference algorithm and stored historical data, verifying their validity. If the data is deemed valid, the controller corrects relevant settings and adjusts the output to achieve dynamic intelligent regulation of power generation. If the data is deemed invalid, the controller maintains its original operating state. For example, the module controller may have standard voltage-current-light curves. In actual use, the light sensor detects the light intensity at a given time point and calculates the approximate power. The actual output power of the photovoltaic cell is then calculated using the voltage and current detection circuits at the photovoltaic cell terminals. If the difference between the two is within 10%, the data is considered normal and valid. Otherwise, it is considered abnormal, the module maintains the operating state of the previous time period, and the anomaly is recorded.

[0083] Intelligent power supply regulation: When the system detects a decrease in photovoltaic module output or an increase in appliance power consumption, it will reduce the charging current to the energy storage battery and supply energy from the battery to the appliance. At this time, the system will adjust the MPPT (Maximum Power Point Tracking) of the circuit control module to ensure the entire system operates at high efficiency and the appliances run stably. When the photovoltaic module output increases or the appliance power consumption decreases, the system will increase the charging current to the energy storage battery within a certain range (by setting the maximum charging current through a current-limiting resistor). When sunlight is insufficient, photovoltaic cell output decreases, and there is demand for appliance use, the energy storage battery will be switched to power the control board and appliances. When the energy storage battery level drops to a warning value, the control board buzzer will sound, and a push notification will be sent to the user's mobile app via Bluetooth or other wireless networks, suggesting that the user turn off the appliance. If the appliance still needs to be used, this operating state will be maintained until the control board detects that the energy storage battery has entered a low-battery range. The system will then send another low-battery notification, disconnect the module output, and automatically enter sleep mode until the photovoltaic cell power reaches the module startup requirement.

[0084] Personalized control mode: Based on user usage, users can independently control the power supply to appliances, turning them on at the required time or putting them into standby or working mode in advance. Alternatively, users can pre-enter relevant function schemes for specific appliances in the personalized services of the software app, making them more convenient to use and reducing power consumption.

[0085] a. Power generation calculation and data validity verification

[0086] The system calculates the real-time power generation every minute:

[0087] ;

[0088] At the same time, the system uses the light intensity read by the light sensor Based on the component's preset efficiency model, an expected power is estimated. .

[0089] Data validity verification:

[0090] Compare the measured power with the estimated power. If If the deviation is within 10%, it is considered valid data, and the controller can fine-tune the operating parameters accordingly. If the deviation exceeds 10%, it is considered abnormal data, the system maintains its original operating state, and the abnormality is recorded for fault diagnosis.

[0091] b. Dynamic prediction and energy management based on lightweight AI

[0092] Based on sensor data, initial setup data, and voltage and current detection circuits in the controller, a complete dataset covering illumination, temperature, force, voltage, current, and power is constructed. A lightweight machine learning model pre-trained in the cloud is used and then fine-tuned using local data. The optimized model is then updated to the MCU to predict power generation data for short-term future events (e.g., 15 minutes). Based on the prediction results and user habits (e.g., "office worker mode"), energy allocation strategies (e.g., charging and discharging priorities, load switching sequence) are dynamically adjusted.

[0093] By integrating high-precision optical sensors, temperature sensors, and force sensors, comprehensive status monitoring is achieved. A noise removal algorithm based on sliding window mean filtering and dynamic threshold culling ensures data quality. All data is uploaded to a software platform (APP) via wireless communication, providing users with an intuitive display of real-time status, predicted curves, and fault alarm information. Users can also remotely control the system and pre-configure strategies through the software platform.

[0094] A lightweight Kalman filter algorithm is used to analyze the future short-term (e.g., 15-minute) power generation. Making predictions. This process consists of two stages:

[0095] Prediction phase:

[0096] ;

[0097] ;

[0098] in, This represents the predicted state value (e.g., power, voltage) at time k. To predict the error covariance, For process noise, Let be the state transition matrix.

[0099] Correction phase:

[0100] ;

[0101] ;

[0102] ;

[0103] in, For Kalman gain, The sensor measurement value. To measure the noise covariance, This is the observation matrix.

[0104] Based on power prediction results and user habits (such as "office worker mode"), the system dynamically adjusts its energy allocation strategy. For example:

[0105] When it is predicted that the light intensity will decrease, the charging current is reduced in advance to conserve battery power.

[0106] When sufficient sunlight is predicted and users do not have immediate power demand, the charging power of the energy storage battery is increased.

[0107] c. Adaptive optimization based on pose recognition

[0108] Using data from mechanical sensors, the installation tilt angle and bending curvature of the components are calculated through a built-in algorithm. .

[0109] This attitude information is used as a new feature input into the above prediction model and control algorithm to adaptively adjust the control parameters of MPPT, so that the system can maintain high performance in different installation scenarios such as planes and curved surfaces.

[0110] By leveraging mechanical sensor data and employing built-in algorithms, the installation tilt angle and bending curvature of the components are calculated. This attitude information is then coupled into the predictive calculation model as a new feature and used to adaptively adjust the MPPT control parameters. This allows the system to automatically adapt to different installation scenarios, such as planar, curved, and vertical installations, achieving optimal performance for "plug-and-play" operation. When flexible photovoltaic modules are bent, their internal stress distribution, light incidence angle, and electrical connection performance between cell units all change, causing drift in their current-voltage (IV) characteristic curves and maximum power point (MPP).

[0111] Traditional MPPT algorithms, based on the assumption of flat installation, cannot effectively track MPP under bending conditions, resulting in decreased power generation efficiency. To address this issue, this system introduces an adaptive optimization mechanism based on attitude recognition, the core logic of which is as follows:

[0112] The system monitors the bending angle θ and radius of curvature R of the module in real time using mechanical sensors (such as strain gauges or flexible angle sensors) integrated on the back of the module. Based on this raw data, the control chip (MCU) calculates the impact of bending on the module's power generation characteristics using a built-in fundamental mechanical model. This model can be expressed as a correction function, taking attitude data as input and outputting corrections to key electrical parameters. For example:

[0113] ;

[0114] in, The expected power under the same illumination and temperature conditions in a flat state. It is an efficiency decay coefficient less than 1, which is pre-calibrated experimentally.

[0115] The corrected power obtained from the above calculation Alternatively, an equivalent voltage / current correction can be input as prior knowledge into the aforementioned Kalman filter prediction algorithm. This allows the system's power generation prediction model to dynamically adapt to changes in the physical shape of the components, thereby obtaining power prediction values ​​that more closely reflect actual bending conditions. .

[0116] The control chip uses the corrected prediction results After AI processing, the search boundary, step size, or reference voltage / current of the MPPT algorithm are dynamically adjusted. Specifically, the initial operating point or search range of the MPPT controller will be set to approach the maximum power point under bending conditions. The expected region, rather than the flat region. This parameter, as a continuous function over a period of time, generates a data table. The AI-assisted algorithm on the software platform processes this data table to obtain a model. When similar situations arise in subsequent use, the system can adjust the MPPT more quickly, saving time. This allows the MPPT algorithm to start its search from a better starting point, significantly reducing power oscillations and misjudgments caused by MPP drift in bending conditions, accelerating the tracking speed, and enabling the system to predict the true maximum power point during bending more quickly and accurately.

[0117] One of the core objectives of this system is to maximize the utilization of limited photovoltaic energy while ensuring stable load operation. To this end, the system monitors the output power of the photovoltaic modules in real time. Real-time power demand of the load It then implements the following intelligent energy management strategy, whose decision-making logic is based on the dynamic balance between the two:

[0118] (1) Energy management when photovoltaic power is excessive ):

[0119] When the system detects the output power of the photovoltaic modules Exceeding the rated or real-time power of the load At this time, the power management module will prioritize meeting the full power demand of the load. Remaining energy... The energy will be intelligently directed to the energy storage battery to charge it. During this process, the MPPT (Maximum Power Point Tracking) algorithm continuously operates to ensure that the photovoltaic modules always operate at their maximum power point, thereby maximizing the total energy input to the system. and remaining energy All are maximized to achieve optimal energy capture efficiency.

[0120] (2) Energy management when photovoltaic power is insufficient ):

[0121] When photovoltaic output power decreases due to factors such as reduced sunlight, shading of modules, or bending, the efficiency drops. Below load requirements When this happens, the system will automatically enter hybrid power supply mode. At this time, the photovoltaic modules will provide all their power. It will be used primarily to supply the load. Power deficit. The power is then supplemented by energy storage batteries. The power management module intelligently switches and controls the batteries to discharge with precise output power, working together with the photovoltaic output to support the stable operation of the load and prevent the load from being interrupted due to insufficient power supply.

[0122] This energy management strategy works in deep synergy with the aforementioned power generation prediction function. For example, when the system predicts that solar radiation will increase ( When sunlight is expected to rise, more charging capacity can be reserved for the battery in advance; conversely, when sunlight is predicted to weaken, the battery can be controlled to discharge with a smaller current in advance to extend the overall power supply time. All energy flow switching and power distribution are automatically completed by the power management module controlled by the microprocessor (MCU), without user intervention, realizing fully automatic intelligent scheduling of "power generation-energy storage-energy consumption".

[0123] Most existing management systems only possess basic MPPT and remote communication functions, lacking forward-looking power generation prediction and intelligent energy dispatch capabilities. They cannot be optimized according to environmental changes and actual user needs, resulting in low overall energy efficiency. Furthermore, users struggle to remotely and in real-time monitor component operating status (such as voltage, current, power, and temperature), and the lack of effective fault warnings and accurate power generation prediction leads to high operation and maintenance costs and a poor user experience. This embodiment achieves dynamic power generation prediction through a lightweight AI model and performs forward-looking energy dispatch based on the prediction results and user profiles. It integrates multiple types of sensors, ensures data quality through noise reduction algorithms, and constructs a software platform with real-time monitoring, fault diagnosis, and remote control functions. Simultaneously, it utilizes mechanical sensor data to identify component installation posture and automatically adjusts prediction model parameters and control strategies accordingly, achieving cross-scenario performance optimization.

[0124] This embodiment utilizes AI dynamic prediction and intelligent control, enabling the system to proactively adjust its operating status and achieve refined management of low-power energy, significantly improving the overall conversion efficiency from solar energy to usable electrical energy. Users can monitor and control the system anytime, anywhere via a mobile app. The system's accurate fault warning and power generation prediction functions greatly reduce subsequent maintenance costs and psychological burden. The newly added mechanical sensors enable the system to automatically identify and optimize performance under different installation scenarios, expanding its application capabilities on irregular curved surfaces and providing unprecedented installation freedom and adaptive performance.

[0125] (3) Status monitoring and fault early warning:

[0126] All data is uploaded to the mobile app via wireless communication.

[0127] The system performs fault diagnosis based on data analysis (such as a continuous abnormal drop in output power, which may indicate that the component is blocked or damaged) and sends alarm information to the user.

[0128] When the energy storage battery power Reduce to the set threshold In such cases, the system will send an alert via the app and suggest shutting down unnecessary loads to maintain system operation.

[0129] In an optional embodiment of the present invention, the flexible photovoltaic module includes a photovoltaic cell layer, a conductive film circuit layer, and a flexible printed circuit layer stacked together; conductive transmission lines are arranged on one surface of the conductive film circuit layer; an integrated controller is attached to one surface of the flexible printed circuit and electrically interconnected with the conductive transmission lines of the conductive film circuit layer.

[0130] In this embodiment, the flexible photovoltaic module is made of organic materials with semiconductor properties and encapsulated in a flexible composite material through a lamination process, exhibiting excellent flexibility. The module differs from traditional photovoltaic modules in that traditional modules typically consist only of the photovoltaic cells themselves, with external electrical connections relying on wires; while this specially designed flexible photovoltaic module not only includes photovoltaic cell materials but also a circuit transmission layer composed of a conductive film and a circuit layer composed of detachable and assembleable flexible printed circuits (controller circuitry). This design maintains the flexibility of the entire photovoltaic module, improving its installation and fit under various non-planar applications; it also significantly increases the integration of the photovoltaic module, enhancing its portability. The conductive film circuit layer is key to achieving "wireless" operation, employing a flexible conductive film with laser-etched conductive lines to replace traditional copper wires in this part. This connection method is flexible, ultra-thin, and can be bent along with the module, offering extremely high reliability.

[0131] The structure of a flexible photovoltaic module is as follows: Figures 2 to 6 As shown, photovoltaic modules have added conductive film layers and flexible circuit layers, such as... Figure 2 As shown; and a protective layer is then encapsulated outside the circuit layer. Conductive film etching is used instead of transmission wires to integrate the main controller onto the back of the flexible photovoltaic panel for high integration and portability.

[0132] Traditional systems employ a split design, with the control unit externally mounted and welded or plugged into components via rigid copper wires. This structure is prone to fatigue-induced connection point breakage in dynamic bending applications, resulting in poor reliability and a bulky size, making it impossible to achieve a slim and integrated design. This embodiment uses flexible circuitry and thermoforming to achieve a conformal connection between the components and the controller, replacing traditional wires and solving the connection reliability problem in flexible applications. The flexible connection fundamentally solves the wire breakage problem, enabling the system to be stably applied in dynamic bending environments. The highly integrated design greatly simplifies the system structure, reduces installation complexity, and achieves true slimness, portability, and "plug and play" functionality.

[0133] This invention also provides an intelligent management method for flexible photovoltaic modules, such as... Figure 7 As shown, the process includes the following steps S1 to S4:

[0134] S1. Using flexible photovoltaic modules to convert light energy into electrical energy;

[0135] S2. Collect multi-dimensional state perception data of the flexible photovoltaic module using an integrated controller. The multi-dimensional state perception data includes the installation posture data of the flexible photovoltaic module.

[0136] S3. The integrated controller obtains the corrected power generation based on the multi-dimensional state perception data, and the pre-trained machine learning model and Kalman filter prediction algorithm predict the corrected power generation value based on the corrected power generation and the multi-dimensional state perception data.

[0137] S4. The integrated controller performs maximum power point tracking on the flexible photovoltaic module based on the corrected power generation prediction value to adaptively adjust the control parameters.

[0138] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0139] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0140] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0141] Specific embodiments have been used to illustrate the principles and implementation methods of this invention. The descriptions of the embodiments above are only for the purpose of helping to understand the method and core ideas of this invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this invention. Therefore, the content of this specification should not be construed as a limitation of this invention.

[0142] Those skilled in the art will recognize that the embodiments described herein are intended to help the reader understand the principles of the invention, and should be understood that the scope of protection of the invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations based on the technical teachings disclosed in this invention without departing from the spirit of the invention, and these modifications and combinations are still within the scope of protection of this invention.

Claims

1. A flexible photovoltaic module intelligent management system, characterized in that, include: Flexible photovoltaic modules are used to convert light energy into electrical energy; An integrated controller is used to collect multi-dimensional state perception data of the flexible photovoltaic module, including the installation posture data of the flexible photovoltaic module; obtain the corrected power generation based on the multi-dimensional state perception data, and use a pre-trained machine learning model and Kalman filter prediction algorithm to predict the corrected power generation value based on the corrected power generation and the multi-dimensional state perception data. The flexible photovoltaic module is subjected to maximum power point tracking based on the corrected power generation prediction value in order to adaptively adjust the control parameters.

2. The intelligent management system for flexible photovoltaic modules according to claim 1, characterized in that, The installation posture data includes the bending angle and radius of curvature of the flexible photovoltaic module.

3. The intelligent management system for flexible photovoltaic modules according to claim 1 or 2, characterized in that, The multi-dimensional state perception data also includes the output voltage and current of the flexible photovoltaic module, the input voltage and current of the integrated controller, the output voltage and current of the integrated controller, and the light intensity, ambient temperature, and backsheet temperature of the flexible photovoltaic module.

4. The intelligent management system for flexible photovoltaic modules according to claim 1, characterized in that, The corrected power generation obtained based on the multi-dimensional state-aware data includes: A corrected model for power generation with respect to installation attitude parameters was calibrated using experimental data; Real-time power generation is calculated based on the electrical data in the multi-dimensional state perception data. The efficiency degradation coefficient is obtained using a modified model based on the installation posture data of the flexible photovoltaic module; The real-time power generation is corrected using the efficiency decay coefficient to obtain the corrected power generation.

5. The intelligent management system for flexible photovoltaic modules according to claim 1, characterized in that, The prediction of the corrected power generation based on the corrected power generation and the multi-dimensional state-aware data, using a pre-trained machine learning model and a Kalman filter prediction algorithm, includes: The state transition matrix of the Kalman filter prediction algorithm is predicted based on the multi-dimensional state-aware data using a pre-trained machine learning model. The optimal state estimation is performed using the Kalman filter prediction algorithm based on the corrected power generation and state transition matrix to obtain the corrected power generation prediction value.

6. The intelligent management system for flexible photovoltaic modules according to claim 1, characterized in that, The pre-trained machine learning model includes: A training dataset is constructed based on historical multi-dimensional state-aware data and corresponding corrected power generation prediction values. Local optimization and updates of machine learning models pre-trained in the cloud are performed using the training dataset.

7. The intelligent management system for flexible photovoltaic modules according to claim 1, characterized in that, After the integrated controller collects multi-dimensional state perception data of the flexible photovoltaic module, it includes: The real-time power generation is calculated based on the output voltage and current of the flexible photovoltaic module. The expected power generation is estimated based on the light intensity and preset efficiency model of the flexible photovoltaic module. Determine whether the ratio of the absolute value of the difference between the real-time power generation and the expected power generation to the expected power generation is less than or equal to a set threshold; if so, determine that the current multi-dimensional state perception data is valid data; otherwise, determine that the current multi-dimensional state perception data is abnormal data and maintain the current control parameters.

8. The intelligent management system for flexible photovoltaic modules according to claim 1, characterized in that, The integrated controller is also used for: Determine whether the real-time power generation of the flexible photovoltaic module is greater than the real-time power of the load; if so, supply power to the load using the real-time power of the load and charge the energy storage module with the remaining energy; otherwise, supply power to the load using the real-time power generation and supply power to the energy storage module for the power difference. Determine whether the corrected predicted power generation is greater than the real-time power generation; if so, increase the charging power to the energy storage module; otherwise, decrease the discharge current of the energy storage module.

9. The intelligent management system for flexible photovoltaic modules according to claim 1, characterized in that, The flexible photovoltaic module includes a photovoltaic cell layer, a conductive film circuit layer, and a flexible printed circuit layer stacked together; conductive transmission lines are arranged on one surface of the conductive film circuit layer; an integrated controller is attached to one surface of the flexible printed circuit and is electrically interconnected with the conductive transmission lines of the conductive film circuit layer.

10. A method for intelligent management of flexible photovoltaic modules, characterized in that, Includes the following steps: Using flexible photovoltaic modules to convert light energy into electrical energy; The flexible photovoltaic module is collected using an integrated controller. The multi-dimensional state perception data includes the installation posture data of the flexible photovoltaic module. The corrected power generation is obtained by using the integrated controller based on the multi-dimensional state perception data, and the corrected power generation prediction value is predicted based on the corrected power generation and the multi-dimensional state perception data by using a pre-trained machine learning model and Kalman filter prediction algorithm. The flexible photovoltaic module is subjected to maximum power point tracking based on the corrected power generation prediction using an integrated controller to adaptively adjust the control parameters.

Images