Light, storage and diesel integrated power supply system and cellular monitoring and power scheduling method thereof

Abstract

The invention relates to the technical field of new energy power supply, and discloses a light, storage and diesel integrated power supply system and a cellular monitoring and power scheduling method thereof. The system comprises a health assessment unit, a self-adaptive degradation unit, a fault-tolerant control unit and a mode management unit by constructing an intelligent operation and maintenance framework of'multi-stage self-adaption-active fault tolerance '. Wherein the health assessment unit is configured with a sound sensor, a vibration sensor and a temperature sensor, acquires system operation state data, and calculates a health index based on a multi-dimensional feature fusion algorithm; the adaptive degradation unit dynamically adjusts the operation mode of the system according to the health index, and supports adaptive switching of full power, optimization, economy, emergency and protection modes; the fault-tolerant control unit ensures that the overall power supply is not influenced by a single-point fault through a modular redundancy design and an intelligent partition isolation technology; the mode management unit is responsible for mode switching strategy optimization and multi-path dynamic scheduling of energy flow.

Description

Technical Field

[0001] The present invention relates to the field of renewable energy power supply technology, and in particular to a highly reliable adaptive photovoltaic-storage-diesel fusion power supply system and a cellular monitoring and power scheduling method thereof, and specifically to health assessment, intelligent energy scheduling, fault diagnosis and maintenance management technology based on multi-dimensional feature fusion. Background Art

[0002] With the transformation of energy structure and the rapid development of smart grid technology, the photovoltaic-storage-diesel integrated power supply system plays an increasingly important role in microgrids. At present, the photovoltaic-storage-diesel system has achieved basic energy scheduling and coordinated control, and has been widely used in remote areas, industrial parks, hospitals, data centers and other scenarios.

[0003] In the prior art, for example, application publication number CN 118432051 A discloses a photovoltaic-storage-diesel fusion power supply system and its power supply method. The system coordinates and controls the operation strategies of the power grid, photovoltaics, diesel generation, and energy storage. When the power grid is out of power, photovoltaics cannot generate electricity, and the battery power is too low, the diesel generator supplies power to the load and charges the battery; when the diesel generator is shut down for refueling, the energy storage system supplies power, thereby achieving continuous and uninterrupted power supply. This technology makes full use of the high dynamic stability control characteristics of power electronics technology, improves the power quality of the power supply through a hybrid inverter, improves the power supply reliability to a certain extent, and avoids restrictions on the use of photovoltaics and energy storage.

[0004] However, the existing solar-storage-diesel hybrid power supply system still faces the following technical challenges in terms of intelligent and refined operation and maintenance: Health assessment mechanism: Although it has basic monitoring functions, it mainly relies on single indicator assessment and lacks intelligent fusion analysis of multi-dimensional features. The assessment accuracy of equipment performance degradation is insufficient, and the accuracy of predictive maintenance needs to be improved. Energy scheduling strategy: The existing scheduling algorithm can achieve basic energy balance and uninterrupted power supply, but there is still room for improvement in multi-objective optimization. The adaptive ability to environmental and load changes is limited, and the overall efficiency of the system can be further improved. Fault handling mechanism: basic fault detection and protection functions are available, but the ability to accurately locate and quickly isolate faults needs to be improved, the intelligence level of the system recovery strategy is insufficient, and the ability to coordinate and handle multiple faults needs to be strengthened; In terms of the level of intelligent operation and maintenance: remote monitoring and basic data analysis have been achieved, but the intelligence level of predictive maintenance is not enough, and there is a lack of deep learning and intelligent decision-making support. There is still a lot of room for optimization of operation and maintenance costs; System adaptability: The existing system has basic mode switching functions, but it lacks adaptability to complex working conditions, lacks a flexible performance tailoring mechanism, and the system reconstruction capability needs to be improved.

[0005] In order to further improve the performance and reliability of the photovoltaic-storage-diesel integrated power supply system, it is necessary to: build an intelligent health assessment system based on multi-dimensional feature fusion, develop an adaptive multi-objective energy optimization scheduling strategy, achieve accurate fault location and rapid isolation, and improve the intelligent operation and maintenance level of the system. Summary of the invention

[0006] The purpose of the present invention is to address the deficiencies of the prior art and to provide a highly reliable adaptive photovoltaic-storage-diesel fusion power supply system and its cellular monitoring and power scheduling method. By constructing a "multi-level adaptive-active fault-tolerant" intelligent operation and maintenance architecture and an innovative cellular monitoring structure, accurate perception of system status, preventive maintenance of faults and efficient coordination of multiple energy sources are achieved, effectively solving the technical problems of traditional power supply systems in monitoring accuracy, fault prevention and energy utilization efficiency.

[0007] In order to achieve one of the above purposes, the present invention adopts the following technical solution: A highly reliable adaptive photovoltaic-storage-diesel fusion power supply system, comprising: Health assessment unit, which includes sound sensors, vibration sensors and temperature sensors, and is used to: collect the operating status data of the photovoltaic system, energy storage system and diesel generator set in real time; calculate the health index of each component through a multi-dimensional feature fusion algorithm, wherein the multi-dimensional features include temperature features, vibration features and acoustic features; and perform graded assessment of the health status of different equipment, including excellent status, good status, general status, poor status and dangerous status; An adaptive degradation unit, configured to: switch between full power mode, optimization mode, economy mode, emergency mode and protection mode based on the health index; and automatically adjust the system operation mode when the health index decreases to achieve smooth degradation of system performance; Fault-tolerant control unit, used to: ensure that single-point failure does not affect the overall power supply through modular redundant design and intelligent partition isolation technology; quickly isolate the fault point when a fault is detected; and ensure the reliability of system communication through dual-bus architecture; A mode management unit is used to: optimize the mode switching strategy and determine the energy allocation scheme under each operating mode; Based on system operating costs, energy transmission losses and reliability risk, multi-path dynamic scheduling of energy flow is achieved, maximizing system operating efficiency while ensuring power supply reliability.

[0008] Preferably, the fault-tolerant control unit comprises: A main controller, a backup controller and a fast isolation circuit, wherein the main controller and the backup controller operate in parallel, and the response time of the fast isolation circuit is less than 5ms.

[0009] Preferably, the energy scheduling optimization target of the mode management unit is: min J = α·Cost + β·Loss + γ·Risk; Among them, Cost is the system operation cost, Loss is the energy transmission loss, Risk is the system reliability risk, α is the cost weight coefficient, β is the loss weight coefficient, γ is the risk weight coefficient, and α + β + γ = 1.

[0010] Beneficial effects: The highly reliable adaptive photovoltaic-storage-diesel fusion power supply system of the present invention constructs an intelligent operation and maintenance architecture of "multi-level adaptation-active fault tolerance", adopts a health assessment mechanism of multi-dimensional feature fusion and an adaptive degradation control strategy, thereby realizing accurate perception of system status and preventive maintenance of faults; through modular redundant design and rapid isolation technology, the fault tolerance and reliability of the system are improved; combined with a multi-objective optimized energy scheduling strategy, the energy utilization efficiency and power supply stability of the system are improved, effectively solving the technical problems of low reliability and high maintenance cost of traditional power supply systems.

[0011] In order to achieve one of the above purposes, the present invention adopts the following technical solution: A photovoltaic panel cellular monitoring system, comprising: A plurality of hexagonal frames, which are closely arranged to form a continuous honeycomb structure and laid on the photovoltaic panel surface, and are used to divide the photovoltaic panel into a plurality of independent monitoring units, and the inner side surface of each hexagonal frame is provided with a sensor; A connecting slot disposed on a plane of each of the hexagonal frames; A reflective wall fixed in the connecting slot; The sensor system disposed in the monitoring unit of each hexagonal frame includes a temperature sensor, a light intensity sensor, and a voltage and current sensor; And a controller, used to: judge the abnormality type of each monitoring unit based on the temperature data, light intensity data and voltage and current data collected by the sensor system, the abnormality type including surface contamination, material aging and reflective wall abnormality; when an abnormality is detected, control the reflective wall of the corresponding monitoring unit to adjust the angle and height to adjust the light distribution of the monitoring unit; by comparing the data of the target monitoring unit with the adjacent monitoring units, judge whether the abnormality is a local problem, a regional problem or a systemic problem.

[0012] Preferably, the reflective wall has three specifications: standard, enhanced and adjustable. The reflective surface of the reflective wall faces the center of the hexagonal frame. The adjustable reflective wall adopts an intelligent driving structure, including: The bottom driving mechanism adopts a bimetallic temperature difference driving device; Angle adjustment mechanism, using micro stepping motor; And a position sensor is used to monitor the working status of the reflective wall in real time.

[0013] Beneficial effects: The photovoltaic panel honeycomb monitoring system of the present invention realizes all-round monitoring of photovoltaic panels by constructing a regular hexagonal honeycomb monitoring network and a multi-dimensional sensor array; with the help of the intelligent reflective wall adjustment mechanism, it can actively adjust the local light distribution; through multi-sensor collaborative monitoring and data fusion analysis, it improves the accuracy and timeliness of fault diagnosis, and effectively solves the technical problems of incomplete coverage and delayed response of traditional monitoring systems.

[0014] In order to achieve one of the above purposes, the present invention adopts the following technical solution: A direct power dispatching system based on a cellular monitoring structure, comprising: The micro group consists of 7 honeycomb units to form a basic monitoring unit, which is used to: realize power characteristic analysis of a single PV module; perform local performance optimization and fault diagnosis; and provide basic power regulation capabilities; The meso-group is a regional management unit composed of three micro-groups, which is used to: coordinate the power balance of each micro-group in the region; manage regional performance optimization and load distribution; and provide medium-scale power dispatch capabilities; The macro group, which is a system-level unit composed of three meso groups, is used to: achieve overall power management of large-scale photovoltaic arrays; perform system-level performance optimization and load balancing; and provide large-scale power scheduling capabilities; The power dispatch controller is used to: monitor the power output status of each group in real time; dynamically adjust the power allocation strategy according to changes in environmental factors such as light and temperature; realize direct and efficient power supply from photovoltaic power generation units to power loads; and automatically perform power redistribution when system parameters are abnormal.

[0015] Beneficial effects: The direct power dispatching system of the present invention realizes hierarchical management from single photovoltaic modules to large-scale photovoltaic arrays by establishing a three-level power management system of micro, meso and macro. It adopts direct power supply mode and dynamic power dispatching strategy to reduce energy conversion links. Through multi-level collaborative control mechanism, it enhances the system's anti-disturbance capability and effectively solves the technical problems of poor dispatching flexibility and large transmission loss of traditional power supply systems.

[0016] Compared with the prior art, the present invention has the following significant beneficial effects: The photovoltaic-storage-diesel fusion power supply system and related methods provided by the present invention adopt a honeycomb monitoring structure and a multi-level adaptive control architecture, and have the following beneficial effects: 1) By establishing an intelligent operation and maintenance architecture of "multi-level self-adaptation and active fault tolerance" and combining a health assessment mechanism with multi-dimensional feature fusion, accurate perception of system status and preventive maintenance of faults are achieved, thus improving the overall reliability of the system; 2) Based on the innovative design of the honeycomb monitoring structure, combined with the intelligent reflective wall adjustment mechanism and multi-dimensional sensor array, all-round monitoring and active adjustment of photovoltaic panels are achieved, improving the monitoring accuracy and fault diagnosis capabilities of the system; 3) Adopting a hierarchical power management system and direct power supply mode, through dynamic power scheduling strategy and multi-level coordinated control, it reduces the energy conversion link and improves the energy utilization efficiency of the system; 4) By building a multi-level monitoring network and dynamic reconstruction mechanism, hierarchical management from micro to macro is achieved, enhancing the system's adaptability and anti-disturbance capabilities; 5) The energy scheduling strategy based on multi-objective optimization achieves efficient coordination of multiple energy sources and precise matching of loads, improving the overall operating efficiency of the system.

[0017] The present invention effectively solves the technical problems existing in traditional photovoltaic-storage-diesel integrated power supply systems, such as low monitoring accuracy, weak fault prevention capability, and low energy utilization efficiency, and provides a new technical solution for the intelligent operation and maintenance of distributed energy systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the figures required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the figures described below are only some embodiments of the present application. For ordinary technicians in this field, other figures can be obtained based on these figures without paying creative work.

[0019] Figure 1 This is a partial system operation flow chart of Example 1; Figure 2 This is a partial system operation flow chart of Example 1; Figure 3 This is a system operation flow chart of Example 2; Figure 4 This is a system operation flow chart of Example 3; Figure 5 This is a system operation flow chart of Example 4; Figure 6 Schematic diagram of laying honeycomb structure for photovoltaic panels.

[0020] Reference numerals: health assessment unit 10, adaptive degradation unit 20, fault-tolerant control unit 30, mode management unit 40, sound sensor 11, vibration sensor 12, temperature sensor 13, temperature sensor array 13-1, vibration sensor array 12-1, sound sensor array 11-1. DETAILED DESCRIPTION

[0021] The following will be combined Figure 1-Figure 6 , the preferred embodiments of the present invention are described in detail. It should be noted that the following description is only a preferred embodiment of the present invention, rather than a limitation of the present invention. Those skilled in the art should understand that various modifications and variations can be made to the present invention without departing from the spirit and scope of the present invention. The scope of protection of the present invention shall be subject to the attached claims.

[0022] Embodiment 1: This embodiment provides a highly reliable adaptive photovoltaic-storage-diesel fusion power supply system and its control method, which realizes the autonomous and reliable operation of the system by building a "multi-level adaptive-active fault-tolerant" intelligent operation and maintenance architecture. Figure 1 As shown, the power supply system adopts a hierarchical distributed control structure, including four functional units: a health assessment unit 10, an adaptive degradation unit 20, a fault-tolerant control unit 30 and a mode management unit 40. Among them, the health assessment unit 10 monitors the status of each component of the system in real time and outputs a health index (Health Index, HI); the adaptive degradation unit 20 dynamically adjusts the operation mode of the system based on the health index, and supports adaptive switching of full power mode, optimization mode, economic mode, emergency mode and protection mode; the fault-tolerant control unit 30 ensures that a single point failure does not affect the overall power supply through modular redundant design and intelligent partition isolation technology; the mode management unit 40 is responsible for the optimization of the mode switching strategy and the multi-path dynamic scheduling of the energy flow, so as to achieve the optimal balance between system performance and reliability.

[0023] Among them, the health assessment unit 10, as the system's state perception center, builds a three-level health assessment system of equipment-component-module. The unit collects the operating status data of the photovoltaic system 01, the energy storage system 02 and the diesel generator set 03 in real time by configuring the sound sensor 11, the vibration sensor 12 and the temperature sensor 13, and calculates the health index (HI) of each level of equipment through a multi-dimensional feature fusion algorithm. Based on the health index, the adaptive degradation unit 20 adopts modular redundant design and intelligent partition isolation technology. When the HI of the equipment is detected to be reduced, it automatically adjusts the operating mode to achieve smooth degradation of system performance. The fault-tolerant control unit 30 ensures the continuous and stable operation of the power supply system in the event of a single point failure through a "1+1" redundant configuration and a fast isolation circuit (response time <5ms). The mode management unit 40 optimizes the energy flow distribution under each operating mode in real time based on the distributed energy management strategy, maximizing the system operation efficiency while ensuring power supply reliability.

[0024] Reference Figure 2 In some specific embodiments, the monitoring configuration of the health assessment unit 10 for the photovoltaic system 01, the energy storage system 02 and the diesel generator set 03 is as follows: The monitoring configuration of photovoltaic system 01 adopts a multi-module adaptive architecture, including temperature sensor array 13-1, vibration sensor array 12-1 and sound sensor array 11-1. Among them, temperature sensor array 13-1 adopts a "three-level adaptive monitoring architecture": intelligent infrared thermal imagers (scanning range 120°, resolution 640×480 pixels) are arranged at the four corners and center of the photovoltaic array, supporting adaptive sampling frequency adjustment based on health index, real-time monitoring of the surface temperature distribution of photovoltaic panel components, and timely detection of hot spot effects; "1+1" redundant design is adopted on the backplane of photovoltaic panel components, and dual-working mode PT100 temperature sensors are arranged every 1 meter to improve monitoring reliability through cross-validation between sensors; a 0.3mm thick intelligent temperature display coating is applied on the surface of key equipment such as junction boxes and inverters, and backup temperature sensors are configured to achieve fault-tolerant monitoring of key nodes. The vibration sensor array 12-1 adopts a master-slave monitoring architecture: two sets of three-axis acceleration sensors (main sensor and backup sensor) are set at the base of the photovoltaic support. By comparing the data difference between the main and backup sensors in real time, cross-validation is triggered when the difference exceeds the preset threshold (such as 10%) to improve data reliability; the vibration sensor array 12-1 is arranged in a "2+1" redundant manner at the top and bottom of the inverter cabinet. The monitoring function can be guaranteed by the redundancy design if any sensor fails. The sound sensor array 11-1 adopts an intelligent redundant arrangement, and three high-precision microphones are staggered at 1 meter around the inverter. The adaptive beamforming technology is used to improve the accuracy of sound feature acquisition.

[0025] The health assessment unit 10 uses a multi-dimensional feature fusion algorithm to calculate the health index (HI) of each component of the system, and its specific implementation method is as follows: 1. Calculation method of health index (HI): The health index ranges from [0,1], where 1 indicates that the device is in the best working state, 0 indicates that the device is completely failed, and intermediate values ​​indicate that the device is in different degrees of performance degradation. The index is calculated by multi-dimensional feature fusion and is used to characterize the overall operating status of the device. The specific calculation process includes two main steps: basic feature extraction and feature fusion calculation: Step 1. Basic feature extraction part, including: (1) Extraction and calculation of temperature characteristics (FT): Arrange temperature sensors at key points of the photovoltaic array to collect real-time temperature data; Calculate the temperature mean Tm: calculate the arithmetic average of the temperature values ​​of N measuring points, that is, Tm = ∑Ti / N, where Ti is the temperature value of the i-th measuring point; Calculate the temperature variance σt: The temperature variance (σt) is calculated by the standard variance calculation formula σt = √[∑(Ti-Tm)² / N] to characterize the degree of discreteness of the temperature distribution, where: Ti represents the specific temperature value of the i-th measuring point (in °C), Tm is the arithmetic mean temperature of all measuring points, calculated by Tm = ∑Ti / N; (Ti-Tm)² in the formula represents the square of the deviation of each measuring point from the average temperature. This design can eliminate the influence of the mutual offset of positive and negative deviations and increase the weight of outliers; ∑(Ti-Tm)² / N represents the average square deviation, which is standardized by dividing by the total number of measuring points N, so that the results of different numbers of measuring points are comparable; finally, the square root √ is taken to return the unit to the original temperature unit and reflect the actual average deviation. For example, for the five measurement points [42, 44, 43, 45, 41]℃ on the photovoltaic panel, first calculate the average temperature of 43℃, then calculate the sum of squared deviations (42-43)² + (44-43)² + (43-43)² + (45-43)² + (41-43)² = 10, divide it by the number of measurement points 5 to get 2, and finally take the square root to get the standard deviation of about 1.4℃, which indicates that the temperature distribution is relatively uniform. This indicator is of great significance in photovoltaic systems and can be used to promptly detect hot spot effects, evaluate heat dissipation performance, and warn of potential component failures; Calculate the temperature gradient Gt: Calculate the temperature change rate between adjacent measuring points using the formula Gt = ΔT / ΔL, where ΔT represents the temperature difference between adjacent measuring points and ΔL represents the actual distance between the two points.

[0026] In order to make the temperature characteristics under different working conditions comparable, the system adopts standardization processing: the temperature mean standardization value Tm_norm is calculated by (Tm - Tm_min) / (Tm_max - Tm_min), where Tm_min and Tm_max are the minimum and maximum operating temperatures allowed by the system respectively; the temperature variance standardization value σt_norm is calculated by 1 - σt / σt_max, where σt_max is the maximum temperature variance allowed by the system; the temperature gradient standardization value Gt_norm is calculated by 1 - Gt / Gt_max, where Gt_max is the maximum temperature gradient allowed by the system. The final temperature feature FT is calculated by weighted fusion: FT = 0.4×Tm_norm + 0.3×σt_norm + 0.3×Gt_norm. This weight configuration reflects the different degrees of influence of mean, variance and gradient on the health status of the equipment. Taking a photovoltaic panel as an example, when the temperatures of the five measuring points are [42, 44, 43, 45, 41]°C, the calculated temperature mean is 43°C, the temperature variance is 1.4°C, and the temperature gradient is 2°C / m; these parameters are standardized relative to the system standard values ​​(standard operating temperature 40°C, maximum allowable variance and maximum gradient), and the results are Tm_norm = 0.92, σt_norm = 0.88, and Gt_norm = 0.85; the final calculated temperature characteristic FT = 0.4×0.92 + 0.3×0.88+ 0.3×0.85 = 0.887, which indicates that the temperature characteristics of the photovoltaic panel are within the healthy range.

[0027] (2) Extraction and calculation of vibration characteristics (FV): The acquisition of vibration characteristics (FV) is based on the data acquisition of the three-axis acceleration sensor, and the sampling frequency is set to 1kHz to ensure data accuracy. Its core parameters include: the vibration amplitude A is calculated by the root mean square method, that is, A = √(∑xi² / N), where xi represents the time domain sampling value. This calculation method can effectively reflect the average energy level of the vibration; the system obtains the spectrum distribution S through fast Fourier transform (FFT) analysis to evaluate the frequency characteristics of the vibration; at the same time, the maximum absolute value within the sampling period is recorded as the acceleration peak value P, which is used to monitor the instantaneous impact.

[0028] To make these parameters comparable, the system uses standardization: the standardized value A_norm of the vibration amplitude is calculated by 1 - A / A_max, where A_max is the maximum vibration amplitude allowed by the system; the standardized value S_norm of the spectrum feature indicates the degree of match between the current spectrum and the spectrum under the standard operating state; the standardized value P_norm of the acceleration peak is calculated by 1- P / P_max, where P_max is the maximum acceleration value allowed by the system. The final calculation of the vibration feature adopts a weighted fusion method: FV = 0.35×A_norm + 0.35×S_norm + 0.3×P_norm. This weight configuration reflects the relative importance of amplitude, spectrum and peak value to the equipment status assessment.

[0029] Taking an actual case as an example: the vibration data collected at a certain time showed that the vibration amplitude was 0.5g (relative to the standard range of 0-1g), the frequency spectrum distribution matched the standard mode by 95%, and the acceleration peak was 1.2g (the maximum value allowed by the system is 2g); after standardization, A_norm = 0.90, S_norm = 0.95, and P_norm = 0.90 were obtained; the vibration characteristic value FV = 0.35×0.90 + 0.35×0.95 + 0.3×0.90 = 0.92 was finally calculated, which shows that the vibration characteristics of the equipment are in a good range.

[0030] (3) Extraction and calculation of acoustic features (FS): Acoustic features (FS) are collected through a high-precision microphone array, and the sampling frequency is set to 44.1kHz to ensure the integrity of the sound signal. Its core parameters include: the sound pressure level L is calculated by the formula L = 20×log10(P / P0), where P represents the measured sound pressure and P0 is the reference sound pressure (the standard value is 20μPa). This logarithmic calculation method is consistent with the perception characteristics of the human ear; the system also extracts the frequency feature F, including the identification of the main frequency and the analysis of the full spectrum, which is used to evaluate the frequency composition of the sound; in addition, the sound wave morphology W is analyzed, and the quality of the sound signal is evaluated by calculating the total harmonic distortion.

[0031] In order to achieve data comparability under different working conditions, the system adopts standardization processing: the standardized value L_norm of the sound pressure level is calculated by 1 - |L - L_std| / L_max, where L_std is the sound pressure level under the standard operating state and L_max is the maximum allowable deviation; the standardized value F_norm of the frequency characteristic indicates the degree of matching between the current frequency characteristic and the standard operating state; the standardized value W_norm of the sound wave morphology is calculated by 1 - distortion, which directly reflects the purity of the sound wave. The final calculation of the acoustic characteristics adopts a weighted fusion method: FS = 0.4×L_norm + 0.3×F_norm + 0.3×W_norm. This weight configuration reflects the relative importance of sound pressure level, frequency characteristics and waveform quality to equipment status assessment.

[0032] Take an actual case as an example: the sound pressure level of a certain device during operation is 65dB (relative to the standard operating value of 60dB), the frequency characteristics match the standard mode by 85%, and the total harmonic distortion is 2% (the system requires less than 3%); after standardization, L_norm = 0.90 (indicating that the sound pressure level is close to the standard value), F_norm = 0.85 (indicating that the frequency characteristics are basically in line with expectations), and W_norm = 0.90 (indicating that the waveform quality is good); the acoustic characteristic value FS = 0.4×0.90+ 0.3×0.85 + 0.3×0.90 = 0.88 is finally calculated, which indicates that the acoustic characteristics of the device are within the normal operating range.

[0033] Step 2. Feature fusion calculation part, using the improved Dempster-Shafer evidence theory for feature fusion: (1) Basic calculation formula of health index: HI = w1·FT + w2·FV + w3·FS; Wherein: w1 is the weight coefficient of temperature feature (FT); w2 is the weight coefficient of vibration feature (FV); w3 is the weight coefficient of acoustic feature (FS).

[0034] (2) Weight coefficient configuration rules: The weight coefficients satisfy the normalization constraint: w1 + w2 + w3 = 1; the value range of each weight coefficient is [0,1]; to maintain feature diversity, the minimum value of a single weight is not less than 0.2; to avoid a single feature dominating, the maximum value of a single weight does not exceed 0.5.

[0035] (3) Dynamic adjustment strategy of weight coefficient: a) Adjustment of sensor under abnormal conditions: When a certain type of sensor data is detected to be abnormal, the weight of the corresponding feature is automatically reduced by 20%-50%; The weights of other features are increased proportionally to ensure that the sum of the weights remains 1; Example: When the temperature sensor is abnormal, w1 decreases from 0.4 to 0.2, and w2 and w3 increase from 0.3 to 0.4 respectively; b) Adjustment based on historical data: Regularly analyze the past 1000 hours of operating data; adjust the weight according to the prediction accuracy of each feature; for every 10% increase in accuracy, the corresponding weight can be increased by 0.05.

[0036] Example: When the vibration feature accuracy reaches 95%, w2 can be increased to 0.4, while w1 and w3 can be reduced in the same proportion; c) Environmental adaptability adjustment: Dynamically adjust weight configuration according to working conditions; increase temperature feature weight in high temperature environment; increase vibration feature weight in high vibration environment; weight adjustment step size does not exceed 0.1 to ensure system stability; Evaluation criteria for health index: According to the numerical range of the health index, the device status is divided into the following levels: [0.9, 1.0]: The device is in excellent condition and all indicators are within the optimal range; [0.75, 0.9): The device is in good condition, and the indicator deviates slightly from the optimal value; [0.6, 0.75): The equipment status is average and the changing trend needs to be closely monitored; [0.3, 0.6): The equipment is in poor condition and maintenance is recommended; [0, 0.3): The device status is dangerous and needs to be handled immediately; 0: The equipment has completely failed and needs to be shut down for maintenance.

[0037] Through the above method, the system can evaluate the health status of the equipment in real time and accurately, providing a reliable basis for preventive maintenance and operation optimization. This method has the characteristics of strong adaptability and high reliability, and can adapt to different operating environments and working conditions.

[0038] 2. Mode switching strategy of the adaptive degradation unit 20: (1) Operation mode definition: Full power mode: The system runs at maximum efficiency and all devices work normally; Optimization mode: The performance of some equipment has slightly degraded, and high-efficiency operation is maintained through optimization control; Economic mode: Equipment performance is significantly reduced, and system reliability is prioritized; Emergency mode: When a core device fails, backup resources are activated to ensure basic power supply; Protection mode: The system is at risk of serious failure and only maintains minimal operation.

[0039] (2) Mode switching criteria: HI ≥ 0.9 is full power mode; 0.75 ≤ HI < 0.9 is optimization mode; 0.6 ≤ HI < 0.75 is economy mode; 0.3 ≤ HI < 0.6 is emergency mode; HI < 0.3 is protection mode.

[0040] 3. Specific implementation of the fault-tolerant control unit 30: (1) "1+1" redundant configuration: the main controller and the backup controller run in parallel; a dual-bus architecture is used to ensure communication redundancy; key sensors adopt dual working modes and support cross-validation.

[0041] (2) Fast isolation circuit design: solid-state switch is used to achieve fast isolation of fault points, with a response time of <5ms; electrical isolation design based on photoelectric coupling, with an isolation voltage of >2500V; integrated overcurrent and overvoltage protection functions, and programmable protection thresholds.

[0042] 4. Distributed energy management strategy: (1) Energy scheduling optimization objective: min J = α·Cost + β·Loss + γ·Risk; Where: Cost is the system operation cost, Cost = Σ(PiEi + Si), Pi is the operating power of the i-th device, Ei is the unit energy cost of the i-th device, Si is the start-stop cost of the i-th device; Loss is the energy transmission loss, Loss = Σ(Ri·Ii²), Ri is the equivalent resistance of the i-th transmission line, Ii is the transmission current of the i-th line; Risk is the system reliability risk, Risk = 1 - Π(1 - ri·hi), Π represents the multiplication of all items, i is the inherent failure rate of the i-th device, hi is the health index impact factor of the i-th device; α: cost weight coefficient, reflecting the importance of economic requirements; β: loss weight coefficient, reflecting the importance of energy efficiency requirements; γ: risk weight coefficient, reflecting the importance of reliability requirements. α + β + γ = 1. For example: the system status at a certain moment is as follows: Cost = 100 yuan / hour (equipment operation cost); Loss = 5kW (transmission loss); Risk = 0.2 (system risk); α = 0.5, β = 0.3, γ = 0.2; Then the objective function value J = 0.5×100 + 0.3×5 + 0.2×0.2 = 51.54; the system will try to minimize this value by adjusting the equipment operating parameters.

[0043] (2) Constraints: Power balance constraint: Ppv + Pbat + Pdie = Pload + Ploss, where: Ppv is the output power of the photovoltaic system; Pbat is the charging and discharging power of the energy storage system; Pdie is the output power of the diesel generator set; Pload is the load demand power; Ploss is the system loss power; Equipment operation constraints: Pmin,i ≤ Pi ≤ Pmax,i, where Pmin,i and Pmax,i are the minimum and maximum operating powers of the i-th equipment, respectively; Energy storage capacity constraint: SOCmin ≤ SOC(t) ≤ SOCmax, SOC(t) is the state of charge of the energy storage system at time t, SOCmin and SOCmax are the minimum and maximum states of charge allowed, respectively; Reliability index constraint: R(t) ≥ Rmin, R(t) is the reliability index of the system at time t; Rmin is the minimum reliability level required by the system.

[0044] (3) Solution method: The system uses the improved particle swarm optimization algorithm (IPSO) to solve the energy scheduling optimization problem. The main steps are as follows: Step 1. Initialization phase: determine the solution dimension: n = number of devices × number of scheduling periods; set the initial population: the size is 50; set the initial position of the particle: to represent the power allocation plan of each device; set the initial speed: randomly generated within the allowed range; Step 2. Iterative optimization phase: Calculate fitness: evaluate the quality of the solution based on the objective function value; Update the optimal solution: track the individual historical optimal and the group global optimal; Update particle state: adjust the particle speed and position based on the inertia weight and acceleration factor; Step 3. Termination conditions: reaching the maximum number of iterations of 200; no significant improvement after 50 consecutive iterations (improvement < 10⁻ 6 ); all constraints are met and the objective function meets the requirements; Through the above optimization process, the system can obtain the optimal energy scheduling solution that meets the constraints. This method has the characteristics of fast convergence speed and high computational efficiency, and is suitable for real-time energy scheduling optimization.

[0045] Example 2 Based on Example 1, the health assessment unit 10 proposes a "photovoltaic panel honeycomb monitoring structure" for the photovoltaic system 01. The structure forms a planar honeycomb structure by laying a regular hexagonal frame array on the surface of the photovoltaic panel, and based on this, a multi-dimensional feature fusion adaptive assessment method is constructed. This design that combines structural innovation with assessment methods breaks through the limitations of traditional single indicator assessment and constructs a multi-mode assessment system that supports performance tailoring.

[0046] See also Figure 6 On the photovoltaic panel surface, the system uses 300 regular hexagonal frames (inscribed circle diameter 0.5 meters) to be closely arranged to form a continuous honeycomb structure. Each frame is 10-15mm wide and 15-20mm high. The inner side of each frame is provided with a mounting groove for installing the sensor. The mounting groove can adopt a dovetail groove structure to ensure that the sensor is firmly installed. The upper plane of each frame is provided with a connecting card slot. The width of the connecting card slot is 12mm and the depth is 8mm. A double-layer step structure is adopted. The upper layer is used for the preliminary positioning of the reflective wall, and the lower layer is provided with an elastic locking mechanism to ensure that the reflective wall is firmly installed. The reflective surface of the reflective wall faces the central area of ​​the hexagonal frame, which is used to adjust the light distribution in this area.

[0047] The connection slot is provided with a reflective wall which is fixed after insertion. The reflective wall has various specifications, including: Standard type: height 50mm, inclination angle 15°, made of high reflectivity aluminum alloy (reflectivity>92%), chrome-plated surface; Enhanced type: height 80mm, inclination angle 30°, composite material structure, front side is directional reflection surface, back side is scattering surface; Adjustable type: height adjustable from 60 to 100 mm, inclination adjustable from 0 to 45°, with telescopic structure design.

[0048] In a preferred embodiment, the reflective wall adopts an intelligent driving structure, including: Bottom drive mechanism: It uses a bimetallic temperature difference drive device. When it detects that the local temperature difference exceeds the preset threshold (the default is 8°C), it automatically triggers the extension of the reflective wall. Angle adjustment mechanism: Use a micro stepping motor with a reducer to achieve 0.5° precise angle adjustment; Position sensor: used to monitor the working status of the reflective wall in real time, including the extension height and tilt angle.

[0049] The system sets up three types of sensors in each honeycomb unit: Temperature sensor: PT100 temperature sensor, measuring accuracy ±0.1℃, measuring range -40℃ to 120℃; Light intensity sensor: photodiode, response time <1ms, sensitivity 10mV / lux; Voltage and current sensor: used to monitor local power generation performance, sampling rate 1kHz.

[0050] When it is detected that the temperature and converted power in a honeycomb unit have dropped significantly (temperature rises by more than 10°C or power generation efficiency drops by more than 15%), the system will automatically compare the data of the six surrounding adjacent units to accurately determine the cause of the abnormality. The specific judgment logic is: If the data of the surrounding units are normal, but the target unit is abnormal, it is determined to be a local problem; If the adjacent area shows a decrease in gradient performance, it is judged to be a regional problem; If the performance of the entire area decreases uniformly, it is considered a systemic problem.

[0051] Based on the above judgment results, the system can accurately identify the following fault types: Surface pollution: characterized by local temperature increase and decreased power generation efficiency; Material aging: characterized by uniform performance decay and slower response characteristics; Abnormal reflective wall: characterized by uneven local light intensity distribution and abnormal temperature distribution.

[0052] Example 3 Based on Example 2, this example proposes a honeycomb monitoring strategy of "dynamic reconstruction of partitions". This strategy makes full use of the geometric characteristics of the honeycomb structure, combines multiple adjacent hexagonal units into monitoring areas of different scales, and realizes multi-level adaptive monitoring from micro to macro.

[0053] See also Figure 4 The system defines three basic monitoring unit combinations: micro-monitoring unit (7-unit combination), meso-monitoring unit (19-unit combination) and macro-monitoring unit (37-unit combination). The micro-monitoring unit consists of a central honeycomb and six adjacent honeycombs around it, which is used to accurately locate local anomalies; the meso-monitoring unit consists of three micro-monitoring units arranged in a triangle, which is used to analyze regional performance changes; the macro-monitoring unit consists of three meso-monitoring units, which is used to evaluate the overall status of the system. The monitoring data of each unit is transmitted to the central controller in real time through the optical fiber bus, with a sampling frequency of 100Hz.

[0054] The dynamic reconstruction process of the system is controlled by an intelligent judgment algorithm. When the temperature gradient exceeds 5°C / m, or the power generation efficiency fluctuates by more than 10%, or an abnormal light intensity distribution is detected (relative standard deviation>15%), the reconstruction mechanism is triggered. The reconstruction process is divided into three levels: local reconstruction (7-unit range), regional reconstruction (19-unit range) and global reconstruction (37-unit range). The reconstruction time is less than 100ms to ensure the continuity of monitoring.

[0055] This embodiment adopts a "layered-progressive" data processing model. At the micro level, the temperature mean (Tavg), standard deviation (σT) and light intensity distribution correlation coefficient (R) of 7 units are calculated; at the meso level, a performance heat map of 19 units is constructed, and the performance gradient (∇P) and directional index (D) are calculated; at the macro level, a state assessment matrix of 37 units is generated, including the health index (HI) and risk warning value (RW). The specific calculation formula is as follows: Micro-level indicators: 1. Temperature uniformity: UI = 1 - σT / Tavg; where: UI is the temperature uniformity index, ranging from [0,1]; σT is the temperature standard deviation, in °C; Tavg is the average temperature, in °C.

[0056] 2. Light intensity correlation: R = Σ(Ii·Ij) / (σi·σj); where: R is the light intensity correlation coefficient, with a value range of [-1,1]; Ii, Ij are the light intensity values ​​of the i-th and j-th measuring points, in lux; σi, σj are the corresponding standard deviations, in lux.

[0057] 3. Local anomaly: LA = 1 - min(UI, R); where LA is the local anomaly, ranging from [0,1]; min(UI, R) means taking the smaller value of UI and R.

[0058] Meso-level indicators: 1. Performance gradient: ∇P = √[(∂P / ∂x)² + (∂P / ∂y)²]; where: ∇P is the performance gradient, unit kW / m²; ∂P / ∂x is the power change rate in the x direction; ∂P / ∂y is the power change rate in the y direction.

[0059] 2. Directional index: D = arctan(∂P / ∂y) / (∂P / ∂x); where: D is the directional index, unit rad; represents the main direction of performance change.

[0060] 3. Regional uniformity: RA = exp(-k·∇P); where: RA is the regional uniformity index, with a value range of [0,1]; k is the attenuation coefficient, with a default value of 0.5 m² / kW; exp represents the natural exponential function.

[0061] Macro-level indicators: 1. Health index: HI = w1·UI + w2·RA + w3·F; where: HI is the health index, with a value range of [0,1]; w1, w2, w3 are weight coefficients, and w1 + w2 + w3 = 1; F is the power factor, which represents the ratio of actual power to rated power. Initial default values: w1=0.4, w2=0.3, w3=0.3.

[0062] 2. Risk warning value: RW = α·LA + β·∇P + γ·(1-HI); where: RW is the risk warning value, with a value range of [0,1]; α, β, γ are risk factors, and α + β + γ = 1; the default value of α is 0.4, the default value of β is 0.3, and the default value of γ is 0.3.

[0063] The system adopts a dynamic risk factor adjustment strategy and makes corresponding adjustments when the following situations occur: when the performance degradation rate is >1% / min: α is increased to 0.6; when a significant gradient is detected (∇P>0.5kW / m²): β is increased to 0.5; when the operating time is >5000h: γ is increased to 0.4, and for every additional 5000h, γ is increased by 0.1, with a maximum of no more than 0.6.

[0064] Based on the above calculation results, the system automatically adjusts the monitoring strategy: when LA>0.2, local reconstruction is initiated; when RA<0.8, regional reconstruction is extended; when HI<0.6, global reconstruction is triggered. The reconstructed monitoring units are interconnected through the optical fiber ring network to achieve real-time data sharing and cross-validation.

[0065] The monitoring effect evaluation of this embodiment uses three key indicators: Response time: less than 200ms from the occurrence of an exception to the completion of reconstruction; Positioning accuracy: The positioning error of the abnormal source is less than the size of a basic unit; Early warning accuracy: no less than 95% (verified based on 1,000 hours of operating data).

[0066] Through this partition dynamic reconstruction strategy, the system realizes all-round and multi-level monitoring of the photovoltaic array, significantly improves the accuracy and timeliness of fault diagnosis, and provides reliable data support for preventive maintenance. Experimental verification shows that this strategy can effectively reduce fault diagnosis time and improve system operation reliability.

[0067] For the "honeycomb monitoring structure" in Example 2 and Example 3, the frame and the photovoltaic panel surface are connected by flexible point contact. 3-4 evenly distributed contact points are set at the bottom of each frame, and weather-resistant UV curing glue is used for bonding and fixing, and the thickness of the glue layer is controlled at 0.8-1.2mm. The main body of the frame is made of PMMA material with a light transmittance greater than 92%, a thickness of 1.5-2.0mm, and the surface is anti-glare treated.

[0068] To prevent rainwater from gathering, a gap of 2-3mm is maintained between the frame and the photovoltaic panel surface, and the edge is designed with a 45° inclination to effectively prevent water accumulation. The overall design ensures that the frame's shading rate is controlled below 3%, minimizing the impact on photovoltaic power generation efficiency.

[0069] Example 4 See also Figure 5 This embodiment proposes a direct power dispatch system based on a cellular monitoring structure. The system realizes a direct power supply mode from the power generation unit to the power load through intelligent grouping and dynamic power allocation of the photovoltaic array. This power supply method avoids the process of energy storage and reconversion in traditional photovoltaic systems, and can significantly improve system efficiency in specific application scenarios. The core of the system is to flexibly combine cellular monitoring units, build a multi-level power supply system, and achieve accurate matching of loads through intelligent scheduling algorithms.

[0070] The dispatching system builds a complete power dispatching solution based on Examples 2 and 3. The "micro-meso-macro" three-level grouping architecture is adopted in the organization of the power supply system, where the micro group consists of 7 honeycomb units (1 central unit plus 6 peripheral units) with a nominal power of about 2kW; the meso group consists of 3 micro groups with a nominal power of about 5.3kW; the macro group consists of 3 meso groups with a nominal power of about 10.4kW. The nominal power of each honeycomb unit is 280W±5%, and the stability of the group output is ensured through precise power monitoring and adjustment.

[0071] The core implementation of the dispatching system includes two key links: dynamic grouping control and direct power dispatching. In terms of dynamic grouping control, the system sets three types of trigger conditions: when the light intensity changes by more than 50W / m² / min, the temperature gradient exceeds 3℃ / m, or the power output fluctuation exceeds 5% / min, the grouping reorganization mechanism will be automatically started. The grouping reorganization process adopts the principle of "nearby priority and balanced distribution" to ensure that the reorganized power output meets the load demand.

[0072] The direct power dispatch system adopts a multi-level real-time matching algorithm, combined with load characteristic identification, group power optimization and dynamic adjustment strategy, which is specifically implemented as follows: 1. Load power characteristics modeling Basic model: P_load(t) = P_base + ΔP(t); Where: P_load is the real-time power demand of the load; P_base is the basic power (static power demand of the load); ΔP(t) is the dynamic power fluctuation; Power prediction: construct a feature sequence {P(tn),...,P(t-1)} based on historical data, and use the sliding window method to predict P_load(t+1); Fluctuation characteristic analysis: Calculate the standard deviation σ = √[Σ(P(t) - P_mean)² / n] to evaluate load stability.

[0073] 2. Group power optimization configuration Total output calculation: P_group(i) = Σ(p_j × η_j); Where: p_j is the real-time output power of the jth photovoltaic unit; η_j is the corresponding power conversion efficiency; Efficiency correction: η_j = η_base × k_temp × k_irr; Among them: η_base is the base conversion efficiency; k_temp is the temperature correction coefficient; k_irr is the light intensity correction coefficient; Group power balancing: Ensure that |p_j - p_mean| ≤ ​​10%, where p_mean is the average power within the group.

[0074] 3. Real-time matching evaluation Matching degree calculation: M = 1 - |P_load - P_group| / P_load; Evaluation cycle: basic evaluation cycle T = 100ms, when ΔP(t) / Δt > 10W / s, trigger fast evaluation (T = 20ms); Matching threshold: Set M_threshold = 0.95, and when M ≥ M_threshold, it is determined to be the best match; 4. Multi-level dynamic adjustment strategy When a power mismatch occurs (M < 0.95), the system adjusts according to the following process: (1) Calculate the real-time power difference: ΔP = P_load - P_group; The system adopts a three-level adjustment strategy according to the size of the power difference |ΔP|: when |ΔP| does not exceed 100W, the first-level adjustment is adopted to maintain the existing group structure and evenly distribute the power difference to each unit in the group, and the adjustment period is 50ms; when |ΔP| is between 100W and 500W, the second-level adjustment is started, the adjacent group coordination mechanism is activated, and the power balance is achieved by calculating the minimum power adjustment amount (that is, |ΔP| divided by the total number of units participating in the coordination), and the adjustment period is extended to 100ms; when |ΔP| exceeds 500W, the third-level adjustment is triggered, and the system will rebuild the group based on the minimum power redundancy principle, and the adjustment period of this process is 200ms. After each adjustment, the system will calculate the new matching degree M_new for verification. If the matching degree decreases, it will fall back to the state before the adjustment; if the system fails to achieve the expected effect after three consecutive adjustments, the abnormal state will be reported to the superior control system.

[0075] The dispatching system uses mature power electronics technology to achieve power regulation. For example, a standard MPPT controller and DC-DC converter combination is used to achieve power regulation of photovoltaic units, where the MPPT controller uses the disturbance observation method and the DC-DC converter uses a bidirectional Buck-Boost topology; power sampling uses a combination of Hall effect current sensors and high-precision voltage divider networks to ensure sampling accuracy; the control system uses the PWM modulation technology commonly used in the industry, with a switching frequency of 20kHz to ensure fast response and stable control.

[0076] After trial operation in an industrial park, the system has shown good results in application scenarios such as data centers, industrial air conditioners, and small industrial equipment. The overall efficiency of the system is 5.8% higher than that of traditional solutions, energy utilization is increased by 7.2%, and operation and maintenance costs are reduced by 12.5%. However, it should be noted that the system has high requirements for load characteristics and environmental conditions. It is recommended to carry out targeted optimization design according to specific scenario requirements in actual applications.

[0077] The basic principles of the present disclosure are described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, effects, etc. mentioned in the present disclosure are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present disclosure. In addition, the specific details disclosed above are only for the purpose of illustration and ease of understanding, and are not limitations. The above details do not limit the present disclosure to the necessity of adopting the above specific details to be implemented.

Claims

1. A highly reliable adaptive photovoltaic-storage-diesel fusion power supply system, characterized in that: include: Health assessment unit, which includes sound sensor, vibration sensor and temperature sensor, is used to collect the operation status data of photovoltaic system, energy storage system and diesel generator set in real time; Calculating the health index of each component by a multi-dimensional feature fusion algorithm, wherein the multi-dimensional features include temperature features, vibration features, and acoustic features; Conduct graded assessments on the health status of different equipment, including excellent status, good status, general status, poor status, and dangerous status; An adaptive degradation unit, configured to: switch between full power mode, optimization mode, economy mode, emergency mode and protection mode based on the health index; and automatically adjust the system operation mode when the health index decreases to achieve smooth degradation of system performance; Fault-tolerant control unit, used to: ensure that single-point failure does not affect the overall power supply through modular redundant design and intelligent partition isolation technology; when a fault is detected, quickly isolate the fault point; Ensure the reliability of system communication through dual bus architecture; A mode management unit is used to: optimize the mode switching strategy and determine the energy allocation scheme under each operating mode; Based on system operating costs, energy transmission losses and reliability risk, multi-path dynamic scheduling of energy flow is achieved, maximizing system operating efficiency while ensuring power supply reliability.

2. The highly reliable adaptive photovoltaic-storage-diesel fusion power supply system according to claim 1 is characterized in that: The fault-tolerant control unit comprises: A main controller, a backup controller and a fast isolation circuit, wherein the main controller and the backup controller operate in parallel, and the response time of the fast isolation circuit is less than 5ms.

3. The highly reliable adaptive photovoltaic-storage-diesel fusion power supply system according to claim 1 is characterized in that: The energy scheduling optimization target of the mode management unit is: min J = α·Cost + β·Loss + γ·Risk; Among them, Cost is the system operation cost, Loss is the energy transmission loss, Risk is the system reliability risk, α is the cost weight coefficient, β is the loss weight coefficient, γ is the risk weight coefficient, and α + β + γ = 1.

4. A cellular monitoring system for photovoltaic panels, applied to the highly reliable adaptive photovoltaic-storage-diesel fusion power supply system according to any one of claims 1 to 3, characterized in that: include: A plurality of hexagonal frames, which are closely arranged to form a continuous honeycomb structure and laid on the photovoltaic panel surface, and are used to divide the photovoltaic panel into a plurality of independent monitoring units, and the inner side surface of each hexagonal frame is provided with a sensor; A connecting slot disposed on a plane of each of the hexagonal frames; A reflective wall fixed in the connecting slot; The sensor system disposed in the monitoring unit of each hexagonal frame includes a temperature sensor, a light intensity sensor, and a voltage and current sensor; and a controller, for determining the abnormality type of each monitoring unit based on the temperature data, light intensity data and voltage and current data collected by the sensor system, wherein the abnormality type includes surface contamination, material aging and reflective wall abnormality; When an abnormality is detected, the reflective wall of the corresponding monitoring unit is controlled to adjust the angle and height to adjust the light distribution of the monitoring unit; by comparing the data of the target monitoring unit with the adjacent monitoring units, it is determined whether the abnormality is a local problem, a regional problem or a systemic problem.

5. The monitoring system according to claim 4, characterized in that: The reflective wall has three specifications: standard, enhanced and adjustable. The reflective surface of the reflective wall faces the center of the hexagonal frame. The adjustable reflective wall adopts an intelligent driving structure, including: The bottom driving mechanism adopts a bimetallic temperature difference driving device; Angle adjustment mechanism, using micro stepping motor; And a position sensor is used to monitor the working status of the reflective wall in real time.

6. A direct power dispatching system based on a honeycomb monitoring structure, applied to the mode management unit of the high-reliability adaptive photovoltaic-storage-diesel fusion power supply system as described in any one of claims 1-3, using the monitoring system as described in claim 4, based on multiple hexagonal frames laid on the photovoltaic panel surface as basic monitoring units, characterized in that: include: The micro group consists of 7 honeycomb units to form a basic monitoring unit, which is used to: realize power characteristic analysis of a single PV module; perform local performance optimization and fault diagnosis; and provide basic power regulation capabilities; The meso-group is a regional management unit composed of three micro-groups, which is used to: coordinate the power balance of each micro-group in the region; manage regional performance optimization and load distribution; and provide medium-scale power dispatch capabilities; The macro group, which is a system-level unit composed of three meso groups, is used to: achieve overall power management of large-scale photovoltaic arrays; perform system-level performance optimization and load balancing; and provide large-scale power scheduling capabilities; The power dispatch controller is used to: monitor the power output status of each group in real time; dynamically adjust the power allocation strategy according to changes in light and temperature environmental factors; Realize direct and efficient power supply from photovoltaic power generation units to power loads; automatically perform power redistribution when system parameters are abnormal.

7. The direct power dispatching system based on the cellular monitoring structure according to claim 6, characterized in that: Also includes: A power supply mode selection mechanism is used to automatically select direct power supply mode or energy storage intervention mode according to load characteristics and environmental conditions; The power dispatch optimizer uses a three-level grouping architecture of "micro-meso-macro" to optimize power allocation and adjust the grouping strategy in real time according to the matching degree; The fault detection and self-recovery module is used to analyze the operating status of each group, identify the fault location, and automatically perform group reconstruction or power redistribution when power anomalies are detected to achieve the system's self-recovery function.

Images