Photovoltaic module dynamic hot spot protection and micro-grid cooperative control optimization system and method

Abstract

The invention belongs to the technical field of photovoltaic module safety protection and micro-grid control, and discloses a photovoltaic module dynamic hot spot protection and micro-grid cooperative control optimization system and method. Four invention points of a dynamic hot spot intelligent protection algorithm, an inverter adaptive harmonic suppression model, a microgrid island rapid detection mechanism and a global cooperative control framework are provided, and two innovative algorithms of hot spot risk assessment and harmonic suppression coefficient adjustment are included. By constructing a protection-suppression-detection global collaborative closed loop, the hot spot protection precision of the photovoltaic module is remarkably improved, the harmonic suppression effect of the inverter is greatly enhanced, the microgrid island detection response speed is remarkably improved, and the method is suitable for distributed photovoltaic, photovoltaic microgrid and other scenes and has a wide application prospect. And the safety of the photovoltaic system and the operation stability of the micro-grid are comprehensively improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of photovoltaic module safety protection and micro-grid control, and discloses a photovoltaic module dynamic hot spot protection and micro-grid collaborative control optimization system and method. BACKGROUND

[0002] In the field of photovoltaic energy technology, the existing technology has the following subproblems:

[0003] Low precision of photovoltaic module hot spot protection: traditional hot spot protection relies on a fixed temperature threshold (such as 70℃ alarm), and only uses single-point temperature monitoring to judge hot spot risk, without considering the coupling relationship between local shadow dynamic distribution (such as cloud shadow moving track, component surface dust accumulation difference) and current mismatch in component series and parallel topology, resulting in hot spot early warning lag, partial potential hot spots (temperature not reaching threshold but power attenuation exceeding 30%) being missed, and high local overheating damage rate of components.

[0004] Poor adaptability of photovoltaic inverter harmonic suppression: the inverter harmonic suppression uses fixed filter parameters (such as preset LC filter cutoff frequency), which is effective for single load type (such as purely resistive load), but cannot adapt to dynamic load changes (such as impact load during motor start, nonlinear load proportion fluctuation), and the harmonic content fluctuates greatly (total harmonic distortion rate 2%-8%) under different working conditions, especially in weak grid scenarios, the harmonic amplification phenomenon is obvious, affecting power quality and equipment life.

[0005] Slow response of photovoltaic micro-grid island detection: island detection relies on passive voltage / frequency mutation judgment (such as voltage deviation exceeding ±5% trigger), without combining micro-grid topology structure (such as distribution of distributed power access points) and load characteristics (such as impact load start-up rules), in weak grid or light load conditions, voltage / frequency changes slowly, detection response delay is more than 200ms, which easily leads to long island operation time, causing safety hazards to maintenance personnel, and may cause equipment overvoltage damage. SUMMARY

[0006] The present application aims to solve the technical problems of low precision of photovoltaic module hot spot protection due to not considering the coupling of shadow dynamic distribution and current mismatch, poor adaptability of photovoltaic inverter harmonic suppression due to fixed filter parameters unable to adapt to dynamic load changes, and slow response of photovoltaic micro-grid island detection due to passive judgment, by constructing a dynamic hot spot protection mechanism, an adaptive harmonic suppression strategy, a fast island detection model, and a global collaborative framework, to improve the safety, power quality, and micro-grid operation reliability of photovoltaic systems.

[0007] A photovoltaic module dynamic hot spot protection and micro-grid collaborative control optimization system, comprising:

[0008] The dynamic hot spot protection unit collects local shadow dynamic parameters, component series and parallel current parameters and temperature distribution data, calculates a hot spot risk level through a dynamic hot spot intelligent protection algorithm, and outputs a hot spot protection instruction.

[0009] The adaptive harmonic suppression unit collects load dynamic parameters, grid impedance parameters and initial harmonic content, calculates a filter coefficient correction amount through an inverter adaptive harmonic suppression model, and outputs an optimized filter coefficient.

[0010] The island rapid detection unit collects micro-grid topology parameters, load characteristics and voltage / frequency dynamic data, analyzes impedance mutation characteristics through a micro-grid island rapid detection mechanism, and outputs an island state detection result.

[0011] The global cooperative control center receives the hot spot protection instruction, the filter coefficient and the island detection result, constructs a global cooperative target and a constraint condition, generates a global optimization control instruction, and coordinates the operation of each unit.

[0012] The formula of the dynamic hot spot intelligent protection algorithm is:

[0013] R=ω S ·Shade(S)+ω I ·Iso(I)+ω T ·Temp(T)

[0014] Wherein, ω S +ω I +ω T =1, S is a local shadow dynamic parameter, I is a component series and parallel current parameter, T is a component temperature distribution; ω S , ω I , ω T are weight coefficients, Shade(·) is a shadow influence function, Iso(·) is a current mismatch function, Temp(·) is a temperature risk function, and the hot spot risk is evaluated by fusing multi-dimensional parameters.

[0015] The formula of the inverter adaptive harmonic suppression model is:

[0016] K=I-AHSM(L,Z,H0)=K0+ΔK·Load(L)·Imped(Z)+α·Dist(H0)

[0017] Wherein K0 is an initial filter coefficient, L is a load dynamic parameter, Z is a grid impedance parameter, H0 is an initial harmonic content, K is an optimized filter coefficient, ΔK is a coefficient correction amount, Load(·) is a load characteristic function, Imped(·) is an impedance influence function, Dist(·) is a harmonic distortion function, and α is a correction weight.

[0018] The modeling process of the dynamic hot spot protection unit comprises: defining local shadow dynamic parameters as shadow area, moving speed and shielding duration, current parameters as branch current deviation rate and reverse voltage value, and temperature distribution as local temperature difference and temperature rise rate, and generating hot spot protection instructions through weighted calculation.

[0019] The modeling process of the adaptive harmonic suppression unit comprises: quantifying load dynamic parameters as nonlinear load proportion and impact current peak value, and grid impedance parameters as fundamental impedance and harmonic impedance characteristics, and outputting optimized harmonic suppression parameters through modified filter coefficients.

[0020] The modeling process of the island rapid detection unit comprises: encoding micro-grid topology parameters, load characteristics and voltage / frequency dynamic data as graph node attributes, constructing a micro-grid-load correlation graph, designing active disturbance and impedance mutation analysis rules, and outputting island state detection results.

[0021] The modeling process of the global collaborative control center comprises: determining global collaborative targets of hot spot protection priority, harmonic suppression target and island detection response time, setting constraint conditions of system operation, constructing a global collaborative control model, and solving to generate global optimization control instructions.

[0022] The application also provides a photovoltaic module dynamic hot spot protection and micro-grid collaborative control optimization method, comprising:

[0023] The dynamic hot spot protection step: collecting local shadow dynamic parameters, module current parameters and temperature distribution data, calculating hot spot risk level through a dynamic hot spot intelligent protection algorithm, and outputting hot spot protection instructions;

[0024] The adaptive harmonic suppression step: collecting load dynamic parameters, grid impedance parameters and initial harmonic content, calculating filter coefficient correction amount through an inverter adaptive harmonic suppression model, and outputting optimized filter coefficients;

[0025] The island rapid detection step: collecting micro-grid topology parameters, load characteristics and voltage / frequency data, analyzing impedance mutation characteristics through a micro-grid island rapid detection mechanism, and outputting island state detection results;

[0026] The global collaborative control step: receiving hot spot protection instructions, filter coefficients and island detection results, constructing global collaborative targets and constraint conditions, and generating global optimization control instructions.

[0027] The dynamic hot spot protection step comprises: quantifying local shadow parameters, current parameters and temperature distribution data as corresponding characteristic values, calculating hot spot risk level through weighted calculation, and generating protection instructions including current adjustment and heat dissipation control according to the risk level.

[0028] The adaptive harmonic suppression step comprises: inputting the load parameter and grid impedance data into a harmonic suppression model, calculating a filter coefficient correction amount, and dynamically adjusting the inverter filter parameter to suppress harmonic amplification under different load conditions.

[0029] Advantages:

[0030] Improved precision of photovoltaic module hot spot protection: The dynamic hot spot intelligent protection algorithm makes the hot spot risk assessment more in line with the actual working conditions, reduces potential hot spot misjudgment, reduces the damage rate of local overheating of the module, especially in cloudy weather with dynamic shadow changes, the timeliness of the protection response is significantly improved, the service life of the module is extended, and the maintenance and replacement cost is reduced.

[0031] Enhanced inverter harmonic suppression effect: The adaptive harmonic suppression model improves the adaptability of harmonic suppression for different load types (such as motors, frequency converters), reduces the total harmonic distortion rate fluctuation range, reduces harmonic amplification in weak grids, improves power quality, enhances compatibility between photovoltaic equipment and the grid, and reduces equipment failures caused by harmonics.

[0032] Improved microgrid island detection responsiveness: The microgrid island rapid detection mechanism shortens the delay of island state identification, reduces the safety risk of maintenance personnel, reduces the overvoltage damage of equipment, and makes the switching between microgrid and main grid more stable. Especially in weak grid and light load conditions, the detection reliability is significantly improved, and the safety of the microgrid operation is enhanced.

[0033] Optimized overall system global operation safety: The global collaborative control framework makes each link work collaboratively, hot spot protection lays the foundation for system safety, harmonic suppression ensures power quality, and island detection avoids grid risks. The three linkages improve the safety redundancy of the photovoltaic system, adapt to complex grid environments and dynamic load changes, and comprehensively improve the overall operation reliability. BRIEF DESCRIPTION OF DRAWINGS

[0034] Figure 1 System overall flowchart;

[0035] Figure 2 Dynamic hot spot intelligent protection algorithm architecture diagram;

[0036] Figure 3 Global collaborative control framework architecture diagram. DETAILED DESCRIPTION

[0037] To solve the problem of low hot spot protection accuracy, a protection method based on shadow dynamics and current mismatch is proposed, and the algorithm formula is

[0038] R=ω S ·Shade(S)+ω I ·Iso(I)+ω T ·Temp(T)

[0039] where ω S +ω I +ω T = 1, S is a local shadow dynamic parameter, I is a component string and parallel current parameter, T is a component temperature distribution; ω S , ω I , ω T are weight coefficients, Shade(·) is a shadow influence function, Iso(·) is a current mismatch function, and Temp(·) is a temperature risk function, and the thermal spot risk is evaluated by fusing multi-dimensional parameters.

[0040] In view of the poor adaptability of harmonic suppression, a suppression model based on load characteristics and grid impedance is constructed, and the algorithm formula is:

[0041] K = I - AHSM(L, Z, H0) = K0 + ΔK·Load(L)·Imped(Z) + α·Dist(H0)

[0042] In the formula, K0 is an initial filter coefficient, L is a load dynamic parameter (such as a nonlinear load ratio and an impact current peak value), Z is a grid impedance parameter (such as a fundamental impedance and a harmonic impedance frequency characteristic), H0 is an initial harmonic content, K is an optimized filter coefficient, ΔK is a coefficient correction amount, Load(·) is a load characteristic function, Imped(·) is an impedance influence function, Dist(·) is a harmonic distortion function, and α is a correction weight. The model dynamically adjusts the filter parameters to achieve harmonic suppression under a wide range of loads.

[0043] In view of the slow response of island detection, a detection strategy based on topological characteristics and load dynamics is designed. A microgrid-load correlation graph is constructed, and the nodes include a distributed power supply access point, a load type, a line impedance, and a voltage / frequency dynamic characteristic. By injecting a small perturbation signal (such as a current perturbation of a specific frequency), the grid impedance mutation characteristics and the load feedback response are monitored, and the topological structure analysis is combined to realize the rapid identification and response of the island state.

[0044] In view of the lack of cooperation of each link of the system, a global cooperation method of thermal spot protection-harmonic suppression-island detection is proposed. The edge terminal collects component state and load data to realize real-time thermal spot protection and dynamic harmonic suppression; the cloud platform analyzes the microgrid topology and operating state to optimize the island detection strategy; the cooperation constraints (such as the high priority of thermal spot protection over harmonic suppression) and the goals (such as the power quality optimization under the premise of safe operation of the microgrid) of the three are established to generate global control instructions, avoiding system conflicts caused by local optimization.

[0045] The dynamic hot spot intelligent protection algorithm breaks through the limitation of single temperature monitoring by fusing shadow dynamics, current mismatch, and temperature distribution multi-dimensional parameters, significantly improves the accuracy of hot spot risk assessment, and realizes the advanced protection of hot spots.

[0046] The inverter adaptive harmonic suppression model dynamically adjusts the filtering parameters based on load characteristics and grid impedance, overcomes the lack of adaptability of fixed parameter suppression, and significantly enhances the harmonic suppression effect under different load conditions, especially improves the harmonic amplification problem in weak grids.

[0047] The micro-grid island rapid detection mechanism combines topology analysis and active disturbance monitoring to solve the response lag problem of passive detection, significantly improves the identification speed and accuracy of island state, and shortens the island operation time.

[0048] The global collaborative control framework realizes the deep collaboration of hot spot protection, harmonic suppression and island detection, the protection strategy provides a safe boundary for suppression, the detection result drives the adjustment of protection and suppression parameters, and the overall safety and operation efficiency of the photovoltaic system are comprehensively improved.

[0049] Advantages:

[0050] Improved accuracy of photovoltaic module hot spot protection: The dynamic hot spot intelligent protection algorithm makes the hot spot risk assessment more in line with the actual working condition, reduces the potential hot spot misjudgment phenomenon, reduces the local overheating damage rate of the module, especially in the scene of dynamic shadow change in cloudy weather, the timeliness of the protection response is obviously improved, the service life of the module is prolonged, and the maintenance and replacement cost is reduced.

[0051] Enhanced inverter harmonic suppression effect: The adaptive harmonic suppression model improves the adaptability of harmonic suppression for different load types (such as motors, frequency converters), reduces the total harmonic distortion rate fluctuation range, reduces the harmonic amplification phenomenon in weak grids, improves the power quality of the grid, enhances the compatibility of photovoltaic equipment and the grid, and reduces equipment failures caused by harmonics.

[0052] Improved micro-grid island detection responsiveness: The micro-grid island rapid detection mechanism shortens the delay of island state identification, reduces the safety risk of maintenance personnel, reduces the overvoltage damage phenomenon of equipment, and makes the switching of micro-grid and main grid more stable, especially in weak grid and light load conditions, the detection reliability is significantly improved, and the operation safety of micro-grid is enhanced.

[0053] Optimized overall system global operation safety: The global collaborative control framework makes each link work collaboratively, hot spot protection lays the foundation for equipment safety, harmonic suppression guarantees power quality, island detection avoids grid risks, and the three linkages improve the safety redundancy of photovoltaic systems, adapt to complex grid environments and dynamic load changes, and comprehensively improve the overall operation reliability. DETAILED DESCRIPTION

[0055] Embodiment 1: Distributed photovoltaic module hot spot protection (solve the problem of low accuracy of hot spot protection)

[0056] Prior art defects: Distributed photovoltaic modules use single-point temperature monitoring (such as module center temperature), in the case of local shadow (such as tree branches blocking the corners of the module), the shadow area current mismatch is not detected, the temperature does not reach the 70℃ threshold, and the protection is not triggered, resulting in the formation of hot spots (power attenuation 40%) in the shadow area, and cracks appear in the local glass of the module after 3 months.

[0057] The present application realizes the steps of:

[0058] The dynamic hot spot intelligent protection algorithm collects local shadow parameters (corner shadow area 0.1m 2 , moving speed 0.5m / h), current parameters (shadow branch current is 30% lower than normal branch, reverse voltage 5V), temperature distribution (shadow area temperature 62℃, 8℃ higher than other areas);

[0059] Calculate the hot spot risk level (shadow impact weight 0.4, current mismatch weight 0.4, temperature risk weight 0.2, comprehensive risk 85 points);

[0060] Trigger active protection (reduce the working current of this branch by 20%, start local cooling fan), inhibit the formation of hot spots.

[0061] Synergistic comparison: The hot spot formation rate of distributed photovoltaic modules in the local dynamic shadow scene is reduced, the local overheating phenomenon of the module is reduced, the damage problem such as glass cracking is reduced, even if the temperature does not reach the traditional threshold, the potential hot spot can be inhibited in time, the safety of the module operation is significantly improved, especially suitable for distributed photovoltaic systems around trees.

[0062] Embodiment 2: Harmonic suppression of industrial and commercial photovoltaic inverters (solve the problem of poor adaptability of harmonic suppression)

[0063] Prior art defects: Industrial and commercial photovoltaic inverters use fixed LC filtering (cutoff frequency 1kHz), when the workshop load changes (such as 10 motor operations during the day and only 2 at night), the harmonic content fluctuates greatly (total harmonic distortion rate 7.5% during the day, 3.2% at night), the harmonic is amplified to 5.8% at night under weak grid conditions, resulting in increased measurement error of precision instruments and affecting product quality.

[0064] The present application realizes the steps of:

[0065] The adaptive harmonic suppression model collects load parameters (nonlinear load ratio 60% during the day, 15% at night), grid impedance (weak grid fundamental impedance 2Ω, harmonic impedance increases with frequency);

[0066] Dynamic adjustment of filter coefficients (30% increase in high-frequency attenuation coefficient during the day, 20% reduction in low-frequency filter strength at night)

[0067] Stable inverter output harmonic content (total distortion rate 3.5% during the day, 2.1% at night), no obvious amplification under weak power grid.

[0068] Synergistic comparison: The harmonic suppression effect of industrial and commercial photovoltaic inverters in the load dynamic change scene is stable, the total harmonic distortion rate fluctuation range is reduced, the harmonic amplification phenomenon in the weak power grid is reduced, the measurement error of precision instruments is reduced, the product quality stability is improved, and the acceptance capacity of the power grid to the photovoltaic system is enhanced.

[0069] Example 3: Photovoltaic micro-grid island detection in industrial park (solve the problem of slow island detection response)

[0070] Defects in the prior art: The photovoltaic micro-grid in the industrial park adopts passive island detection (voltage deviation ±5% trigger), under the weak power grid light load working condition (main grid voltage fluctuation 3%, load only 20% of rated value), after the main grid is disconnected, the voltage slowly rises, the detection response is delayed by 250ms, resulting in too long island operation time and overvoltage damage of the capacitor compensation device.

[0071] The implementation steps of the present application are:

[0072] The micro-grid island rapid detection mechanism collects the micro-grid topology (2 photovoltaic access points, 1 tie line) and load characteristics (light load mainly with lighting load, no impact characteristics);

[0073] Inject a specific frequency current disturbance (0.5Hz small current fluctuation), monitor the impedance mutation of the tie line (the impedance rises from 1.2Ω to 5Ω after the main grid is disconnected);

[0074] Combine topology analysis to identify island state, response delay is shortened to 80ms, photovoltaic grid-connected switch is immediately disconnected.

[0075] Synergistic comparison: The island detection response speed of the photovoltaic micro-grid in the industrial park under the weak power grid light load working condition is significantly accelerated, the island operation time is shortened, the overvoltage damage phenomenon of the capacitor compensation device is reduced, the safety risk of maintenance personnel entering the isolated area is reduced, the switching of the micro-grid and the main grid is more stable, and the operation reliability is improved.

[0076] Example 4: Photovoltaic array maximum power point tracking (MPPT) adaptive control

[0077] The traditional MPPT adopts fixed step perturbation observation method, when the cloud layer is quickly blocked, the light intensity fluctuates sharply (such as from 1000W / m 2 to 300W / m 2) time, the tracking lag is obvious, the maximum power point capture efficiency is low, the array power generation loss is more than 15%, and it is easy to fall into a suboptimal working point under local shadow. In view of the problems of tracking efficiency and insufficient dynamic response, an adaptive MPPT algorithm based on illumination gradient-power characteristics is proposed.

[0078] Algorithm principle:

[0079] Multi-scale detection: the instantaneous characteristics of photovoltaic array output power P and voltage U are obtained through high-frequency sampling (10 kHz), the illumination change gradient G (the rate of change of illumination intensity per unit time) is calculated, and G> 500 W / (m 2 ·s) is defined as a severe fluctuation scene.

[0080] Dynamic step adjustment: a step decision model is constructed, when G≤200W / (m 2 ·s), a small step (0.5V) is used for fine search; when 200<G≤500, a medium step (1V) is used to balance speed and accuracy; when G>500, a global scanning mode (step 5V) is switched to quickly locate the power peak value region.

[0081] Local shadow identification and avoidance: the shadow boundary is identified through the power difference △P of adjacent components, a power-voltage surface model is established, and an improved particle swarm algorithm is used to optimize on the surface model, skipping the suboptimal peak point caused by local shadow.

[0082] Implementation process:

[0083] Real-time monitoring of illumination gradient and power curve slope, when severe fluctuations or multiple peak characteristics of power curve are detected, the dynamic step and global optimization mechanism are started, the current optimal working voltage instruction is output, and the DC / DC converter is driven to adjust the array output.

[0084] Synergistic effect comparison: the maximum power point tracking accuracy of the photovoltaic array under rapidly changing illumination and local shadow scene is significantly improved, the power generation loss is greatly reduced, the efficiency decay caused by misjudging the suboptimal working point is avoided, and it is especially suitable for photovoltaic systems under cloudy weather and complex shading environment.

[0085] Example 5: charge-discharge synergistic optimization of photovoltaic energy storage system

[0086] Defects of prior art: the photovoltaic energy storage system adopts a fixed charge-discharge strategy (such as full-power charging when the illumination is sufficient, and fixed-power discharging at night), without considering the peak-valley price difference, load fluctuation and battery cycle life, resulting in high energy storage operation cost (0.1 yuan increase in degree electric cost), accelerated battery degradation (cycle life shortened by 20%), and difficulty in coping with sudden load growth.

[0087] Implementation steps of the present application:

[0088] To solve the problem of poor adaptability of charging and discharging strategy, a multi-objective prediction-based energy storage collaborative optimization model is constructed.

[0089] Algorithm principle:

[0090] Multi-objective modeling: The input parameters include short-term photovoltaic output prediction Ppv(t), load prediction P1(t), time-of-use electricity price C(t), and current SOC (state of charge) and SOH (state of health) of the battery.

[0091] Constraint construction: The constraint conditions are defined as "battery SOC ∈ [20%, 90%]", "charging and discharging power ≤ rated power", and "single discharge depth ≤ 70% (to prolong the life)".

[0092] Rolling optimization solution: Model predictive control (MPC) is adopted, and the prediction data is updated every 15 minutes. The optimal charging and discharging plan is solved within the prediction time domain (4 hours). The remaining power is charged during the low electricity price period, discharged during the high electricity price period or when the photovoltaic output is insufficient, and 20% capacity is reserved to cope with sudden load.

[0093] Implementation process:

[0094] The edge node collects photovoltaic output, load and battery state data in real time, combines with cloud time-of-use electricity price information, and generates a charging and discharging power curve for the next 4 hours through MPC algorithm rolling, dynamically adjusts the charging and discharging rate of the energy storage system, and triggers the emergency discharge mechanism when the actual load exceeds the prediction value by 10%.

[0095] Efficiency comparison: The comprehensive operation cost of the photovoltaic energy storage system is significantly reduced, the battery cycle life is effectively prolonged, the adaptability to load fluctuations and electricity price policy is enhanced, and the system economy is improved.

[0096] Example 6: Enhanced control of photovoltaic grid-connected inverter low voltage ride through

[0097] Defects of prior art: The photovoltaic grid-connected inverter adopts a fixed low voltage ride through (LVRT) strategy (such as maintaining grid connection when the voltage drops to 80% of the rated value, and disconnecting when the voltage drops to 50%), without considering the dynamic characteristics of voltage drop depth and duration. When the voltage drops to 30% and the duration is long (> 2 seconds), it is easy to disconnect from the grid due to overcurrent protection action, resulting in a sudden decrease in photovoltaic output when the power grid recovers, affecting system stability.

[0098] Implementation steps of the present application:

[0099] To solve the problem of insufficient LVRT capability, an adaptive ride through control strategy based on voltage drop characteristics is designed.

[0100] Algorithm principle:

[0101] Drop feature recognition: Real-time monitoring of grid voltage amplitude U(t) and phase θ(t), calculating drop depth D=(U_N-U(t)) / U_N(U_N is rated voltage) and drop duration T, constructing "depth-time" feature matrix, dividing mild (D≤30%), moderate (30%<D≤60%), severe (D>60%) drop levels.

[0102] Hierarchical control strategy:

[0103] Mild drop: maintain current active output, support grid voltage through reactive compensation (inject 20% rated reactive power);

[0104] Moderate drop: quickly reduce active output (to 50%), increase reactive power injection (40% rated reactive power), activate DC side capacitor energy buffer;

[0105] Severe drop: immediately cut off 80% active output, full power output reactive to suppress voltage further drop, while starting DC side unloading circuit to avoid capacitor overvoltage, only when the drop duration exceeds the threshold (such as 3 seconds) and the voltage has no recovery trend, the grid disconnection is performed.

[0106] Implementation process:

[0107] The inverter detects voltage drop characteristics in real time, matches the corresponding control strategy, dynamically adjusts the active / reactive output ratio and DC side energy management, until the grid voltage recovers to more than 90% of the rated value, and then gradually restores the active output.

[0108] Efficiency comparison: The low voltage ride-through capability of photovoltaic grid-connected inverters in different depth and duration of voltage drop scenarios is significantly enhanced, avoiding unnecessary grid disconnection, ensuring the stability of photovoltaic output during grid failure, and improving the support capability of photovoltaic system to grid failure.

[0109] Example 7: Active-reactive power coordinated control of distributed photovoltaic cluster

[0110] Defects of prior art: The active and reactive power outputs of each inverter in the distributed photovoltaic cluster are controlled independently, without global coordination, resulting in large voltage fluctuations at the grid connection point (±5% of rated value), high line loss rate (increased by 2%), and inability to respond to reactive power auxiliary service instructions (such as voltage regulation requirements) of the grid dispatch, resulting in low overall operating efficiency of the cluster.

[0111] Implementation steps of the present application:

[0112] To solve the problem of insufficient cluster coordination, an active-reactive power coordinated control method based on voltage sensitivity is proposed.

[0113] Algorithm principle:

[0114] Sensitivity modeling: Through offline calculation and online correction, the sensitivity matrix S of the active / reactive power output of each inverter and the grid-connected point voltage is established, where S_ij represents the influence quantity of the unit reactive power output of inverter i on the voltage of node j, which is used to quantify the regulation effect.

[0115] Hierarchical control architecture:

[0116] Cloud layer: Receive grid dispatching instructions (such as grid-connected point voltage target Uref), calculate the total reactive power demand Qtotal of the cluster = ∑(S_ij 1 ·(Uref-Uj));

[0117] Edge layer: According to the current active power output P_i of each inverter (to avoid the influence of reactive power regulation on active power generation), the electrical distance from the grid-connected point (near-end inverters are allocated more reactive power tasks), the Qtotal is proportionally allocated to each inverter while the reactive power output of a single inverter is constrained ≤ 30% of the rated capacity.

[0118] Active power coordination: When the grid-connected point voltage is out of limit and the reactive power regulation space is insufficient, the inverter with the greatest influence on the voltage is determined through sensitivity analysis, and its active power output is appropriately reduced (≤ 10%) to assist voltage recovery.

[0119] Implementation process:

[0120] The cloud platform calculates the total reactive power demand in real time and distributes it to each edge node, and the edge node allocates reactive power instructions according to the local inverter state and sensitivity coefficient, dynamically adjusts the power factor of the inverter, and realizes precise voltage control and active power output coordination at the cluster level.

[0121] Efficiency comparison: The voltage stability of the distributed photovoltaic cluster grid-connected point is significantly improved, the fluctuation range is greatly reduced, the line loss rate is reduced, and the ability to respond to grid reactive power auxiliary services is enhanced, the overall operation efficiency of the cluster and the friendliness to the grid are enhanced, and it is suitable for high penetration rate distributed photovoltaic access to distribution networks.

[0122] Example 8: Building Integrated Photovoltaics (BIPV) temperature adaptive regulation

[0123] Defects of prior art: BIPV components (such as photovoltaic curtain walls) use fixed heat dissipation design (such as natural ventilation), without considering building orientation, sunshine angle and indoor and outdoor temperature difference, resulting in high component operating temperature (exceeding 65℃), reduced power generation efficiency (efficiency decreases by 0.5% per 1℃ increase), and increased heat transfer to the room in summer, resulting in a 10% increase in air conditioning load.

[0124] Implementation steps of the present application:

[0125] Aiming at the problem of poor temperature regulation adaptability, a BIPV temperature self-adaptive regulation system based on multi-field coupling is designed.

[0126] Algorithm principle:

[0127] Temperature field modeling: A correlation model of BIPV component temperature T and influencing factors is constructed, and the input parameters include: solar irradiance G, ambient temperature Ta, indoor set temperature Tis, component inclination angle θ, and ventilation fan speed v.

[0128] Dual-target decision: Taking "component working temperature ≤ 55℃ (to ensure power generation efficiency)" and "indoor and outdoor heat transfer power Q ≤ design value (to reduce air conditioning load)" as the target, the regulation priority is defined: in summer (Tis < 26℃), Q is preferentially controlled, and in winter (Tis > 20℃), power generation efficiency is preferentially ensured (allowing T to rise to 60℃ appropriately).

[0129] Adaptive execution: A fuzzy control algorithm is adopted to dynamically adjust the ventilation fan speed (0-1000rpm) and the power of semiconductor refrigeration piece (only started at extreme high temperature), and to coordinate with the building shading system (such as deploying the sunshade board when the irradiance is >800W / m 2 ).

[0130] Implementation process:

[0131] The component surface and backboard are installed with temperature sensors to monitor the working temperature and heat transfer power in real time, and the edge controller outputs device regulation instructions according to the fuzzy control rules to realize the coordination of heat dissipation and building heat management.

[0132] Synergistic comparison: The working temperature of BIPV component under different seasons and solar conditions is precisely controlled, the power generation efficiency decay is significantly reduced, and the adverse effects on the building indoor thermal environment are effectively reduced, the air conditioning system energy consumption is reduced, and the comprehensive energy efficiency of the building photovoltaic integrated system is improved.

[0133] Example 9: Frequency stability control of photovoltaic micro-grid isolated network operation

[0134] Defects of prior art: When the photovoltaic micro-grid isolated network operates, due to the fluctuation of photovoltaic output (±20%) and the sudden change of load (such as motor starting), the system frequency deviation is large (±0.5Hz), which exceeds the allowed range (±0.2Hz), resulting in that sensitive load (such as precision instruments) cannot work normally, and in serious cases, the protection shutdown is triggered.

[0135] Implementation steps of the present application:

[0136] Aiming at the problem of poor isolated network frequency stability, a collaborative control strategy based on virtual inertia and droop characteristics is proposed.

[0137] Algorithm principle:

[0138] Virtual inertia simulation: Introduce virtual inertia control for photovoltaic inverters, detect the frequency change rate df / dt, output active compensation amount ΔP = J·df / dt (J is the virtual inertia coefficient, which is dynamically adjusted according to system capacity, and the larger the load, the larger the value of J), simulate the inertia response of synchronous generator, and suppress the rapid drop / rise of frequency.

[0139] Adaptive droop control: Design a variable slope droop characteristic curve P = k·(fN-f), where the droop coefficient k is dynamically adjusted according to the current photovoltaic output margin (Pmax-Ppv): when the margin is >30%, k takes a small value (allowing a small range of frequency fluctuations to prioritize photovoltaic utilization); when the margin is <10%, k takes a large value (enhancing the frequency regulation strength).

[0140] Energy storage cooperation: When the frequency deviation exceeds 0.3Hz, trigger the energy storage system to respond quickly and output instantaneous active support (discharge or charge), until the virtual inertia and droop control pull the frequency back to the allowed range.

[0141] Implementation process:

[0142] The microgrid central controller monitors the system frequency and photovoltaic output margin in real time, dynamically adjusts the virtual inertia coefficient and droop coefficient of the inverter, and when a large frequency deviation is detected, the energy storage system is linked to perform active compensation, maintaining frequency stability.

[0143] Efficiency comparison: The frequency stability of photovoltaic microgrid in isolated network operation mode is significantly improved, the frequency deviation is controlled within the allowed range, effectively avoiding abnormal sensitive load and system shutdown caused by frequency fluctuation, enhancing the independent operation ability and power supply reliability of microgrid.

[0144] Example 10: Early warning and positioning of photovoltaic array failure

[0145] Prior art defects: Photovoltaic array failure detection relies on regular manual inspection or threshold alarm (such as current below 10% of rated value), which cannot identify hidden failures (such as diode aging, loose wiring) early, leading to expansion of failures (such as a single component failure causing a 5% decrease in series branch efficiency), and time-consuming fault location (average troubleshooting time 2 hours).

[0146] Implementation steps of the present application:

[0147] To solve the problems of failure early warning lag and positioning difficulty, a fault diagnosis model based on multi-dimensional features is constructed.

[0148] Algorithm principle:

[0149] Feature extraction: Collect multi-dimensional operation data of photovoltaic array - time domain features (mean, variance) of string current I, voltage U, power P, frequency domain features (harmonic content), component temperature distribution T(x, y), and construct normal operation feature library.

[0150] Abnormality detection: Use the Isolation Forest algorithm to calculate the deviation S of real-time features from the normal feature library, identify implicit abnormalities (S>0.6 triggers early warning, which has not reached the traditional threshold), and distinguish fault types (such as short circuit, open circuit, performance degradation).

[0151] Fault location: Combine with array topology structure, construct "feature-position" correlation map, use feature difference of adjacent strings (such as I1-I2>5%) to locate fault string, and then lock specific fault component through component-level temperature distribution (abnormal temperature rise / fall of fault component).

[0152] Implementation process:

[0153] Edge nodes collect and analyze operation data in real time, calculate feature deviation S, send early warning information to the monitoring platform when S exceeds the early warning threshold, and output possible fault location and type combined with topology map to guide operation and maintenance personnel to accurately investigate.

[0154] Efficiency comparison: The early identification ability of photovoltaic array implicit faults is significantly enhanced, the fault expansion phenomenon is effectively curbed, the fault location time is greatly shortened, the operation and maintenance efficiency is improved, the power generation loss caused by faults is reduced, and the operation and maintenance cost is reduced.

Claims

1. A photovoltaic module dynamic hot spot protection and microgrid collaborative control optimization system, characterized in that, Comprising: a dynamic hot spot protection unit that collects local shadow dynamic parameters, component string and parallel current parameters, and temperature distribution data, calculates a hot spot risk level through a dynamic hot spot intelligent protection algorithm, and outputs a hot spot protection instruction; an adaptive harmonic suppression unit that collects load dynamic parameters, grid impedance parameters, and initial harmonic content, calculates a filter coefficient correction amount through an inverter adaptive harmonic suppression model, and outputs an optimized filter coefficient; an island rapid detection unit that collects microgrid topology parameters, load characteristics, and voltage / frequency dynamic data, analyzes impedance mutation characteristics through a microgrid island rapid detection mechanism, and outputs an island state detection result; a global collaborative control center that receives the hot spot protection instruction, the filter coefficient, and the island detection result, constructs a global collaborative target and a constraint condition, generates a global optimization control instruction, and coordinates the operation of each unit.

2. The system of claim 1, wherein, The formula of the dynamic hot spot intelligent protection algorithm is: R = ω S • Shade (S) + ω I • I SO (I) + ω T • Temp (T) where ω S +ω I +ω T = 1, S is a local shadow dynamic parameter, I is a component series-parallel current parameter, T is a component temperature distribution; ω S , ω I , ω T are weight coefficients, Shade(·) is a shadow influence function, I SO (·) is a current mismatch function, and Temp(·) is a temperature risk function, which fuses multi-dimensional parameters to evaluate hot spot risk.

3. The system of claim 1, wherein, The formula of the inverter adaptive harmonic suppression model is: K = I - AHSM(L, Z, H0) = K0 + △K·Load(L)·Imped(Z) + α·Dist(H0) where K0 is an initial filter coefficient, L is a load dynamic parameter, Z is a grid impedance parameter, H0 is an initial harmonic content, K is an optimized filter coefficient, △K is a coefficient correction amount, Load(·) is a load characteristic function, Imped(·) is an impedance influence function, Dist(·) is a harmonic distortion function, and α is a correction weight.

4. The system of claim 1, wherein, The modeling process of the dynamic hot spot protection unit includes defining local shadow dynamic parameters as shadow area, moving speed, and shading duration, current parameters as branch current deviation rate and reverse voltage value, and temperature distribution as local temperature difference and temperature rise rate, and generating a hot spot protection instruction through weighted calculation.

5. The system of claim 1, wherein, The modeling process of the adaptive harmonic suppression unit includes quantifying load dynamic parameters as nonlinear load proportion and impact current peak value, and grid impedance parameters as fundamental impedance and harmonic impedance characteristics, and outputting optimized harmonic suppression parameters by correcting filter coefficients.

6. The system of claim 1, wherein, The modeling process of the island rapid detection unit includes encoding microgrid topology parameters, load characteristics, and voltage / frequency dynamic data as graph node attributes, constructing a microgrid-load correlation graph, designing active disturbance and impedance mutation analysis rules, and outputting an island state detection result.

7. The system of claim 1, wherein, The modeling process of the global collaborative control center includes determining global collaborative targets of hot spot protection priority, harmonic suppression target, and island detection response time, setting constraint conditions for system operation, constructing a global collaborative control model, and solving to generate a global optimization control instruction.

8. A method for dynamic hot spot protection of photovoltaic modules and coordinated control optimization with microgrid, characterized in that, Comprising: a dynamic hot spot protection step of collecting local shadow dynamic parameters, component current parameters, and temperature distribution data, calculating a hot spot risk level through a dynamic hot spot intelligent protection algorithm, and outputting a hot spot protection instruction; an adaptive harmonic suppression step of collecting load dynamic parameters, grid impedance parameters, and initial harmonic content, calculating a filter coefficient correction amount through an inverter adaptive harmonic suppression model, and outputting an optimized filter coefficient; Island fast detection step: Collecting microgrid topology parameters, load characteristics and voltage / frequency data, analyzing impedance mutation characteristics through microgrid island fast detection mechanism, and outputting island state detection results; Global cooperative control step: Receiving hot spot protection instructions, filter coefficients and island detection results, building global cooperative target and constraint conditions, and generating global optimization control instructions.

9. The method of claim 8, wherein, The dynamic hot spot protection step includes: quantifying local shadow parameters, current parameters and temperature distribution data into corresponding characteristic values, calculating hot spot risk level through weighted calculation, and generating protection instructions including current adjustment and heat dissipation control according to risk level.

10. The method of claim 8, wherein, The adaptive harmonic suppression step includes: inputting load parameters and grid impedance data into harmonic suppression model, calculating filter coefficient correction amount, and dynamically adjusting inverter filter parameters to suppress harmonic amplification under different load conditions.

Images