Distributed photovoltaic power station intelligent monitoring method and system

Abstract

The invention relates to the technical field of new energy intelligent decentralized control, and discloses a distributed photovoltaic power station intelligent monitoring method and system, and the method comprises the steps: collecting electrical state parameters, and carrying out the normalization processing; performing anomaly detection and island detection based on a multi-dimensional deviation fusion mechanism; generating a power support parameter through a self-adaptive reactive power injection mechanism; under the island mode, loading voltage frequency double-droop combined control; and after the main network is recovered, grid-connected reconnection is realized based on a phase progressive phase locking mechanism. Compared with the prior art that multi-dependence centralized master station judgment and scheduling cannot realize fault isolation and independent operation in time especially under the condition that main network communication is abnormal or frequency fluctuation occurs in a far-end island edge node, the technical problem that fault isolation and independent operation cannot be realized in time in the prior art is solved. Autonomous monitoring of an individual power station level and stable operation of an island are realized, and the stability of the system in a weak network or island operation scene is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of new energy intelligent distributed control technology, and particularly relates to a distributed photovoltaic power station intelligent monitoring method and system. BACKGROUND

[0002] At present, as an important part of new renewable energy, distributed photovoltaic power stations have been widely deployed in industrial parks, town roofs and remote areas and other types of scenes. The traditional monitoring architecture usually relies on a centralized master station for unified scheduling and state identification. The master station collects the operating parameters of all sub-stations and centrally processes them, and then issues operating control instructions to each power station. However, under the conditions of multi-station grid connection or large-scale deployment, the centralized master control has the following problems: on the one hand, the instability of the communication link and the bandwidth bottleneck can easily lead to state reporting delay, and thus cause fault identification lag. For example, when a power station has power fluctuations, frequency deviation or inverter abnormalities, the master station may not be able to receive and process the relevant abnormal information in time due to delay or network congestion, leading to problems such as fault propagation and local grid voltage instability. On the other hand, under the conditions of inaccessible master control platform, abnormal master grid frequency or island state triggering, the existing system lacks on-site sensing and response capability, and cannot realize basic operating guarantee mechanisms such as local autonomous decision-making, fast reactive power injection and Droop regulation, which may eventually lead to power station disconnection, system reclosing failure, and affect photovoltaic output efficiency and power quality. In addition, although some existing solutions attempt to introduce edge computing and local strategy issuance, their control logic still relies on master station authorization or periodic synchronization, and cannot realize truly distributed autonomous response and control loop in emergency situations, especially in unattended or remote island deployment scenarios.

[0003] Therefore, there is an urgent need for a distributed photovoltaic power station intelligent monitoring method and system that can still automatically complete operating state identification, reactive power regulation, island switching and grid reconnection control based on distributed control mechanism under the conditions of abnormal master control system, communication link or serious grid fluctuation, so as to enhance dynamic autonomy and adapt to the future development needs of distributed new energy power grids. SUMMARY

[0004] In view of the above technical deficiencies, the purpose of the present application is to provide a distributed photovoltaic power station intelligent monitoring method, which aims to solve the technical problem that the existing technology relies too much on centralized master station for judgment and scheduling, and especially under the conditions of abnormal master grid communication or frequency fluctuation of remote island edge nodes, cannot realize fault isolation and independent operation in time.

[0005] To solve the above technical problems, the present application adopts the following technical scheme: the present application provides a distributed photovoltaic power station intelligent monitoring method, The distributed photovoltaic power station intelligent monitoring method comprises: Step S10: Collecting electrical state parameters at time t in real time at each photovoltaic power station i through local sensors, normalizing the electrical state parameters by using an improved local fluctuation balanced normalization method, and outputting an electrical state multi-source data set; Step S20: Performing power prediction and residual analysis, voltage and frequency deviation detection, and island detection by using a multi-dimensional deviation fusion detection mechanism based on prediction residual driving according to the electrical state multi-source data set, and outputting a comprehensive abnormality detection score and an island determination index; Step S30: When the comprehensive abnormality detection score is within a preset abnormality detection score range or the island determination index is within a preset island determination index range, determining a current target reactive power adjustment amount by using an impedance feedback driven adaptive reactive power injection regulation mechanism, and outputting an adaptive power support parameter set U; Step S40: Obtaining a coupling factor representing the state of the main grid from the adaptive power support parameter set U, and starting an island operation mode of the photovoltaic power station i when the coupling factor is lower than a preset decoupling threshold, and loading a voltage and frequency double droop joint control mechanism based on the current electrical state multi-source data set to realize local power self-balancing; Step S50: Continuously monitoring the voltage and frequency state of the photovoltaic power station i and the main grid frequency state of the distributed main grid, and updating the output state of the grid-connected flag bit based on the voltage and frequency state and the main grid frequency state.

[0006] Preferably, in step S10, the electrical state parameters include port voltage , current , frequency , reactive power , active power , ambient temperature , illumination , and energy storage state ; the improved local fluctuation balanced normalization method introduces an adaptive window method based on local fluctuation, and dynamically adjusts the mean and standard deviation calculation interval in the normalization process based on local data fluctuation in the power station; wherein the local data fluctuation in the power station includes power fluctuation and illumination change.

[0007] Preferably, in step S20, the steps of power prediction and residual analysis and voltage and frequency deviation detection specifically include: Power prediction and residual analysis: obtaining historical power data from the electrical state multi-source data set, predicting the power output in the next time period by using a time series prediction network LSTM, obtaining predicted power, comparing the difference between actual power and predicted power, and calculating power residual based on a prediction error adaptive reconstruction method, wherein, Perror,i(t) is the power error of the ith photovoltaic power station at time t, which is used to represent the difference between the actual power and the predicted power; Pactual,i(t) is the actual power of the ith photovoltaic power station at time t; Ppredicted,i(t) is the predicted power of the ith photovoltaic power station at time t; is an adaptive adjustment coefficient, which is used to adjust the influence degree of the change rate of the prediction error on the power error; when the power error is greater than a preset power error threshold, it is determined that the power output is abnormal; Voltage and frequency deviation detection: the voltage deviation is calculated based on the symmetric relative deviation rate determination method, wherein, is the voltage deviation of the ith photovoltaic power station at time t, is a preset microgrid reference voltage; is the port voltage value measured by the ith photovoltaic power station at time t; β is a voltage change weight adjustment coefficient; Frequency deviation is calculated based on the perturbation stability domain deviation method, wherein, is the frequency deviation of the ith photovoltaic power station at time t; is the frequency value measured by the ith photovoltaic power station at time t; is the nominal frequency of the power grid; is a frequency fluctuation sensitivity adjustment coefficient, which is used to adjust the response degree of the frequency rapid fluctuation.

[0008] Preferably, in step S20, the step of island detection specifically comprises: identifying whether the power grid and the photovoltaic power station are off-grid through phase disturbance, and defining the island determination index of the ith photovoltaic power station at time t based on the weighted sum of voltage change, frequency change and voltage phase angle change , wherein, , and are weight coefficients of respective parameters, which are used to adjust the contribution degree of voltage change, frequency change and phase change to the island determination index according to the normalization of the corresponding parameter dimension; and are the voltage phase angles of the ith power station at time t and the previous time t−1, respectively.

[0009] Preferably, in step S30, the adaptive power support parameter set U includes target reactive power set value, voltage limiting parameter, reactive power regulation response rate, maximum power support capability and coupling factor representing the state of the main grid.

[0010] Preferably, in step S40, the step of implementing local power self-balancing based on the current electrical state multi-source data set to load the voltage and frequency double droop joint control mechanism, specifically comprises: According to the current electrical state multi-source data set and the adaptive power support parameter set U generated in step S30, corresponding target reference frequency values and target reference voltage values are generated and used as dynamic set points of the frequency control chain and the voltage control chain; A current real-time frequency value is obtained, a frequency control deviation is constructed based on the difference between the current real-time frequency value and the target reference frequency value, a frequency control coefficient is dynamically set according to the frequency control deviation, the active power output of the photovoltaic inverter is adjusted, and a first droop control chain is constructed; A current real-time voltage value is obtained, a voltage control deviation is constructed based on the difference between the current real-time voltage value and the target reference voltage value, a voltage control coefficient is dynamically set according to the voltage control deviation, the reactive power output of the photovoltaic inverter is adjusted, and a second droop control chain is constructed; A power error index is constructed based on the current reactive power output, the active power output and the load power demand, and the control coefficients in the first droop control chain and the second droop control chain are corrected in a dynamic correction mode based on error feedback, the closed-loop adjustment of the dual-droop joint control mechanism is completed, and local power self-balancing is realized.

[0011] Preferably, in step S50, the voltage frequency state of the photovoltaic power station i and the main grid frequency state of the distributed main grid are continuously monitored, and the step of updating the output state of the grid connection flag based on the voltage frequency state and the main grid frequency state specifically comprises: continuously monitoring the voltage frequency state of the photovoltaic power station i and the main grid frequency state of the distributed main grid, and when it is detected that the main grid frequency returns to the reference frequency range and the phase difference between the voltage frequency state and the main grid frequency state meets the preset gradual synchronization condition, a phase gradual phase-locked reconnection mechanism is used to update the output state of the grid connection flag.

[0012] The application also provides a distributed photovoltaic power station intelligent monitoring system, which comprises: A state normalization processing module is configured to collect electrical state parameters at time t in real time at each photovoltaic power station i through a local sensor, normalize the electrical state parameters by using an improved local fluctuation balancing normalization method, and output an electrical state multi-source data set; An abnormality detection and island determination module is configured to perform power prediction and residual error analysis, voltage and frequency deviation detection and island detection by using a multi-dimensional deviation fusion detection mechanism based on predicted residual error driving based on the electrical state multi-source data set, and output a comprehensive abnormality detection score and an island determination index; A reactive power regulation parameter generation module is configured to determine a current target reactive power adjustment amount by using an impedance feedback driven adaptive reactive power injection regulation mechanism when the comprehensive abnormality detection score is within a preset abnormality detection score range or the island determination index is within a preset island determination index range, and output an adaptive power support parameter set U. An island control activation module is configured to obtain a coupling factor representing a state of a main grid from an adaptive power support parameter set U, and when the coupling factor is lower than a preset decoupling threshold, an island operation mode of the photovoltaic power station i is started, and a voltage-frequency double droop joint control mechanism is loaded based on a current electrical state multi-source data set to realize local power self-balancing. A grid synchronization monitoring module is configured to continuously monitor a voltage-frequency state of the photovoltaic power station i and a main grid frequency state of the distributed main grid, and update an output state of a grid connection flag based on the voltage-frequency state and the main grid frequency state.

[0013] The application further provides a distributed photovoltaic power station intelligent monitoring device, which comprises a memory, a processor, and a distributed photovoltaic power station intelligent monitoring program stored in the memory and capable of running on the processor.

[0014] The application further provides a computer program product comprising a distributed photovoltaic power station intelligent monitoring program, which, when executed by a processor, implements the distributed photovoltaic power station intelligent monitoring method.

[0015] The application has the advantages that: the application adopts a decentralized control mechanism, so that each photovoltaic power station can independently decide and adjust an operation strategy, and in the absence of a central control, the stability of the power grid and the efficient operation of the power station are ensured, and the single-point failure risk in the traditional centralized control mode is avoided.

[0016] The application combines decentralized control and multi-source data-driven real-time state perception, can timely switch to an island mode and realize self-balancing when the main grid is abnormal, ensures the reliability and stability of the power station under complex environmental conditions, and significantly improves the self-healing ability of power under local disturbance. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description only constitute some embodiments of the application, and for those skilled in the art, other drawings can also be obtained without creative labor based on these drawings.

[0018] Figure 1 It is a flowchart of the first embodiment of the distributed photovoltaic power station intelligent monitoring method of the application.

[0019] Figure 2 It is a voltage fluctuation comparison diagram under different control conditions of the first embodiment of the distributed photovoltaic power station intelligent monitoring method of the application.

[0020] Figure 3 This is a schematic diagram of the equipment for a distributed photovoltaic power station intelligent monitoring method according to the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Example 1: As Figure 1 The diagram shown is a flowchart of the first embodiment of the intelligent monitoring method for distributed photovoltaic power stations of the present invention, which presents the first embodiment of the intelligent monitoring method for distributed photovoltaic power stations of the present invention.

[0023] In the first embodiment, the intelligent monitoring method for distributed photovoltaic power stations includes: Step S10: In each photovoltaic power station i, the electrical state parameters at time t are collected in real time by local sensors. The electrical state parameters are normalized by an improved local fluctuation equilibrium normalization method, and the multi-source electrical state dataset is output. It's important to note that the "improved local fluctuation equilibrium normalization method" is not simply a scaling operation on multidimensional physical quantities. Instead, it incorporates the disturbance patterns and frequency of change of electrical quantities within the actual operating environment of a photovoltaic power plant, constructing a "disturbance-aware normalization strategy." Specifically, in traditional methods, all parameters are typically linearly mapped to their full-cycle extreme values. However, this approach ignores the dynamic nonlinear characteristics of photovoltaic power generation, such as high-frequency disturbances caused by sudden changes in sunlight intensity, cloud cover, or rapid load switching. The aforementioned method introduces a short-time sliding window to capture the current parameter change trend and uses local fluctuations as parameter adjustment factors to dynamically expand or compress the normalization interval, allowing various electrical parameters to better retain their relative change characteristics in the short term. The purpose of normalization is expanded from "stretching to a standard scale" to "maximizing the preservation of fluctuation trends without distorting the original energy characteristics."

[0024] Understandably, this normalization method can still extract the inherent correlation trend between parameters even when faced with rapid fluctuations in electrical state parameters (such as sudden voltage increases or frequency shifts), preventing the normalization operation itself from masking changes in key signals. In other words, the normalization result is not only used for scale unification but also carries characteristic information such as the rate of change and fluctuation frequency. This ensures that the subsequent anomaly detection mechanism receives "true perceived expression" rather than "compressed and harmonicized" parameters, thereby improving the accuracy and timeliness of the overall perception and decision-making of the intelligent monitoring system.

[0025] It should be understood that, compared with the traditional global extreme difference normalization method, the present application emphasizes the "adaptability" and "fluctuation sensitivity" of normalization. The traditional method will fail under extreme operating conditions, for example, the sampling power value is close to zero under extremely low light conditions at night, and if the maximum power of the whole day is still used as the normalization reference, the processing result will make the local fluctuation be submerged and cannot identify the actual small fault precursor. In the present method, the normalization interval is adjusted in real time based on the local window, which can automatically adapt to different data distribution characteristics in different operating stages, so that the feature pattern can be stably extracted even in low light operation or energy storage access fluctuation stage, thereby greatly improving the expression ability and applicability of normalization processing in abnormal operating scenarios.

[0026] Step S20: power prediction and residual analysis, voltage and frequency deviation detection and island detection are performed according to the electrical state multi-source data set using a multi-dimensional deviation fusion detection mechanism based on prediction residual driving, and a comprehensive abnormal detection score and island determination index are output; It should be noted that the "multi-dimensional deviation fusion detection mechanism based on prediction residual driving" means that the active power is short-term predicted in a time series modeling manner, the dynamic residual between the actual value and the predicted value is extracted, the real-time voltage and frequency change trend is continuously tracked, and the time sequence characteristics of these deviation signals are fused to construct a unified abnormal score model and island determination rule. Specifically, "power residual" is used to identify significant deviations between power generation capacity and model expectations; "voltage deviation" reflects power quality fluctuation amplitude; and "frequency deviation" is sensitive to grid-side power imbalance and decoupling trend. Through the multi-channel input structure, it is uniformly input into the fusion layer to comprehensively judge whether the current power station is in an unstable state or a decoupling risk rising state. This mechanism not only focuses on the change of a single feature, but also emphasizes the interactive consistency and deviation trend coordination expression between multi-source data, so as to realize more accurate abnormal identification.

[0027] It can be understood that the technical effect of the fusion detection mechanism is that it can use "power prediction error" as a guide signal to guide the subsequent detection process. When a certain type of abnormality occurs (such as cloud shadow blocking or energy storage interference), the prediction model error will first increase, and then the stability indicators in the voltage and frequency channels will deteriorate. Therefore, through the residual driving mode, the response can be awakened in advance, forming a time foresight warning capability for abnormal development.

[0028] It should be understood that, compared with the traditional single-threshold triggered voltage / frequency detection mechanism, the fusion model of the application determines the abnormal score based on the consistency of the trend between multiple channels. The traditional method is prone to false negatives or misjudgments in voltage sudden change or transient frequency jitter scenarios, especially in non-fault fluctuation scenarios (such as load disturbance). The method introduces a dynamic indicator, "power prediction residual", which can be used as an abnormal trigger factor to guide the scoring mechanism. Its fusion score is not a static logical judgment, but a dynamic identification method for trend, structure and residual collaborative mode, which adapts to the self-organizing abnormal judgment logic under the islanding trend, and significantly enhances the local autonomy of distributed control.

[0029] Step S30: When the integrated abnormality detection score is within the preset abnormality detection score range or the islanding determination index is within the preset islanding determination index range, determine the current target reactive power adjustment amount by the impedance feedback driven adaptive reactive power injection regulation mechanism, and output the adaptive power support parameter set U; It should be noted that the "impedance feedback driven adaptive reactive power injection regulation mechanism" means that when the operating state abnormality or the obvious decoupling trend of the power grid is detected, the photovoltaic power station i does not depend on the master control center, but dynamically estimates the required reactive power support capability according to the equivalent impedance change trend of the local power grid, and combines the current voltage margin and frequency stability to autonomously calculate the optimal reactive power adjustment amount. This mechanism takes "local perception-adaptive response" as the core logic, measures the changes of equivalent resistance and equivalent reactance through continuous feedback, and adjusts the reactive power support strategy in real time, thereby forming a "power support parameter set U" containing target reactive power, injection direction, voltage control sensitivity and other parameters.

[0030] It can be understood that the introduction of the impedance feedback mechanism enables the photovoltaic power station to infer the equivalent impedance from the measured voltage disturbance response, so as to perceive the risk level of the main grid strength, load mutation or local decoupling. This local inference mechanism without the support of full network communication has the ability of "fault area self-identification and local flexible support", which effectively supports power quality restoration during voltage sag and dynamic voltage pull-up control during main grid voltage undershoot. The adaptive injection strategy can also dynamically adjust the injection rate and target value to avoid oscillation or reactive power backflow caused by overcompensation.

[0031] It should be understood that the traditional centralized reactive power control scheme needs to rely on the SCADA system or the remote master strategy to push the control instruction, which cannot respond in time in the communication link interruption or island scene, and is easy to cause the expansion of reactive power shortage, rapid deterioration of voltage and other problems. The application adopts impedance feedback logic to establish a local adaptive loop, which is completely executed by the distributed photovoltaic power station in a closed loop, avoiding the delay and failure risk caused by relying on central control. Compared with the reactive power control based on fixed voltage margin, the application scheme can realize more fine-grained and more scene-matched dynamic adjustment, effectively improving the voltage stability and the support ability of the grid side.

[0032] Step S40: obtaining a coupling factor representing the main grid state from the adaptive power support parameter set U, when the coupling factor is lower than a preset decoupling threshold, starting the island operation mode of the photovoltaic power station i, and loading a voltage frequency double droop joint control mechanism based on the current electrical state multi-source data set to realize local power self-balancing; It should be noted that the "coupling factor" is a quantitative index for representing the dynamic electrical connection strength between the photovoltaic power station i and the upper main grid, which is calculated according to the phase correlation between the photovoltaic power station output power change and the bus voltage frequency disturbance, and its core reflects whether the local power regulation can effectively affect the voltage and frequency response. The low coupling factor indicates that the main grid responsiveness is weak, and there may be physical disconnection, impedance distortion, and serious bus voltage sag, etc., triggering the necessity of islanding. The "double droop joint control mechanism" refers to: respectively adopting P-V droop control law in the voltage regulation channel and Q-f droop law in the frequency channel, and maintaining local power balance and stable operation through joint regulation, which is suitable for island control scene in the decoupled state of the grid.

[0033] It can be understood that the coupling factor is equivalent to a "signaling signal of electrical connectivity", which has dynamic and real-time characteristics, and is estimated by tracking the sensitivity of voltage frequency response to reactive power and active power, and is an adaptive connection quality perception index. This mechanism does not need to rely on external communication channels, but can construct the coupling estimation quantity through local sampling, greatly improving the real-time and robustness of decoupling judgment. The introduction of double droop control means that the photovoltaic power station can independently control its output voltage and frequency after being disconnected from the main grid, and can quickly compensate the response to local load changes, realizing dynamic stable operation in the "quasi-microgrid state".

[0034] It should be understood that, compared to traditional solutions that directly disconnect after islanding detection or passively maintain operation using a single droop mode, this invention introduces a main grid coupling factor to construct a priori judgment mechanism. This allows for early intervention in control preparation before signs of coupling degradation appear, improving the predictability and flexibility of switching islanding modes. Simultaneously, the dual droop joint control, while actively supporting the system, balances voltage stability and frequency consistency, overcoming the problems of weak regulation capability and slow dynamic tracking in single droop mode. This achieves steady-state control and dynamic equilibrium under islanding conditions with multiple load types and concurrent disturbances at multiple time scales.

[0035] For example, such as Figure 2 As shown, the voltage fluctuation performance of the traditional single droop control mechanism and the dual droop joint control mechanism described in this invention were tested in islanded mode for ten typical test conditions numbered T1 to T10. In each test, the voltage stability response of each photovoltaic power station after external disturbances (such as load changes or grid disconnection) were recorded. As can be seen from the figure, the voltage fluctuation amplitude under the traditional single droop mechanism is generally between 5.2% and 8.1%, with large fluctuations and slow response, which can easily lead to control mis-triggering and misjudgment by the energy storage system. However, after adopting the voltage and frequency dual droop joint control mechanism proposed in this invention, the voltage fluctuation amplitude is significantly reduced, and the overall control is between 3.1% and 4.6%, with a fluctuation amplitude reduction of more than 30%. This shows that the solution of this invention can effectively improve the stability of islanded operation and the robustness of voltage regulation, enhance the ability to suppress disturbances, and is particularly suitable for distributed photovoltaic systems with autonomous response requirements, achieving high-performance decentralized control without relying on central coordination.

[0036] Step S50: Continuously monitor the voltage and frequency status of photovoltaic power station i and the main grid frequency status of the distributed main grid, and update the output status of the grid connection flag bit based on the voltage and frequency status and the main grid frequency status.

[0037] It should be noted that the "voltage frequency status" refers to the effective value of the AC output voltage, frequency value, and rate of change collected by local sensors at photovoltaic power station i, while the "grid frequency status" is obtained from reference frequency data from nearby public access nodes via the communication module. Updating the "grid connection flag" means setting a status register bit indicating "whether grid connection conditions are met" based on the determination results of the current frequency difference and phase difference, which is used to drive the execution of subsequent grid connection commands.

[0038] It can be understood that, by continuously comparing the local output state with the main network synchronization state, the flexible transition from the island mode to the grid-connected mode is realized, the grid-connected impact caused by frequency jump and phase inconsistency is effectively avoided, and the undisturbed reconnection ability of the distributed photovoltaic power station is ensured. The mechanism has adaptive reconnection judgment ability, rather than passive waiting for superior dispatching instructions, thereby improving the autonomy in the decentralized control scenario.

[0039] It should be understood that, compared with the reconnection mechanism of "timed detection + manual reset + fixed threshold judgment" in the traditional scheme, the present application introduces a logic based on master-slave phase progressive locking and frequency adaptive return judgment. When the frequency difference and the phase difference simultaneously satisfy the preset synchronization condition, the grid-connected flag is automatically set, the slow coincidence at a lower synchronization rate is supported, the reconnection process has progressive and high fault tolerance, and thus it is suitable for actual deployment environments with large voltage disturbance, unstable network communication or delayed dispatching response.

[0040] Embodiment two: In addition, the present application provides a distributed photovoltaic power station intelligent monitoring system which adopts the distributed photovoltaic power station intelligent monitoring method in the above embodiment, and can solve the technical problem of a distributed photovoltaic power station intelligent monitoring. Compared with the prior art, the beneficial effects of the distributed photovoltaic power station intelligent monitoring system provided by the present application are the same as those of the distributed photovoltaic power station intelligent monitoring method provided by the above embodiment, and other technical features in the distributed photovoltaic power station intelligent monitoring system are the same as those disclosed in the above embodiment method, and thus will not be repeated here.

[0041] Embodiment three: The present application provides a distributed photovoltaic power station intelligent monitoring device, please refer to Figure 3A distributed photovoltaic power station intelligent monitoring device includes at least one processor, and a memory connected to the at least one processor in communication; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the distributed photovoltaic power station intelligent monitoring method in the above-mentioned embodiment one. The distributed photovoltaic power station intelligent monitoring device in the embodiment of the present application can include, but is not limited to, mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (Personal Digital Assistant), PADs (Portable Application Description), PMPs (Portable Media Player), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), and the like, and fixed terminals such as digital TVs, desktop computers, and the like. A distributed photovoltaic power station intelligent monitoring device is only an example, and should not bring any limitation to the function and use range of the embodiment of the present application. A distributed photovoltaic power station intelligent monitoring device can include a processing device 1001 (such as a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to programs stored in a read-only memory 1002 or loaded from a storage device 1003 to a random access memory 1004. In the random access memory 1004, various programs and data required for the operation of a distributed photovoltaic power station intelligent monitoring device are also stored. The processing device 1001, the read-only memory 1002, and the random access memory 1004 are connected to each other through a bus 1005. An I / O interface 1006 is also connected to the bus. Generally, the following systems can be connected to the I / O interface 1006: input devices 1007 including, for example, touch screens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, and the like; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, and the like; storage devices 1003 including, for example, magnetic tapes, hard disks, and the like; and communication devices 1009. The communication device 1009 can allow a distributed photovoltaic power station intelligent monitoring device to communicate with other devices wirelessly or by wire to exchange data. Although a distributed photovoltaic power station intelligent monitoring device with various systems is shown in the figure, it should be understood that all the shown systems are not required to be implemented or possessed. More or fewer systems can be alternatively implemented or possessed.

[0042] Embodiment four: the application further provides a computer program product comprising a computer program which, when executed by a processor, implements the steps of a distributed photovoltaic power station intelligent monitoring method as described above. The computer program product provided by the application can solve the technical problem of a distributed photovoltaic power station intelligent monitoring. Compared with the prior art, the beneficial effects of the computer program product provided by the application are the same as those of the distributed photovoltaic power station intelligent monitoring method provided by the above-described embodiments, and are not described here in detail.

[0043] In particular, according to the embodiments disclosed by the application, the processes described above with reference to the flowcharts can be implemented as a computer software program. For example, the embodiments disclosed by the application include a computer program product comprising a computer program carried on a computer readable medium, the computer program comprising program code for executing the method shown in the flowchart. In such embodiments, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a read-only memory 1002. When the computer program is executed by a processing device 1001, the above-mentioned functions defined in the method of the embodiments disclosed by the application are executed.

[0044] It should be understood that various parts of the application disclosed can be realized in hardware, software, firmware or a combination thereof. In the description of the above-described embodiments, specific features, structures, materials or characteristics can be combined in any appropriate manner in any one or more embodiments or examples.

[0045] Obviously, those skilled in the art can make various modifications and variations to the application without departing from the spirit and scope of the application. Thus, if these modifications and variations of the application fall within the scope of the claims of the application and their equivalent technologies, the application also intends to include these modifications and variations.

Claims

1. A distributed photovoltaic power station intelligent monitoring method, characterized in that, The method comprises: Step S10: collecting electrical state parameters at time t in real time at each photovoltaic power station i through local sensors, normalizing the electrical state parameters by using an improved local fluctuation equalization normalization method, and outputting an electrical state multi-source data set; Step S20: performing power prediction and residual analysis, voltage and frequency deviation detection, and island detection by using a multi-dimensional deviation fusion detection mechanism based on a prediction residual driving mechanism according to the electrical state multi-source data set, and outputting a comprehensive abnormality detection score and an island determination index; Step S30: when the comprehensive abnormality detection score is within a preset abnormality detection score range or the island determination index is within a preset island determination index range, determining a current target reactive power adjustment amount by using an impedance feedback driven adaptive reactive power injection regulation mechanism, and outputting an adaptive power support parameter set U; Step S40: obtaining a coupling factor representing a main grid state from the adaptive power support parameter set U, and starting an island operation mode of the photovoltaic power station i when the coupling factor is lower than a preset decoupling threshold, and loading a voltage-frequency double droop joint control mechanism based on the current electrical state multi-source data set to realize local power self-balancing; Step S50: continuously monitoring a voltage-frequency state of the photovoltaic power station i and a main grid frequency state of a distributed main grid, and updating an output state of a grid-connected flag based on the voltage-frequency state and the main grid frequency state.

2. The method of claim 1, wherein the method further comprises: In step S10, the electrical state parameters include port voltage , current , frequency , reactive power , active power , ambient temperature , illumination and energy storage state ; the improved local fluctuation equalization normalization method introduces a local fluctuation-based adaptive window method, which dynamically adjusts the mean and standard deviation calculation interval in the normalization process based on local data fluctuations in the power station; wherein the local data fluctuations in the power station include power fluctuations and illumination changes.

3. The method of claim 1, wherein the method further comprises: In step S20, the steps of power prediction and residual analysis and voltage and frequency deviation detection specifically comprise: Power prediction and residual analysis: obtain historical power data from the electrical state multi-source data set, predict the power output of the next time period through the time series prediction network LSTM to obtain the predicted power, compare the difference between the actual power and the predicted power, and calculate the power residual based on the prediction error adaptive reconstruction method, wherein, is the power residual of the ith power station at time t, which is used to represent the difference between the actual power and the predicted power; is the actual power of the ith power station at time t; is the predicted power of the ith power station at time t; is an adaptive adjustment coefficient, which is used to adjust the influence degree of the change rate of the prediction error on the power residual; when the power residual is greater than the preset power residual threshold, it is determined that the power output is abnormal; Voltage and frequency deviation detection: voltage deviation is calculated based on the symmetrical relative shift rate determination method, wherein, is the voltage deviation of the ith photovoltaic power station at time t, is a preset microgrid reference voltage; is the measured port voltage value of the ith photovoltaic power station at time t; and β is a voltage change weight adjustment coefficient. The frequency deviation is calculated based on the perturbation stability domain deviation method, wherein, is the frequency deviation of the ith photovoltaic power station at time t; is the frequency value measured at time t by the ith photovoltaic power station; is the nominal frequency of the electrical grid; is the frequency fluctuation sensitivity adjustment coefficient, which is used to adjust the response degree to the frequency rapid fluctuation.

4. The method of claim 3, wherein the method further comprises: In step S20, the step of island detection specifically comprises: identifying whether the power grid and the photovoltaic power station are off-grid through phase disturbance, and defining an island determination index of the ith photovoltaic power station at time t based on a weighted sum of voltage variation, frequency variation and voltage phase angle variation , wherein, , and are weight coefficients of respective parameters, used for adjusting the contribution degree of voltage variation, frequency variation and phase variation to the island determination index according to the normalization of the corresponding parameter item dimension; and are voltage phase angles of the ith power station at time t and the previous time t−1, respectively.

5. The method of claim 1, wherein the method further comprises: In step S30, the adaptive power support parameter set U includes a target reactive power set value, a voltage limiting parameter, a reactive power regulation response rate, a maximum power support capability, and a coupling factor representing a main grid state.

6. The intelligent monitoring method for a distributed photovoltaic power station according to claim 1, characterized in that, In step S40, the step of loading a voltage-frequency double droop joint control mechanism based on the current electrical state multi-source data set to realize local power self-balancing specifically comprises: According to the current electrical state multi-source data set and the adaptive power support parameter set U generated in step S30, corresponding target reference frequency and target reference voltage values are generated and used as dynamic set points of a frequency control chain and a voltage control chain; A current real-time frequency value is obtained, a frequency control deviation is constructed based on the difference between the current real-time frequency value and the target reference frequency value, a frequency control coefficient is dynamically set according to the frequency control deviation, the active power output of the photovoltaic inverter is adjusted, and a first droop control chain is constructed; A current real-time voltage value is obtained, a voltage control deviation is constructed based on the difference between the current real-time voltage value and the target reference voltage value, a voltage control coefficient is dynamically set according to the voltage control deviation, the reactive power output of the photovoltaic inverter is adjusted, and a second droop control chain is constructed; A power error index is constructed based on the current reactive power output, the active power output, and the load power demand, and the control coefficients in the first droop control chain and the second droop control chain are corrected by using an error feedback dynamic correction method, a closed-loop adjustment of the double droop joint control mechanism is completed, and local power self-balancing is realized.

7. The intelligent monitoring method for a distributed photovoltaic power station according to claim 1, characterized in that, In step S50, the voltage frequency state of the photovoltaic power station i and the main grid frequency state of the distributed main grid are continuously monitored, and the output state of the grid connection flag is updated based on the voltage frequency state and the main grid frequency state. The specific steps include: continuously monitoring the voltage frequency state of the photovoltaic power station i and the main grid frequency state of the distributed main grid, and when the main grid frequency is restored to the reference frequency range and the phase difference between the voltage frequency state and the main grid frequency state meets the preset gradual synchronization condition, the phase gradual phase-locked reconnection mechanism is used to update the output state of the grid connection flag.

8. A distributed photovoltaic power station intelligent monitoring system applied to the distributed photovoltaic power station intelligent monitoring method in any one of claims 1 to 7, characterized in that, The distributed photovoltaic power station intelligent monitoring system comprises: A state normalization processing module is configured to collect electrical state parameters of each photovoltaic power station i at a time t in real time through a local sensor, normalize the electrical state parameters by using an improved local fluctuation balancing normalization method, and output an electrical state multi-source data set. An abnormality detection and island determination module is configured to perform power prediction and residual analysis, voltage and frequency deviation detection, and island detection by using a multi-dimensional deviation fusion detection mechanism based on a prediction residual driving mechanism based on the electrical state multi-source data set, and output a comprehensive abnormality detection score and an island determination index. A reactive power regulation parameter generation module is configured to determine a current target reactive power adjustment amount by using an impedance feedback driven adaptive reactive power injection regulation mechanism when the comprehensive abnormality detection score is within a preset abnormality detection score range or the island determination index is within a preset island determination index range, and output an adaptive power support parameter set U. An island control activation module is configured to obtain a coupling factor representing a main grid state from the adaptive power support parameter set U, and start an island operation mode of the photovoltaic power station i when the coupling factor is lower than a preset decoupling threshold, and load a voltage frequency double droop joint control mechanism based on the current electrical state multi-source data set to realize local power self-balancing. A grid synchronization monitoring module is configured to continuously monitor the voltage frequency state of the photovoltaic power station i and the main grid frequency state of the distributed main grid, and update the output state of the grid connection flag based on the voltage frequency state and the main grid frequency state.

9. A distributed photovoltaic power station intelligent monitoring device, characterized in that, The distributed photovoltaic power station intelligent monitoring device comprises a memory, a processor, and a distributed photovoltaic power station intelligent monitoring program stored in the memory and executable on the processor, and the distributed photovoltaic power station intelligent monitoring program is executed by the processor to implement the distributed photovoltaic power station intelligent monitoring method in any one of claims 1 to 7.

10. A computer program product, characterised in that, The computer program product comprises a distributed photovoltaic power station intelligent monitoring program, and the distributed photovoltaic power station intelligent monitoring program is executed by the processor to implement the distributed photovoltaic power station intelligent monitoring method in any one of claims 1 to 7.

Images