Intelligent operation and maintenance management system for distributed photovoltaic project

Abstract

The invention discloses an intelligent operation and maintenance management system for a distributed photovoltaic project, and relates to the technical field of project operation and maintenance management. 18 types of sensing terminals are deployed, 5G + LoRa networking is adopted, millisecond-level snapshot is carried out when abnormity occurs, and data are uploaded after being processed by edge nodes; the digital twin operation and maintenance center constructs a 1: 1 holographic model, maps the equipment state in real time, and simulates the fault condition; the adaptive evolution fault diagnosis module fuses multi-source data to generate a fault tree; the dynamic scheduling module integrates resources to generate an order dispatching scheme and supports unmanned aerial vehicle inspection; the energy efficiency self-adaptive optimization module is linked with energy storage peak regulation and optimizes an MPPT strategy; and the active safety protection module quickly disposes and gives an alarm when detecting abnormity. The method improves the operation and maintenance efficiency and the power generation efficiency of the photovoltaic project, reduces the cost, guarantees the safety, excavates the comprehensive value, and is suitable for various distributed photovoltaic projects.

Description

Technical Field

[0001] This invention relates to the field of project operation and maintenance management technology, and in particular to an intelligent operation and maintenance management system for distributed photovoltaic projects. Background Technology

[0002] Distributed photovoltaic (PV) projects face numerous challenges in operation and maintenance management due to their wide distribution and complex installation environments. Currently, most projects employ a model combining periodic inspections with manual troubleshooting, which suffers from slow response times and high costs. PV modules, constantly exposed to the outdoors, are susceptible to dust accumulation, shadows, and extreme weather, leading to reduced power generation efficiency. Traditional maintenance methods struggle to monitor these dynamic factors in real time, often only addressing issues reactively after a significant drop in power generation. Equipment fault diagnosis relies heavily on manual experience; however, the fault characteristics of inverters and modules vary greatly across different brands and models, resulting in low accuracy of universal diagnostic methods and frequent misdiagnosis and missed diagnoses, thus extending maintenance cycles.

[0003] Existing operation and maintenance (O&M) systems are mostly limited to data collection and simple alarms, lacking refined management of the entire equipment lifecycle. Predictive maintenance capabilities are weak, failing to identify progressive failures such as component aging and inverter performance degradation in advance, often waiting until equipment fails before replacement, increasing unplanned downtime. Resource scheduling is inefficient, with low matching between O&M personnel, equipment, and failure points, frequently resulting in resource mismatches such as "people waiting for failures" or "failures waiting for people," especially in distributed projects with dispersed deployments, leading to low O&M efficiency. Furthermore, most systems do not integrate value-added services such as carbon emission reduction and peak-valley arbitrage, making it difficult to fully explore the comprehensive value of photovoltaic projects.

[0004] With the surge in distributed photovoltaic (PV) installations, traditional operation and maintenance (O&M) models can no longer meet the demands for large-scale, high-reliability management. Small-scale owner projects lack professional O&M teams, resulting in delayed fault handling; large-scale industrial and commercial projects involve numerous and diverse equipment from various brands, making unified management difficult. Simultaneously, the grid's requirements for distributed PV are increasing, and traditional systems struggle to respond quickly to grid dispatch commands, posing grid connection safety risks. These issues hinder the sustainable development of distributed PV, necessitating the construction of an intelligent O&M management system with comprehensive perception, intelligent diagnostics, dynamic dispatching, and energy efficiency optimization capabilities to achieve safe, efficient, and economical operation of PV projects. Summary of the Invention

[0005] This invention proposes an intelligent operation and maintenance management system for distributed photovoltaic projects to solve the problems mentioned in the prior art.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: an intelligent operation and maintenance management system for distributed photovoltaic projects, comprising:

[0007] The global sensing module adopts a star + mesh hybrid topology architecture and deploys sensing terminals including photovoltaic modules, combiner boxes, inverters, and environmental monitoring. It uses 5G slicing technology to prioritize data transmission, and the edge nodes have built-in abnormal pattern recognition algorithms to filter invalid data in real time.

[0008] Digital Twin Operation and Maintenance Hub: Constructs a holographic twin model containing geographic information, equipment parameters, and meteorological data, supporting synchronization between physical equipment and the virtual model; integrates component temperature fields and inverter heat dissipation flow fields for dynamic simulation; integrates historical fault cases for deep learning to generate virtual fault injection models to provide a training environment for operation and maintenance personnel;

[0009] The adaptive evolutionary fault diagnosis module adopts a three-layer architecture of feature extraction, pattern matching, and root cause reasoning. The bottom layer generates a fault fingerprint database through attention feature fusion; the middle layer introduces a transfer learning algorithm to adapt to fault identification of inverters and components across brands; and the root layer builds a knowledge graph to automatically trace the fault propagation path.

[0010] Dynamic resource scheduling platform: Integrates operation and maintenance resources to create digital profiles, including skill tags, equipment compatibility, and historical performance; adopts reinforcement learning algorithms to optimize dispatching, and generates scheduling plans by comprehensively considering the scope of fault impact, repair urgency, and resource accessibility; supports collaborative inspection by drone swarms, and realizes full-area scanning of the power plant through three-dimensional path planning;

[0011] Energy efficiency adaptive optimization hub: Real-time data collection from the generation and load sides to build a dynamic matching model; When the self-consumption rate is <65%, the energy storage system is linked to perform adaptive charging and discharging control, while adjusting the photovoltaic output curve; Equipped with a dust accumulation identification system, combined with precipitation probability prediction to generate the optimal working sequence of the cleaning robot; Dynamically optimize the MPPT tracking strategy to compress the tracking cycle to 1 second when clouds pass by.

[0012] Active safety protection module: Constructs a closed-loop system of perception-decision-execution, including arc monitoring device, dual verification of microwave radar and video analysis, and insulation monitoring instrument; adopts digital filtering technology to eliminate power grid harmonic interference, and identifies potential DC arcing and grounding fault hazards; when a safety threat is detected, it is handled in a graded manner according to the scope of impact, and simultaneously activates a three-level alarm mechanism of audible and visual alarm, operation and maintenance terminal push, and responsible person SMS notification.

[0013] Furthermore, it also includes:

[0014] Predictive health management module and intelligent supply chain collaboration module: Based on equipment degradation characteristics and combined with environmental stress, a remaining life prediction model is constructed to provide early warning of failure risks up to 12 months in advance; a material-level aging acceleration algorithm is developed to shorten the 25-year life assessment cycle of components to 6 months of laboratory testing, and automatically generate differentiated maintenance plans; it connects to the equipment supplier's ERP system to share spare parts inventory data in real time and dynamically adjusts safety stock using replenishment algorithms; a blockchain system for recycling used components is established to record data from the entire process of dismantling and material reuse; and it supports cross-project spare parts allocation.

[0015] Value analysis and interaction modules: Integrate power generation revenue, carbon asset revenue, subsidy revenue, and operation and maintenance cost data to construct a full life cycle value assessment model; generate a sensitivity analysis matrix to provide quantitative basis for project expansion and equipment upgrades; develop a virtual power plant aggregation function to virtualize distributed photovoltaic projects as a single dispatch unit to participate in grid peak shaving and improve peak-valley arbitrage revenue; support real-time annotation of maintenance steps, holographic projection of equipment status display, and natural language interaction on the glasses; set up a four-level permission system for visitors, operation and maintenance, management, and auditing, and use blockchain for operation log storage to support one-click traceability of parameter modification records.

[0016] Furthermore, the digital twin operation and maintenance center is equipped with a spatiotemporal dimension expansion function, enabling access to meteorological station data within a 5km radius, and combining it with a solar trajectory model to predict the distribution of solar radiation received by components over the next 72 hours; it trains a shadow prediction model based on historical shadow data to provide early warning of power generation loss caused by temporary shading 4 hours in advance; and it supports virtual capacity expansion simulation to simulate the impact of adding new components on the overall system efficiency and output the optimal capacity expansion scheme.

[0017] Furthermore, the adaptive evolution fault diagnosis module includes a fault self-healing subsystem, which automatically performs repair operations for common minor faults: adjusting the branch switch of the combiner box to achieve current balancing, remotely restarting the inverter fan to restore normal speed, or adjusting the direction of the cooling fan to optimize heat dissipation efficiency; the self-healing process is recorded throughout; and for faults that cannot be self-healed, a maintenance guide including 3D disassembly diagrams, tool lists, and safety precautions is automatically generated.

[0018] Furthermore, the dynamic resource scheduling platform introduces a digital employee collaboration mechanism to generate a virtual operation and maintenance assistant, which helps engineers complete repetitive tasks such as data verification, report generation, and compliance checks; it also develops an autonomous repair function for drones, which can be equipped with a robotic arm to replace simple parts.

[0019] Furthermore, the predictive health management module develops a material performance degradation monitoring subsystem, which monitors the crosslinking degree of EVA film and the water vapor transmission rate of the backsheet in real time through micro sensors implanted inside the component; combined with outdoor exposure test data, it establishes a correlation model between material performance and power generation efficiency, provides early warning of material aging risk 6 months before power decay; and automatically pushes material maintenance suggestions to extend the actual service life of the component.

[0020] Furthermore, the value analysis module includes a carbon footprint tracking subsystem that records carbon emission data throughout the entire process from component production, transportation, installation to operation and maintenance; it uses the life cycle assessment (LCA) method to calculate the carbon emissions per kilowatt-hour of photovoltaic power generation and generates a carbon reduction report that complies with the ISO 14064 standard; and it connects to the carbon trading platform API interface to support automatic carbon asset declaration and automatic carbon revenue transfer.

[0021] Furthermore, the energy efficiency adaptive optimization center is equipped with grid-friendly control functions, responds to grid dispatch commands in real time, and adjusts active power output and reactive power compensation. A voltage sag tolerance enhancement algorithm is developed to maintain grid-connected operation when the grid voltage drops to 60% of the rated value.

[0022] Furthermore, the proactive security protection module integrates a cybersecurity protection subsystem, adopts a zero-trust architecture, and encrypts all data transmissions using the national standard SM4 encryption; it also includes a built-in intrusion detection system (IDS); and a firmware security update mechanism is developed to ensure uninterrupted system operation during device program updates.

[0023] Compared with existing technologies, the beneficial effects of this invention are:

[0024] The all-domain perception module covers multi-dimensional data of components, devices, and environment, and combined with 5G slicing technology, it ensures priority transmission of key data, providing real-time basis for accurate decision-making. The digital twin operation and maintenance center builds a holographic model that can dynamically simulate equipment status and fault conditions, providing a virtual environment for operation and maintenance training and simulating scenarios such as shadows and capacity expansion in advance, helping to optimize decision-making.

[0025] The adaptive evolutionary fault diagnosis module significantly improves the accuracy and efficiency of fault identification through a multi-layer architecture and knowledge graph. Combined with the fault self-healing subsystem, it can automatically repair common minor faults and reduce manual intervention. The dynamic resource scheduling platform integrates multi-dimensional resources, optimizes dispatching and inspection paths through intelligent algorithms, and introduces digital employees and drone collaboration, which greatly improves the utilization rate and response speed of operation and maintenance resources.

[0026] The adaptive energy efficiency optimization center achieves dynamic matching between power generation and load, and improves power generation efficiency by combining dust accumulation identification and MPPT strategy optimization. The active safety protection module constructs a multi-layer protection system to ensure safe system operation while enhancing grid friendliness. The predictive health management and supply chain collaboration module extends equipment life and reduces operation and maintenance costs, while the multi-dimensional value analysis module explores value-added benefits such as carbon emission reduction, comprehensively improving the overall benefits of photovoltaic projects. It is suitable for distributed photovoltaic projects of different scales and types. Attached Figure Description

[0027] Figure 1 This is a schematic block diagram of an intelligent operation and maintenance management system for distributed photovoltaic projects proposed in this invention;

[0028] Figure 2 This is a schematic diagram illustrating the fault diagnosis module identification efficiency of an intelligent operation and maintenance management system for distributed photovoltaic projects proposed in this invention. Detailed Implementation

[0029] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0031] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified. Furthermore, the terms "installed," "connected," and "linked" should be interpreted broadly; for example, they may refer to a fixed connection, a detachable connection, or an integral connection; they may refer to a mechanical connection or an electrical connection; they may refer to a direct connection or an indirect connection through an intermediate medium; and they may refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances. The invention will now be described in further detail with reference to the accompanying drawings.

[0032] Reference Figures 1 to 2 A smart operation and maintenance management system for distributed photovoltaic projects, comprising:

[0033] The global sensing module adopts a hybrid "star + mesh" topology architecture, deploying 18 types of intelligent sensing terminals, covering photovoltaic modules (string-level current sensors, 0-25A range, ±0.02A accuracy; distributed fiber optic temperature measurement, spatial resolution 1m, temperature measurement accuracy ±0.2℃), combiner boxes (multi-channel current acquisition modules, sampling rate 1kHz), inverters (high-frequency power sensors, 0-800kW range, ±0.1%FS accuracy), and environmental monitoring (full-spectrum radiometer, 350-1100nm band coverage, ±2W / m). 2 Accuracy; Atmospheric visibility meter, measurement range 10m-50km; Data transmission priority is achieved through 5G slicing technology (fault data preempts bandwidth, transmission latency <10ms), edge nodes have built-in abnormal pattern recognition algorithm, which can filter invalid data in real time (rejection rate ≥90%) and compress it for uploading (compression ratio 10:1).

[0034] Digital Twin Operation and Maintenance Hub: Constructs a holographic twin model containing geographic information (accuracy ±0.05m), equipment parameters (3D to component level), and meteorological data (updated minute-by-minute), supporting millisecond-level synchronization between physical equipment and the virtual model; integrates component temperature field and inverter heat dissipation flow field for dynamic simulation, with simulation results deviating from measured data by <2%; integrates historical fault cases (over 100,000) for deep learning to generate a virtual fault injection model, which can simulate system response under 300+ fault conditions, providing an immersive training environment for operation and maintenance personnel.

[0035] The adaptive evolutionary fault diagnosis module adopts a three-layer architecture of "feature extraction - pattern matching - root cause reasoning". The bottom layer generates a fault fingerprint database (containing 500+ types of fault features) through attention feature fusion (infrared thermal image texture + acoustic spectrum + vibration waveform); the middle layer introduces a transfer learning algorithm, which can identify faults in inverters (adaptability ≥95%) and components (adaptability ≥98%) across brands; the root layer constructs a knowledge graph (containing 1000+ causal association rules), which can automatically trace the fault propagation path (such as "diode grid voltage fluctuation → inverter overcurrent → capacitor aging → string power failure"), with fault location accuracy at the component level (error <0.5m) and diagnosis time <10 seconds.

[0036] Dynamic Resource Scheduling Platform: Integrates operation and maintenance resources (500+ certified engineers, 20+ drones, 30+ special vehicles) to create digital profiles, including skill tags (100+ sub-fields), equipment compatibility, and historical performance; adopts reinforcement learning algorithms for dispatch optimization, and generates scheduling plans by comprehensively considering the scope of fault impact (covering installed capacity), repair urgency (power generation loss rate), and resource accessibility (real-time traffic conditions), improving response speed by 40%; supports collaborative inspection of drone clusters (simultaneous dispatch of 5 drones), and achieves full-area scanning of the power station through three-dimensional path planning (repetition coverage rate <3%), improving inspection efficiency by 3 times.

[0037] Energy efficiency adaptive optimization hub: Real-time acquisition of data from the generation side (module power curve, inverter efficiency) and the load side (electricity waveform, peak and valley periods) to build a dynamic matching model; When the self-consumption rate is <65%, the energy storage system is linked to perform multi-period charging and discharging control (charging from 11:00 to 15:00 in the afternoon and discharging from 18:00 to 21:00 in the evening), while adjusting the photovoltaic output curve (achieving smooth output through reactive power compensation); Equipped with a dust accumulation intelligent recognition system (based on image grayscale value analysis), combined with precipitation probability prediction (for the next 48 hours), the optimal working sequence of the cleaning robot is generated to ensure that the module power generation efficiency is restored to more than 99% of the initial value after cleaning; Dynamically optimize the MPPT tracking strategy, compressing the tracking cycle to 1 second under the condition of rapid cloud passing, and improving the maximum power point tracking accuracy to 99.5%.

[0038] Active safety protection module: Constructs a closed-loop system of "perception-decision-execution", including an arc monitoring device (detection sensitivity 0.1A), intelligent fence (microwave radar + video analysis dual verification), and insulation monitoring instrument (measurement range 0-1000MΩ); adopts digital filtering technology to eliminate power grid harmonic interference (total harmonic distortion rate <3%), accurately identifies hidden dangers such as DC arcing and grounding faults; when a safety threat is detected, it is handled in a graded manner according to the scope of impact (local fault isolation time <0.05s, emergency shutdown time of the entire station <0.2s), and simultaneously activates a three-level alarm mechanism of audible and visual alarm (sound pressure level ≥95dB), operation and maintenance terminal push, and responsible person SMS notification.

[0039] This invention also includes:

[0040] Predictive Health Management Module and Supply Chain Intelligent Collaboration Module: The predictive health management module constructs a remaining life prediction model based on equipment degradation characteristics (growth rate of dark spot area in component EL images, fluctuation range of inverter IGBT junction temperature) and environmental stress (cumulative ultraviolet radiation, diurnal temperature range), which can provide early warning of potential failure risks up to 12 months in advance; it develops a material-level aging acceleration algorithm to shorten the 25-year life assessment cycle of components to 6 months of laboratory testing; and it automatically generates differentiated maintenance plans (e.g., components in high-irradiance areas are inspected every 2 years, and components in shaded areas are inspected every 1 year). The supply chain intelligent collaboration module connects to the ERP systems of 50+ equipment suppliers, shares spare parts inventory data in real time (updated every minute), and uses an intelligent replenishment algorithm (considering historical failure rates and transportation cycles) to dynamically adjust safety stock; it establishes a blockchain system for the recycling of used components, recording data for the entire process of dismantling and material reuse (tamper-proof), achieving a recycling rate of ≥95%; and it supports cross-project spare parts allocation (based on geographical location and fault urgency), with an allocation response time of <2 hours.

[0041] The module includes a multi-dimensional value analysis module and an immersive interactive module. The multi-dimensional value analysis module integrates data on power generation revenue (time-of-use electricity price difference), carbon asset revenue (real-time carbon trading price), subsidy revenue (policy updates), and operation and maintenance costs (manpower + materials + energy consumption) to construct a full life-cycle value assessment model. It generates a sensitivity analysis matrix (weighting the impact of factors such as irradiance, electricity price, and discount rate on revenue), providing quantitative basis for project expansion and equipment upgrades. It also develops a virtual power plant aggregation function, which can virtualize 10+ distributed photovoltaic projects into a single dispatch unit to participate in grid peak shaving, improving peak-valley arbitrage profits. The immersive interactive module supports AR remote guidance (real-time annotation of maintenance steps on smart glasses), holographic projection of equipment status (360° unobstructed view of internal structure), and natural language interaction (supporting 200+ operation and maintenance commands with an accuracy rate ≥98%). It sets up a four-level permission system (visitor, operation and maintenance, management, and audit), and operation logs are stored on blockchain (retained for 10 years), supporting one-click traceability of any parameter modification record (including the modifier, time, and terminal information).

[0042] In this invention, the digital twin operation and maintenance center has a spatiotemporal dimension expansion function, which can access meteorological station data (wind speed, precipitation, cloud movement speed) within a 5km radius and combine it with a solar trajectory model (calculation accuracy ±0.05°) to predict the distribution of solar radiation received by components in the next 72 hours; based on historical shadow data (more than 3 years) to train a shadow prediction model, it can provide an early warning of power generation loss caused by temporary shading (such as vehicle parking, bird flocks) 4 hours in advance; it supports virtual capacity expansion simulation to simulate the impact of adding new components (different power, tilt angle) on the overall system efficiency and output the optimal capacity expansion scheme (including cable current carrying capacity verification and inverter capacity matching).

[0043] In this invention, the adaptive evolutionary fault diagnosis module includes a fault self-healing subsystem, which can automatically perform repair operations for more than 20 common minor faults (such as string current deviation <5% and abnormal inverter fan speed): achieving current balancing by adjusting the combiner box branch switch, remotely restarting the inverter fan to restore normal speed, or adjusting the direction of the cooling fan to optimize heat dissipation efficiency; the entire self-healing process is recorded (including operation instructions, execution results, and parameter change curves), with a self-healing success rate of ≥85%; for faults that cannot be self-healed, a maintenance manual including 3D disassembly diagrams, tool lists, and safety precautions is automatically generated, reducing the reliance on senior technicians.

[0044] In this invention, the dynamic resource scheduling platform introduces a digital employee collaboration mechanism, which can generate an AI virtual operation and maintenance assistant to help engineers complete repetitive tasks such as data verification (automatically checking the accuracy of sensor readings), report generation (automatically filling in test data according to national standard formats), and compliance checks (automatically marking non-compliant items against 100+ operation and maintenance standards), reducing manual operation by more than 60%. Furthermore, the invention develops an autonomous repair function for drones, equipped with a robotic arm (positioning accuracy ±0.5mm) to complete simple component replacements (such as fuses and connectors), reducing repair time by 50% compared to manual repair.

[0045] In this invention, the predictive health management module develops a material performance degradation monitoring subsystem, which uses micro-sensors (size < 5 mm) implanted inside the component to monitor in real time the crosslinking degree of the EVA film (measurement accuracy ±1%) and the water vapor transmission rate of the backsheet (measurement range 0-10 g / m³). 2 • day); Combine outdoor exposure test data (1000+ module samples) to establish a correlation model between material performance and power generation efficiency, which can provide early warning of material aging risk 6 months before power decay; automatically push material maintenance suggestions (such as coating the backsheet with an anti-UV coating) to extend the actual service life of the module by 3-5 years.

[0046] In this invention, the multidimensional value analysis module includes a carbon footprint tracking subsystem that records carbon emission data throughout the entire process from component production (silicon purification, cell manufacturing), transportation (distance, mode), installation (energy consumption) to operation and maintenance (electricity consumption); it uses the life cycle assessment (LCA) method to calculate the carbon emissions per kilowatt-hour of photovoltaic power generation (accuracy ±5gCO2eq / kWh) and generates a carbon reduction report that conforms to the ISO14064 standard; it connects to the carbon trading platform API interface to support automatic carbon asset declaration (declaration accuracy 100%) and achieve automatic carbon revenue transfer.

[0047] In this invention, the energy efficiency adaptive optimization hub has grid-friendly control functions, which can respond to grid dispatch commands in real time (response time < 1 second), adjust active power output (adjustment range 0-100% rated power, adjustment accuracy ±2%), and reactive power compensation (leading / lagging 0-100% rated reactive power); a voltage sag tolerance enhancement algorithm is developed, which can still maintain grid-connected operation (duration ≥ 2 seconds) when the grid voltage drops to 60% of the rated value; and harmonic active suppression technology is adopted to make the total harmonic distortion rate of the inverter output current < 2%, which meets the highest level requirements of GB / T14549-93.

[0048] In this invention, the proactive security protection module integrates a cybersecurity protection subsystem, adopts a zero-trust architecture (verifying terminal identity every 10 minutes), and encrypts all data transmissions using the national standard SM4 encryption (256-bit encryption strength). It has a built-in intrusion detection system (IDS) capable of identifying over 300 types of network attack characteristics (such as SQL injection and man-in-the-middle attacks), with an attack detection rate ≥99%. A firmware security update mechanism (using OTA differential upgrade) is developed to ensure uninterrupted system operation during device program updates (100% update success rate) and prevent malicious code injection.

[0049] A specific implementation method for an intelligent operation and maintenance management system for distributed photovoltaic projects.

[0050] Example 1: Commercial and Industrial Distributed Solar Rooftop Project (10MW)

[0051] A 10MW distributed photovoltaic project in an industrial park covers the rooftops of 15 factory buildings, employing a hybrid network of multi-brand modules (Brand A 330W, Brand B 340W) and inverters (Brand C 50kW, Brand D 60kW). During system deployment, a cascade current sensor (0-25A range) is installed in every string of 20 modules connected in series in the global sensing module. Distributed fiber optic temperature sensors (1m spatial resolution) are embedded between modules. The combiner box incorporates a multi-channel current acquisition module (1kHz sampling rate). High-frequency power sensors are connected to the inverters. Full-spectrum radiometers and atmospheric visibility meters are deployed around the factory area. 5G slicing technology is used to divide the fault data transmission channels (latency <10ms). Edge nodes perform anomaly filtering on the collected temperature and current data (92% rejection rate), compress the data, and upload it to the cloud.

[0052] The digital twin operation and maintenance center imports factory CAD drawings (accuracy ±0.05m) to construct a holographic model, including the installation location, tilt angle (30°), orientation (due south), and cable routing of each component. It synchronizes the component temperature field (generated from fiber optic temperature measurement data to create a thermal map) and inverter cooling fan speed (1500-3000rpm) in real time, with simulation deviation <2%. The simulation measures the shading effect of a 30m high cooling tower on the western component during the summer hours of 12:00-14:00, showing a 12% output shading loss rate. It recommends installing an adjustable bracket (adjustment range 20°-40°), which reduces the loss to 3% after adjustment.

[0053] During the operation of the adaptive evolutionary fault diagnosis module, a 15% current deviation was detected in a string of a C-brand inverter (exceeding the threshold). The bottom-level edge node located the suspicious area within 1 second; the middle layer called infrared thermal imaging (1280×1024 resolution) and acoustic fingerprint data, and identified diode failure within 30 seconds; the knowledge graph traced back to "diode failure → recent frequent voltage fluctuations → capacitor aging", generating a maintenance guide (including 3D disassembly diagram). The fault self-healing subsystem automatically adjusted the combiner box branch switch for another string with a 5% current deviation, restoring balance within 5 minutes.

[0054] After receiving a repair order, the dynamic resource scheduling platform matched an engineer within 3km who was skilled in repairing Brand C inverters (90% skill match), and planned an optimal route to avoid the morning rush hour (reducing time by 22%). Simultaneously, two drones (with a 40-minute flight time) were dispatched to conduct a cluster inspection of 15 rooftops. The 3D path planning reduced the overlap coverage to 2%, and the onboard 20-megapixel camera identified 3 microcracks (0.2mm) and 10 bird droppings (8cm in area). 2 Automatically associate maintenance work orders.

[0055] The energy efficiency adaptive optimization center monitored that the photovoltaic output (8MW) far exceeded the load (3MW) between 11:00 AM and 3:00 PM, with a self-consumption rate of 62%. This triggered a charging of the 500kWh energy storage system (charge / discharge efficiency of 92%), while simultaneously adjusting the inverter's reactive power output curve to smooth it out. The intelligent dust accumulation identification system, through image grayscale analysis, found that the dust coverage of the north-side modules was 25%. Combined with a 48-hour forecast of no precipitation, two cleaning robots (efficiency 1000 modules / hour) were dispatched, and after cleaning, the efficiency was restored to the initial value of 99.2%.

[0056] The active safety protection module detects that the insulation resistance in a certain area drops to 0.8MΩ (below the threshold of 1MΩ), isolates the branch within 0.05 seconds, activates a 95dB audible and visual alarm, and simultaneously pushes the alarm to the maintenance terminal and the responsible person's mobile phone.

[0057] Operation and maintenance metrics Traditional Operations and Maintenance System maintenance Fault response time 4 hours 30 minutes Annual power generation 12 millionkWh 13.2 million kWh Number of maintenance personnel 15 people 8 people Unplanned downtime 80 hours / year 12 hours / year

[0058] Traditional O&M relies on manual inspections, resulting in delayed fault response. A 10MW project requires 15 people for maintenance, with 80 hours of unplanned downtime annually and a power generation of 12 million kWh. This system, through comprehensive perception and intelligent diagnostics, reduces fault response time to 30 minutes and unplanned downtime to 12 hours; energy efficiency optimization measures increase power generation by 10% (to 13.2 million kWh); dynamic scheduling and drone inspections reduce manpower to 8 people, significantly lowering labor costs. Digital twins and predictive maintenance proactively mitigate the risks of shadow wear and equipment aging, ensuring stable system operation through proactive safety protection, fully demonstrating the advantages of intelligent O&M in large-scale industrial and commercial projects.

[0059] Example 2: Residential Distributed Photovoltaic Project (5kW)

[0060] A farmer's 5kW residential photovoltaic (PV) project is installed on the roof of his house, using 20 250W modules (tilted at 25°), a 5kW inverter, and a 2kWh energy storage battery. A global sensing module is used, with current sensors (0-15A) installed at the string end, temperature sensors (±0.2℃ accuracy) attached to the module surface, and a small irradiance meter installed outside the yard wall. Data is transmitted to the edge node via LoRa (latency <50ms), and invalid data is filtered out (88% rejection rate) before being uploaded.

[0061] A simplified model of the rooftop photovoltaic system (accuracy ±0.1m) is constructed using a digital twin operation and maintenance center to synchronize module temperature (25-60℃) and inverter output power (0-5kW). Simulations are performed for the winter hours of 9:00-15:00, considering the shading effect of trees on the eastern modules, resulting in an 8% output loss rate. Pruning branches (keeping a height <2m) is recommended to reduce the loss to 1.5%.

[0062] The adaptive evolution fault diagnosis module detected a sudden drop of 30% in string current and located the faulty string in 1 second; it called up the component photos taken by the mobile APP (with user assistance) and combined them with historical data (large temperature fluctuations in the past 3 days) to diagnose poor contact in the junction box in 10 seconds; and generated a maintenance manual with image annotations (steps: power off → open junction box → re-tighten terminals).

[0063] After receiving the work order, the dynamic resource scheduling platform matches maintenance personnel (with an 85% skill matching rate) within 5km of the township and plans the optimal route on rural roads (avoiding field paths). Because the fault is simple, drones are not dispatched; instead, maintenance personnel carry basic tools (insulated wrench, multimeter) to the location.

[0064] The energy efficiency adaptive optimization center monitors that during the daytime (10:00-14:00), the photovoltaic output (4-5kW) exceeds the household load (1-2kW), with a self-consumption rate of 60%, and instructs the energy storage battery to charge (fully charged between 11:00-14:00); in the evening (18:00-20:00), the load increases (2-3kW), and the battery discharges to supplement power supply. MPPT tracking compresses the cycle to 1 second in cloudy weather, increasing power generation.

[0065] The active safety protection module detected that the grounding resistance rose to 4.5Ω (standard 4Ω) during thunderstorms, triggering an early warning (sound pressure level 90dB) and sending a text message suggesting that the system be checked after the rain. It also detected a sudden drop in grid voltage of 15% and adjusted the inverter's reactive power output within 0.5 seconds to maintain grid-connected operation.

[0066] The predictive health management module analyzed the component's EL image and found that the dark spot area of ​​one component was increasing by 0.5% monthly, with a predicted remaining lifespan of 8 years. It was recommended to focus on monitoring it after 3 years. The supply chain collaboration module showed that nearby township service points had two units of this model in stock, which could meet the replacement needs.

[0067]

[0068]

[0069] Traditional residential photovoltaic (PV) maintenance suffers from a large service radius, requiring two days for fault repair and incurring an average annual maintenance cost of 600 yuan, while generating only 5500 kWh of electricity. This system, through lightweight sensing and simplified diagnostic processes, completes repairs within four hours. Homeowners can assist by taking photos via a mobile app, increasing user engagement. Energy efficiency optimization increases power generation to 6000 kWh. Predictive maintenance and nearby spare parts inventory reduce average annual costs to 200 yuan. Digital twin simulation of shadow effects provides a simple solution (like pruning branches), and proactive safety protection adapts to rural power grid fluctuations, fully meeting the low-cost, easy-to-operate maintenance needs of residential projects and promoting the widespread adoption of distributed PV in rural areas.

[0070] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. An intelligent operation and maintenance management system for distributed photovoltaic projects, characterized in that, include: The global sensing module adopts a star + mesh hybrid topology architecture and deploys sensing terminals including photovoltaic modules, combiner boxes, inverters, and environmental monitoring. It uses 5G slicing technology to prioritize data transmission, and the edge nodes have built-in abnormal pattern recognition algorithms to filter invalid data in real time. Digital Twin Operation and Maintenance Hub: Constructs a holographic twin model containing geographic information, equipment parameters, and meteorological data, supporting synchronization between physical equipment and the virtual model; integrates component temperature fields and inverter heat dissipation flow fields for dynamic simulation; integrates historical fault cases for deep learning to generate virtual fault injection models to provide a training environment for operation and maintenance personnel; The adaptive evolutionary fault diagnosis module adopts a three-layer architecture of feature extraction, pattern matching, and root cause reasoning. The bottom layer generates a fault fingerprint database through attention feature fusion; the middle layer introduces a transfer learning algorithm to adapt to fault identification of inverters and components across brands; and the root layer builds a knowledge graph to automatically trace the fault propagation path. Dynamic resource scheduling platform: Integrates operation and maintenance resources to create digital profiles, including skill tags, equipment compatibility, and historical performance; uses reinforcement learning algorithms to optimize dispatching, and generates scheduling plans by comprehensively considering the scope of fault impact, repair urgency, and resource accessibility; Supports collaborative inspection by drone swarms, enabling full-area scanning of the power plant through 3D path planning; Energy efficiency adaptive optimization hub: Real-time data collection from the generation and load sides to build a dynamic matching model; When the self-consumption rate is <65%, the energy storage system is linked to perform adaptive charging and discharging control, while adjusting the photovoltaic output curve; Equipped with a dust accumulation identification system, combined with precipitation probability prediction to generate the optimal working sequence of the cleaning robot; Dynamically optimize the MPPT tracking strategy to compress the tracking cycle to 1 second when clouds pass by. Active safety protection module: Constructs a closed-loop system of perception-decision-execution, including arc monitoring device, dual verification of microwave radar and video analysis, and insulation monitoring instrument; adopts digital filtering technology to eliminate power grid harmonic interference, and identifies potential DC arcing and grounding fault hazards; when a safety threat is detected, it is handled in a graded manner according to the scope of impact, and simultaneously activates a three-level alarm mechanism of audible and visual alarm, operation and maintenance terminal push, and responsible person SMS notification.

2. The intelligent operation and maintenance management system for distributed photovoltaic projects according to claim 1, characterized in that, Also includes: Predictive health management module and intelligent supply chain collaboration module: Based on equipment degradation characteristics and combined with environmental stress, a remaining life prediction model is constructed to provide early warning of failure risks up to 12 months in advance; a material-level aging acceleration algorithm is developed to shorten the 25-year life assessment cycle of components to 6 months of laboratory testing and automatically generate differentiated maintenance plans; it connects with equipment supplier ERP systems to share spare parts inventory data in real time and uses replenishment algorithms to dynamically adjust safety stock; a blockchain system for recycling used components is established to record data of the entire process of dismantling and material reuse; and it supports cross-project spare parts allocation.

3. The intelligent operation and maintenance management system for distributed photovoltaic projects according to claim 1, characterized in that, Also includes: Value analysis module and interaction module: Integrate data on power generation revenue, carbon asset revenue, subsidy revenue, and operation and maintenance costs to build a full life cycle value assessment model; generate a sensitivity analysis matrix to provide quantitative basis for project expansion and equipment upgrades; develop a virtual power plant aggregation function to virtualize distributed photovoltaic projects as a single dispatch unit to participate in grid peak shaving and improve peak-valley arbitrage revenue; Supports real-time annotation of repair steps, display of holographic projection device status, and natural language interaction on the glasses; A four-level permission system is set up for visitors, operations and maintenance, management, and auditing. Operation logs are stored on blockchain and support one-click traceability of parameter modification records.

4. The intelligent operation and maintenance management system for distributed photovoltaic projects according to claim 1, characterized in that, The digital twin operation and maintenance center is equipped with a spatiotemporal dimension expansion function to access meteorological station data within a 5km radius, and combine it with a solar trajectory model to predict the distribution of solar radiation received by components in the next 72 hours; it also trains a shadow prediction model based on historical shadow data to provide an early warning of power generation loss caused by temporary shading 4 hours in advance. It supports virtual capacity expansion simulation, which simulates the impact of adding new components on the overall system efficiency and outputs the optimal expansion scheme.

5. The intelligent operation and maintenance management system for distributed photovoltaic projects according to claim 1, characterized in that, The adaptive evolution fault diagnosis module includes a fault self-healing subsystem, which automatically performs repair operations for common minor faults: adjusting the branch switch of the combiner box to achieve current balance, remotely restarting the inverter fan to restore normal speed, or adjusting the direction of the cooling fan to optimize heat dissipation efficiency; the self-healing process is recorded in its entirety; for faults that cannot be self-healed, a maintenance guide including 3D disassembly diagrams, tool lists, and safety precautions is automatically generated.

6. The intelligent operation and maintenance management system for distributed photovoltaic projects according to claim 1, characterized in that, The dynamic resource scheduling platform introduces a digital employee collaboration mechanism to generate a virtual operation and maintenance assistant, which helps engineers complete repetitive tasks such as data verification, report generation, and compliance checks; it also develops an autonomous repair function for drones, which can be equipped with a robotic arm to replace simple parts.

7. The intelligent operation and maintenance management system for distributed photovoltaic projects according to claim 2, characterized in that, The predictive health management module develops a material performance degradation monitoring subsystem, which monitors the crosslinking degree of EVA film and the water vapor transmission rate of the backsheet in real time through micro sensors implanted inside the component; combined with outdoor exposure test data, it establishes a correlation model between material performance and power generation efficiency, and provides early warning of material aging risk 6 months before power decay; it automatically pushes material maintenance suggestions to extend the actual service life of the component.

8. The intelligent operation and maintenance management system for distributed photovoltaic projects according to claim 3, characterized in that, The value analysis module includes a carbon footprint tracking subsystem that records carbon emission data throughout the entire process from component production, transportation, installation to operation and maintenance; it uses the life cycle assessment (LCA) method to calculate the carbon emissions per kilowatt-hour of photovoltaic power generation and generates a carbon reduction report that complies with the ISO 14064 standard; and it connects to the carbon trading platform API interface to support automatic carbon asset declaration and automatic carbon revenue transfer.

9. The intelligent operation and maintenance management system for distributed photovoltaic projects according to claim 1, characterized in that, The energy efficiency adaptive optimization center is equipped with grid-friendly control functions, responds to grid dispatch commands in real time, and adjusts active power output and reactive power compensation. A voltage sag tolerance enhancement algorithm is developed to maintain grid-connected operation when the grid voltage drops to 60% of the rated value.

10. The intelligent operation and maintenance management system for distributed photovoltaic projects according to claim 1, characterized in that, The active security protection module integrates the cybersecurity protection subsystem, adopts a zero-trust architecture, and encrypts all data transmissions using the national standard SM4 encryption; it has a built-in intrusion detection system (IDS); and it has developed a firmware security update mechanism to ensure that the system does not run interrupted during device program updates.

Images