Photovoltaic power station DC side fault prediction method and system

Abstract

The invention relates to the related technical field of power station fault prediction, in particular to a photovoltaic power station direct current side fault prediction method and system, and the method comprises the steps: extracting time domain fluctuation features and frequency domain harmonic features; performing cooperative control optimization on the filter and the reactive power compensation device; and combining the local electrical characteristics and the global electrical characteristics to obtain fusion characteristic data, performing comparative analysis on the fusion characteristic data and normal operation characteristic data, evaluating a fault risk index and performing fault risk reminding. The technical problems that photovoltaic power station direct current side fault prediction is mostly limited to a single fault type, cross-side fault recognition of intermediate equipment lags behind, and the photovoltaic power station fault prevention requirement cannot be met are solved, fusion feature data are obtained through feature selection downsampling, deep fusion of local and global features is carried out, and the fault prevention requirement of the photovoltaic power station is met. According to the protection parameters of the DC circuit breaker and the AC circuit breaker, cooperative control optimization is carried out on the filter and the reactive power compensation device, a cross-side fault feature linkage mechanism is established, and the effectiveness of fault prevention is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power station fault prediction, and in particular to a photovoltaic power station DC side fault prediction method and system. BACKGROUND

[0002] As a key link of energy conversion and transmission of photovoltaic power stations, the DC side connects photovoltaic arrays and core devices such as inverters, and its faults (such as leakage, overcurrent, harmonic distortion, etc.) may cause device damage, power generation drop, even fire accidents and other safety accidents. With the development of large-scale and centralized photovoltaic power stations, the DC side circuit structure is becoming more and more complex, and the device operating environment is changing, which puts higher requirements on early prediction and accurate prevention and control of faults.

[0003] Current photovoltaic power station DC side fault prediction is mostly limited to a single fault type (such as only monitoring overcurrent or leakage), and the identification of cross-side faults of intermediate devices such as inverters and rectifiers lags behind. In addition, conventional fault prediction mostly relies on fixed threshold judgment and simple feature extraction, which is difficult to capture hidden fault signs such as harmonic distortion and voltage oscillation, and cannot meet the timeliness requirements of power station fault prevention.

[0004] In summary, the existing technology has the technical problem that photovoltaic power station DC side fault prediction is mostly limited to a single fault type, and the identification of cross-side faults of intermediate devices lags behind, which cannot meet the requirements of photovoltaic power station fault prevention. SUMMARY

[0005] The present application provides a photovoltaic power station DC side fault prediction method and system, which aims to solve the technical problem that photovoltaic power station DC side fault prediction in the prior art is mostly limited to a single fault type, and the identification of cross-side faults of intermediate devices lags behind, which cannot meet the requirements of photovoltaic power station fault prevention.

[0006] In view of the above problems, the technical scheme of the present application is as follows: The first aspect of the present application provides a photovoltaic power station DC side fault prediction method, wherein the method comprises: performing feature processing on DC side electrical parameters, extracting time domain fluctuation features and frequency domain harmonic features; based on a first one-way fault feature mapping set corresponding to DC-to-AC conversion and a second one-way fault feature mapping set corresponding to AC-to-DC conversion, combining the leakage action value of the DC circuit breaker and the overcurrent protection threshold of the AC circuit breaker, and performing collaborative control optimization on the filter and the reactive power compensation device; at the same time, using the time domain fluctuation features and the frequency domain harmonic features as local electrical features, using the device state features corresponding to the intermediate device as global electrical features, combining the features, and using feature selection downsampling to obtain fusion feature data; comparing and analyzing the fusion feature data with normal operation feature data in a data storage unit, evaluating a fault risk index and giving a fault risk warning.

[0007] Preferably, based on the photovoltaic array string, the DC side raw data including string voltage, branch current, insulation resistance is obtained; the DC side raw data is time sequence aligned to form the DC side electrical parameters including string voltage sequence, branch current sequence and insulation resistance sequence.

[0008] Preferably, according to the inverter, DC circuit breaker and AC circuit breaker in the intermediate equipment between the DC side of the photovoltaic power station and the AC side of the photovoltaic power station, a first one-way fault feature mapping set is set; the tripping coil current of the DC circuit breaker is associated with the circuit breaking fault.

[0009] Preferably, the inverter output current harmonic distortion rate, the DC circuit breaker tripping coil current mutation and the AC circuit breaker contact temperature anomaly are taken as the first fault feature; after the first fault feature is triggered, the filter corresponding first cooperative control action and the second cooperative control action corresponding to the reactive power compensation device are set; the rolling optimization is performed according to the first cooperative control action and the second cooperative control action to balance the filtering effect and the reactive power compensation efficiency.

[0010] Preferably, after the first fault feature is triggered, the main harmonic frequency is identified and the resonant frequency of the filter is adjusted as the first cooperative control action; based on the DC bus voltage fluctuation amplitude, the compensation capacity of the reactive power compensation device is dynamically adjusted as the second cooperative control action.

[0011] Preferably, according to the rectifier, DC circuit breaker and AC circuit breaker in the intermediate equipment between the DC side of the photovoltaic power station and the AC side of the photovoltaic power station, a second one-way fault feature mapping set is set; the action time of the AC circuit breaker is associated with the reverse flow fault.

[0012] Preferably, the rectifier input current reverse flow, the AC circuit breaker action time delay and the DC bus capacitor voltage oscillation are taken as the second fault feature; after the second fault feature is triggered, the filter corresponding third cooperative control action and the fourth cooperative control action corresponding to the reactive power compensation device are set; according to the third cooperative control action and the fourth cooperative control action, the real-time linkage of the filter and the reactive power compensation device is performed using the industrial Ethernet.

[0013] Preferably, after the first fault feature is triggered, the filter is switched to the reverse flow suppression mode to block the reverse current path as the third cooperative control action; according to the DC bus capacitor voltage oscillation, the reactive power compensation device is switched from the constant voltage mode to the damping mode for oscillation suppression as the fourth cooperative control action.

[0014] Preferably, the filter adopts an LC filter topology, and the filter parameters are adjusted according to the U-th harmonic content and V-th harmonic content in the frequency domain harmonic characteristics; the compensation capacity of the reactive power compensation device is dynamically configured according to the power factor target value, and a pulse synchronization mechanism is used in conjunction with the filter parameters to perform coordinated control of the filter and the reactive power compensation device.

[0015] In a second aspect, this application provides a DC-side fault prediction system for a photovoltaic power station, comprising: a feature processing module for performing feature processing on DC-side electrical parameters to extract time-domain fluctuation features and frequency-domain harmonic features; a collaborative control module for optimizing the collaborative control of filters and reactive power compensation devices based on a first unidirectional fault feature mapping set corresponding to DC-to-AC conversion and a second unidirectional fault feature mapping set corresponding to AC-to-DC conversion, combined with the leakage current action value of the DC circuit breaker and the overcurrent protection threshold of the AC circuit breaker; a feature selection module for simultaneously using the time-domain fluctuation features and frequency-domain harmonic features as local electrical features and the equipment status features corresponding to intermediate equipment as global electrical features, and using feature selection downsampling to obtain fused feature data; and a fault risk warning module for comparing and analyzing the fused feature data with the normal operation feature data in the data storage unit, evaluating the fault risk index, and providing fault risk warnings.

[0016] In summary, one or more technical solutions provided in this application achieve the following technical effects: extracting the time-domain fluctuation characteristics and frequency-domain harmonic characteristics of DC-side electrical parameters as local electrical characteristics, combining them with the equipment status characteristics of intermediate equipment as global electrical characteristics, and obtaining fused feature data through feature selection and downsampling; performing deep fusion of local and global features; optimizing the coordinated control of filters and reactive power compensation devices based on the protection parameters of DC circuit breakers and AC circuit breakers; establishing a cross-side fault feature linkage mechanism; and improving the effectiveness of fault prevention. Attached Figure Description

[0017] Figure 1 This application provides a flowchart illustrating a method for predicting DC-side faults in photovoltaic power plants.

[0018] Figure 2 This application provides a schematic diagram of the structure of a DC-side fault prediction system for a photovoltaic power plant.

[0019] Explanation of reference numerals in the attached diagram: Feature processing module M100, collaborative control module M200, feature selection module M300, fault risk warning module M400. Detailed Implementation

[0020] Example 1: The present application will be described in detail below with reference to the accompanying drawings, as follows... Figure 1 As shown, this application provides a method for predicting DC-side faults in a photovoltaic power plant, wherein the method includes: S1: Characteristic processing of direct current side electrical parameters, extracting time domain fluctuation characteristics and frequency domain harmonic characteristics; S2: Based on the first one-way fault feature mapping set corresponding to the conversion of direct current to alternating current, the second one-way fault feature mapping set corresponding to the conversion of alternating current to direct current, combined with the leakage action value of the direct current circuit breaker and the overcurrent protection threshold of the alternating current circuit breaker, the filter and the reactive power compensation device are cooperatively controlled and optimized.

[0021] Specifically, characteristic processing refers to a series of mathematical transformations and statistical analyses on the collected direct current side electrical parameters to extract key features that can reflect the operating state of the equipment. Time domain fluctuation characteristics refer to the variation characteristics of electrical parameters on the time axis, such as mean, variance, peak, valley, and rate of change. These characteristics can intuitively reflect the stability and volatility of electrical parameters in the time dimension. Frequency domain harmonic characteristics are obtained by converting time domain signals to frequency domain through Fourier transform and other methods, analyzing the amplitude and phase of different frequency components in the signal, and mainly focusing on harmonic content, i.e., the amplitude ratio of integer multiple frequency components higher than the fundamental frequency, such as 5th, 7th, 11th, etc. Harmonic distortion rate, etc., are used to reveal the complex changes of electrical parameters in the frequency dimension, especially the harmonic distortion degree, which is useful for judging the operating state and fault type of power electronic equipment.

[0022] The first one-way fault feature mapping set and the second one-way fault feature mapping set are mapping relationship sets established for fault features in the conversion process of direct current and alternating current. Through a large number of experimental data and fault mode analysis, specific fault types are associated with corresponding electrical feature change patterns, providing a basis for fault identification. For example, in the process of converting direct current to alternating current, the fault mode of the inverter may cause an increase in output current harmonic distortion rate, a sudden change in the tripping coil current of the direct current circuit breaker, etc. The corresponding relationship between these characteristic changes and fault types constitutes the first one-way fault feature mapping set. In the process of converting alternating current to direct current, the fault of the rectifier may cause a delay in the action time of the alternating current circuit breaker, oscillation of the direct current bus capacitor voltage, etc., which correspondingly forms the second one-way fault feature mapping set.

[0023] The synergistic control optimization refers to synchronously and coordinately controlling the filter and the reactive power compensation device according to the leakage operating value of the DC circuit breaker and the overcurrent protection threshold of the AC circuit breaker, so as to optimize the performance of the filter and the reactive power compensation device; the filter is used to reduce the harmonic pollution in the power system and improve the power quality; the reactive power compensation device is used to adjust the reactive power of the system, maintain the voltage stability and improve the power factor of the system; through the synergistic control, the working states of the filter and the reactive power compensation device can be adjusted in time according to the protection parameters of the circuit breaker when the fault occurs or the fault sign appears, so that the filter and the reactive power compensation device can cooperate with each other to better cope with the fault condition and improve the stability and reliability of the system.

[0024] The execution step is: the characteristic processing is performed on the DC side electrical parameters, the time domain fluctuation characteristics and the frequency domain harmonic characteristics are extracted, specifically, the time sequence alignment and the characteristic processing are performed on the DC side original data such as the string voltage and the branch current, the time domain fluctuation characteristics of the string voltage are found, the voltage mean value produces a certain fluctuation when the light intensity changes, if the fluctuation range is about ±5% of the rated voltage, the variance is small when the light is stable, and the variance increases when the environmental temperature changes greatly or there is local shading, the frequency domain harmonic characteristic analysis of the branch current is performed, for example, when the inverter is normally operated, the 5th harmonic content is about 3% of the fundamental wave, the 7th harmonic content is about 2%, when the inverter appears a slight fault, the 5th harmonic content increases to 5%, the 7th harmonic content increases to 3.5%, the harmonic distortion rate increases from 4.2% in the normal time to 6.8%, preferably, the local electrical characteristics can reflect the subtle changes of the DC side electrical parameters, and provide a basis for early fault warning.

[0025] Based on the first one-way fault feature mapping set and the second one-way fault feature mapping set, the filter and the reactive power compensation device are synergistically controlled and optimized in combination with the protection parameters of the circuit breaker, specifically, when the DC circuit breaker tripping coil current mutation is detected, the mutation amount is more than 1.2 times of the normal working time, and the inverter output current harmonic distortion rate is more than 8% (about 5% in normal operation), the first one-way fault feature mapping set is used to judge that the inverter fault may occur, the deep fusion of the local characteristics and the global characteristics is realized, the cross-side fault feature linkage mechanism is established, the problems of lagging in the cross-side fault recognition and disconnection of the protection measures in the traditional method are overcome, the accuracy and the advance of the DC side fault prediction are improved, the fault can be accurately identified in the early stage and effective protection measures can be taken in time, and the risk of equipment damage and power generation loss is reduced.

[0026] S3: At the same time, the time domain fluctuation characteristics and the frequency domain harmonic characteristics are used as the local electrical characteristics, the device state characteristics corresponding to the intermediate device are used as the global electrical characteristics, the fusion feature data is obtained by using the feature selection downsampling; S4: The fusion feature data is compared and analyzed with the normal operation feature data in the data storage unit, the fault risk index is evaluated, and the fault risk warning is performed.

[0027] Specifically, the local electrical features refer to time-domain fluctuation features and frequency-domain harmonic features directly extracted from the DC side electrical parameters, reflecting the variation characteristics of the DC side electrical parameters in time and frequency dimensions, and capable of capturing subtle fluctuations and distortion of the electrical parameters; the global electrical feature combination is a feature set formed by integrating the device state features of intermediate devices (such as inverters, rectifiers, DC circuit breakers, AC circuit breakers, etc.), which covers the operating parameters, fault indication signals and protection action states of the intermediate devices, and reflects the health status and operation trend of the intermediate devices as a whole, providing a global perspective for cross-side fault identification; feature selection downsampling aims to select a feature subset that is most relevant and representative for fault prediction from a large number of local and global electrical features, while reducing data dimension and redundant information, improving the efficiency and accuracy of subsequent data analysis and processing, ensuring that the fused feature data retains key fault feature information while reducing data size and complexity, and helping to improve the performance and real-time performance of the fault prediction model.

[0028] The fused feature data is a comprehensive feature data set formed by combining the selected local electrical features and global electrical features, which integrates the detailed changes of the DC side electrical parameters and the global state information of the intermediate devices, providing a more comprehensive and accurate data basis for fault risk assessment; the normal operation feature data in the data storage unit refers to the feature data of the DC side of the photovoltaic power station under normal operation state, including the normal range, distribution law and change mode of local electrical features and global electrical features, which is obtained by long-term monitoring of the normal operation conditions of the photovoltaic power station; the fault risk index is a quantitative index calculated by comparing and analyzing the fused feature data and the normal operation feature data using a pre-set fault risk assessment model, used to represent the possibility of fault occurrence of the current DC side, the higher the fault risk index, the greater the probability of fault occurrence, and the higher the alert and attention required; the fault risk reminder is based on the evaluation result of the fault risk index, when the fault risk index exceeds the set threshold, the system automatically triggers the corresponding reminder mechanism, timely informs the operation and maintenance personnel to take corresponding inspection and maintenance measures, to prevent the occurrence of faults or reduce the impact of faults.

[0029] The extracted time domain fluctuation features and frequency domain harmonic features are taken as local electrical features, and the corresponding device state features of the intermediate device are collected to form a global electrical feature combination. Specifically, the local electrical features include the time domain fluctuation mean and variance of the string voltage, the frequency domain harmonic content (such as 5th and 7th harmonic ratios) of the branch current, and the change rate of the insulation resistance, etc. The global electrical feature combination includes the efficiency, temperature, and fault alarm times of the inverter, the opening coil current and contact wear degree of the DC circuit breaker, and the action time and overcurrent protection action times of the AC circuit breaker. Through feature selection and down-sampling of these local and global features, the features with the greatest contribution to fault prediction are selected, effectively improving the data processing efficiency.

[0030] The fusion feature data is compared and analyzed with normal operation feature data in the data storage unit. The normal operation feature data is a feature interval determined based on long-term stable operation of the photovoltaic power station. Through comparison and analysis, a fault risk assessment model constructed by a machine learning algorithm (such as support vector machine, random forest, etc.) is used to determine the fault risk index. For example, if the fault risk index is 82 (full score is 100, indicating a very high risk), a fault risk reminder is triggered, and the DC side of the photovoltaic power station is checked and maintained in a timely manner. The deep fusion of local and global features is realized, and the detailed changes of the DC side electrical parameters and the global state information of the intermediate device are fully utilized. Through comparison and analysis with normal operation feature data, the fault risk is accurately assessed, overcoming the defects of traditional methods that rely only on single features and are difficult to capture implicit fault signs, effectively improving the accuracy and advance of fault prediction, providing a strong guarantee for the safe and stable operation of the photovoltaic power station, and reducing the loss of power generation and the risk of equipment damage caused by faults.

[0031] Further, the method described in the present application includes: Based on the photovoltaic array string, DC side raw data including string voltage, branch current, and insulation resistance are obtained. The DC side raw data is time-aligned to form DC side electrical parameters including string voltage sequence, branch current sequence, and insulation resistance sequence.

[0032] Specifically, the photovoltaic array string refers to a power generation unit formed by connecting multiple photovoltaic panels in series, which is the basic component of a photovoltaic power station for photoelectric conversion, and can convert solar energy into direct current energy; the direct current side raw data is the untreated electrical parameter data collected directly from the electrical connection point of the photovoltaic array string, mainly including string voltage, branch current and insulation resistance, etc. These data truly reflect the electrical state of the photovoltaic array string during operation; the string voltage refers to the direct current voltage value between the two ends of the photovoltaic array string, which is closely related to factors such as solar irradiance, environmental temperature and photovoltaic panel performance; the branch current refers to the direct current flowing through the photovoltaic array string, which is also affected by conditions such as light and temperature, and is closely related to the connection method and characteristics of each photovoltaic panel in the string; the insulation resistance refers to the insulation resistance value between the photovoltaic array string and the ground, which is used to evaluate the insulation performance of the photovoltaic system to prevent electric shock accidents.

[0033] Time alignment is a data preprocessing operation aimed at arranging and calibrating the direct current side raw data collected at different time points according to a unified time sequence, ensuring the consistency and synchronization of electrical parameter data in the time dimension, so as to facilitate subsequent data analysis and feature extraction; the direct current side electrical parameters are complete data sets containing string voltage sequence, branch current sequence and insulation resistance sequence after time alignment processing, which record the changes of string voltage, branch current and insulation resistance over time, providing basic data support for subsequent fault feature extraction and analysis.

[0034] Execution steps: based on the photovoltaic array string, use various sensors on the spot of the photovoltaic power station, such as voltage sensors, current sensors and insulation resistance testers, to collect direct current side raw data such as string voltage, branch current and insulation resistance; the raw data is transmitted to the data processing unit through the data acquisition system; the collected direct current side raw data is processed by time alignment, arranging and calibrating the string voltage, branch current and insulation resistance data with different time stamps according to a unified time sequence to form direct current side electrical parameters containing string voltage sequence, branch current sequence and insulation resistance sequence. Further, through time alignment, each parameter data that may have a time deviation is adjusted to a unified time sequence in seconds, so that each time point corresponds to accurate string voltage, branch current and insulation resistance values, forming a neat data sequence, providing accurate and synchronous basic data for subsequent feature processing and fault prediction, ensuring the reliability and effectiveness of fault feature extraction.

[0035] Further, based on the first one-way fault feature mapping set corresponding to the conversion of direct current to alternating current, the method described in the present application includes: According to the inverter, the DC circuit breaker and the AC circuit breaker in the intermediate equipment between the DC side of the photovoltaic power station and the AC side of the photovoltaic power station, a first one-way fault feature mapping set is arranged; the tripping coil current of the DC circuit breaker is associated with the circuit breaking fault.

[0036] Specifically, the intermediate equipment is a key component connecting the DC side and the AC side, mainly including an inverter, a DC circuit breaker and an AC circuit breaker, etc.; the function of the inverter is to convert the DC power generated by the photovoltaic array into AC power to supply the AC load or to be connected to the power grid; the DC circuit breaker is used to protect the DC side circuit, and when overcurrent, short circuit and other faults are detected on the DC side, the circuit can be quickly cut off to prevent the fault from expanding; the AC circuit breaker is used to protect the AC side circuit, and when abnormal current occurs on the AC side, the AC circuit is cut off in time to ensure the safe operation of the whole system; the first one-way fault feature mapping set is a fault feature database, which is established through a large number of experiments, simulations and statistical analysis of actual operation data, and is used to associate and map specific fault types and corresponding electrical characteristic change modes to provide a basis for rapid and accurate fault identification, wherein the tripping coil current of the DC circuit breaker is a key electrical characteristic parameter and is directly associated with the circuit breaking fault; when a fault occurs on the DC side, such as overload, short circuit, etc., the tripping coil current of the DC circuit breaker will change, for example, the current will suddenly increase or fluctuate abnormally, thereby triggering the action of the circuit breaker to cut off the fault circuit, so the change of the tripping coil current can be used as an important basis for judging whether the DC circuit breaker has a fault and the severity of the fault.

[0037] The execution step is: according to the inverters, DC circuit breakers and AC circuit breakers in the intermediate equipment between the DC side and the AC side of the photovoltaic power station, a first one-way fault feature mapping set is set, specifically, through long-term monitoring and fault case analysis, electrical feature data of the inverters, DC circuit breakers and AC circuit breakers in normal operation and various fault states are collected, including inverter output current harmonic distortion rate, DC circuit breaker tripping coil current waveform, AC circuit breaker contact temperature change, etc.; through statistical analysis of these data, the first one-way fault feature mapping set is established, for example, when the inverter has a slight fault, the output current harmonic distortion rate may increase from 5% in normal time to 8%; when the DC circuit breaker has a poor contact fault, the tripping coil current waveform will have periodic fluctuations, and the fluctuation amplitude reaches 1.2 times of the normal current; when the contact resistance of the AC circuit breaker increases due to contact wear, the contact temperature may be higher than the normal operation temperature by more than 15℃; the correspondence between these feature changes and specific fault types is recorded in detail in the first one-way fault feature mapping set; based on the first one-way fault feature mapping set, the fault recognition and collaborative control optimization provide accurate fault feature basis, so that the system can quickly and accurately identify the fault type of the intermediate equipment, take corresponding control measures in time, and effectively improve the accuracy and reliability of the DC side fault prediction of the photovoltaic power station, and ensure the safe and stable operation of the photovoltaic power station.

[0038] Further, in combination with the leakage action value of the DC circuit breaker and the overcurrent protection threshold value of the AC circuit breaker, the filter and the reactive power compensation device are collaboratively controlled and optimized, and the method described in the application comprises: The inverter output current harmonic distortion rate, the DC circuit breaker tripping coil current mutation and the AC circuit breaker contact temperature anomaly are taken as the first fault feature; after the first fault feature is triggered, the first collaborative control action corresponding to the filter and the second collaborative control action corresponding to the reactive power compensation device are set; according to the first collaborative control action and the second collaborative control action, rolling optimization is performed to balance the filtering effect and the reactive power compensation efficiency.

[0039] Specifically, the inverter output current harmonic distortion rate refers to the proportion of harmonic components in the output current when the inverter converts direct current into alternating current, which reflects the quality of the inverter output current. Normally, the harmonic distortion rate is low, and when the inverter fails, the harmonic distortion rate will increase significantly. The DC circuit breaker tripping coil current mutation refers to the abnormal change of the tripping coil current of the DC circuit breaker during operation, which is usually manifested as a sudden increase or fluctuation of the current. It is an important feature of DC side circuit failure and is related to short circuit, overload and other faults. The abnormal temperature of the AC circuit breaker contact refers to the temperature of the contact part of the AC circuit breaker exceeding the normal range, which is caused by poor contact, overcurrent and other reasons, and also indicates that there may be potential faults on the AC side. The first fault feature takes these three key features as trigger conditions, and when they exceed the set normal threshold, it is determined that the first fault feature is triggered. The first cooperative control action corresponding to the filter refers to a specific control operation on the filter to improve the filtering effect, such as adjusting the resonant frequency of the filter, adding more filter branches, etc. The second cooperative control action corresponding to the reactive power compensation device refers to the control operation on the reactive power compensation device to improve the efficiency of reactive power compensation, such as dynamically adjusting the output capacity of the reactive power compensation device, changing the compensation mode, etc. Rolling optimization is a dynamic control strategy that continuously adjusts the control parameters of the filter and reactive power compensation device based on real-time monitoring data to maximize the efficiency of reactive power compensation while ensuring the filtering effect.

[0040] Execution steps: According to the first one-way fault feature mapping set, it is judged that the inverter may fail, at which time the first cooperative control action of the filter is triggered, the resonant frequency of the filter is adjusted to better filter out harmonics of a specific frequency, and the second cooperative control action of the reactive power compensation device is set to dynamically increase the compensation capacity to stabilize the DC bus voltage. Through the above steps, the development of the fault is effectively suppressed, the impact of the fault on the power generation of the photovoltaic power station is reduced, the effectiveness of fault prevention is improved, and the safe and stable operation of the photovoltaic power station is ensured.

[0041] Real-time monitoring of inverter output current harmonic distortion, DC circuit breaker tripping coil current and AC circuit breaker contact temperature, for example, in the normal operation of photovoltaic power station, the inverter output current harmonic distortion is about 5%, the DC circuit breaker tripping coil current is about 10A, and the AC circuit breaker contact temperature is maintained below 60℃; When the system detects that the inverter output current harmonic distortion suddenly rises to 8%, the DC circuit breaker tripping coil current surges to 15A, and the AC circuit breaker contact temperature rises to 75℃, the first fault feature is determined to be triggered, at this time, the first cooperative control action of the filter is set, such as adjusting the resonance frequency of the filter from 50Hz to 55Hz, to better filter out harmonics of a specific frequency; At the same time, the second cooperative control action of the reactive power compensation device is set, such as increasing the output capacity of the reactive power compensation device from 80kvar to 100kvar, to improve the reactive power support of the system.

[0042] According to the real-time monitoring data, a rolling optimization algorithm is used to adjust the control parameters of the filter and the reactive power compensation device once in a fixed time period (such as 1S) to balance the filtering effect and the reactive power compensation efficiency. Specifically, by continuously optimizing the parameter adjustment amount of the filter and the capacity configuration of the reactive power compensation device, the filtering effect and the reactive power compensation efficiency are improved, the rapid response and effective treatment of the fault are realized, the operation of the filter and the reactive power compensation device is optimized through cooperative control, the stability and reliability of the system are improved, and the influence of the fault on the normal operation of the photovoltaic power station is reduced.

[0043] Further, the first cooperative control action corresponding to the filter and the second cooperative control action corresponding to the reactive power compensation device are set, and the method described in the application comprises: After the first fault feature is triggered, the main harmonic frequency is identified and the resonance frequency of the filter is adjusted as the first cooperative control action; Based on the DC bus voltage fluctuation amplitude, the compensation capacity of the reactive power compensation device is dynamically adjusted as the second cooperative control action.

[0044] Specifically, the first fault feature trigger refers to when the monitored inverter output current harmonic distortion, DC circuit breaker tripping coil current mutation, and AC circuit breaker contact temperature anomaly exceed the set normal threshold, the system determines that the first fault feature has occurred, and the corresponding cooperative control action needs to be started; the main harmonic frequency refers to the dominant frequency component in the harmonic component of the inverter output current, which is usually the frequency with the highest harmonic content. It reflects the main source of harmonic pollution. Further, the resonant frequency of the filter is adjusted by changing the electrical parameters of the filter, such as inductance, capacitance, etc., so that its resonant frequency matches the main harmonic frequency, thereby enhancing the filter's ability to filter out the main harmonic; the compensation capacity of the reactive power compensation device refers to the size of the reactive power that the reactive power compensation device can provide, which is used to adjust the voltage level and power factor of the system. Further, the compensation capacity of the reactive power compensation device is dynamically adjusted according to the fluctuation of the DC bus voltage to stabilize the DC bus voltage and improve power quality.

[0045] Execution steps: After the first fault feature trigger, the inverter output current is analyzed by fast Fourier transform (FFT), and the main harmonic frequency is identified. For example, when the system detects the first fault feature trigger, FFT analysis finds that the main harmonic frequency in the inverter output current is the 5th harmonic (250Hz); at this time, the resonant frequency of the filter is adjusted as the first cooperative control action, i.e. the resonant frequency of the filter is adjusted from the original 50Hz to 250Hz, so that it can more effectively filter out the 5th harmonic and improve the filtering effect; at the same time, the fluctuation amplitude of the DC bus voltage is monitored. Assuming that the DC bus voltage fluctuation amplitude reaches ±10V, based on this fluctuation amplitude, the compensation capacity of the reactive power compensation device is dynamically adjusted as the second cooperative control action, for example, the compensation capacity of the reactive power compensation device is increased from the original 100kvar to 150kvar to stabilize the DC bus voltage; through the above steps, rapid response and accurate processing of faults are realized. By adjusting the resonant frequency of the filter and the compensation capacity of the reactive power compensation device, harmonics can be effectively filtered out, the DC bus voltage can be stabilized, the power quality and stability of the system can be improved, the impact of faults on the operation of the photovoltaic power station can be reduced, and the effectiveness of fault prevention and the reliability of the system can be improved.

[0046] Further, based on the second one-way fault feature mapping set corresponding to the conversion of electrical energy to DC power, the method described in the present application includes: According to the rectifier, DC circuit breaker and AC circuit breaker in the intermediate equipment between the DC side of the photovoltaic power station and the AC side of the photovoltaic power station, a second one-way fault feature mapping set is set; the action time of the AC circuit breaker is associated with the reverse flow fault.

[0047] Specifically, the intermediate device refers to a key component connecting the DC side and the AC side, including a rectifier, a DC circuit breaker and an AC circuit breaker. The rectifier is used to convert AC power into DC power, and is usually applied to the scene where AC power is accessed to the DC side in a photovoltaic power station, such as the charging and discharging process of an energy storage system. The DC circuit breaker is used to protect the DC circuit and cut off the circuit when overcurrent, short circuit or other abnormalities are detected. The AC circuit breaker is used to protect the AC circuit and prevent the AC side fault from affecting the entire system. The second one-way fault feature mapping set is a database established through experiments, simulations and operation data analysis, which is used to associate the AC side fault features with specific electrical parameter change patterns. The reverse flow fault refers to the current flowing from the AC side to the DC side abnormally, which is usually related to the failure of the rectifier or abnormal voltage of the AC side, and may cause equipment damage and system instability. The action time of the AC circuit breaker refers to the time required from detecting the fault to cutting off the circuit by the AC circuit breaker, which is closely related to the severity of the reverse flow fault.

[0048] The execution step is as follows: according to the rectifier, the DC circuit breaker and the AC circuit breaker in the intermediate device between the DC side and the AC side of the photovoltaic power station, the second one-way fault feature mapping set is set. Specifically, by monitoring and analyzing, when the reverse flow fault occurs in the AC side, the rectifier input current flows reversely, the AC circuit breaker action time is delayed, and the DC bus capacitor voltage oscillation is significant. The second one-way fault feature mapping set is established to associate these features with the reverse flow fault type. For example, when the rectifier input current flows reversely by 1.5 times of the normal current, the AC circuit breaker action time is delayed by more than 20 milliseconds, and the DC bus capacitor voltage oscillation amplitude is more than 5%, the mapping set determines that it is a reverse flow fault. In the above step, the second one-way fault feature mapping set is set, which provides a basis for subsequent fault identification and collaborative control optimization, can quickly identify the reverse flow fault, effectively improves the accuracy and reliability of the DC side fault prediction of the photovoltaic power station, and ensures the safe and stable operation of the system.

[0049] Further, in combination with the leakage action value of the DC circuit breaker and the overcurrent protection threshold of the AC circuit breaker, the filter and the reactive power compensation device are collaboratively controlled and optimized. The method described in the application comprises: The rectifier input current flows reversely, the AC circuit breaker action time is delayed, and the DC bus capacitor voltage oscillation are taken as the second fault features. After the second fault features are triggered, the third collaborative control action corresponding to the filter and the fourth collaborative control action corresponding to the reactive power compensation device are set. According to the third collaborative control action and the fourth collaborative control action, the real-time linkage of the filter and the reactive power compensation device is performed using the industrial Ethernet.

[0050] Specifically, the rectifier input current reverse flow refers to the current direction being opposite to the normal flow direction designed, and the current should normally flow from the alternating current side to the direct current side, and if it is reversed, it means that there is a fault; the alternating current circuit breaker action time delay refers to the time from the detection of the fault current by the alternating current circuit breaker to the cutting off of the circuit exceeding the normal range, which may cause the fault to expand; the direct current bus capacitor voltage oscillation refers to the fluctuation of the voltage across the direct current bus capacitor, which is usually manifested as the periodic change of the voltage amplitude, and may be caused by system disturbance or fault, and these features together constitute the second fault feature, which is used to judge whether a specific fault occurs in the system: when the triggering condition is met, the third cooperative control action and the fourth cooperative control action will be executed in turn, that is, the operating parameters of the filter and the reactive power compensation device are adjusted respectively to cope with the fault; the industrial Ethernet is an Ethernet variant suitable for industrial environments, which has the characteristics of strong anti-interference ability and high transmission rate, and is commonly used to connect and control industrial equipment to realize data transmission and cooperative work between devices.

[0051] The execution step is to judge whether the second fault feature is triggered, for example, when the photovoltaic power station is normally operated, the rectifier input current direction is correct, the alternating current circuit breaker action time is about 20 milliseconds, and the direct current bus capacitor voltage is stable within a fluctuation range of ±5%; when it is monitored that the rectifier input current reverse flow reaches 1.2 times the rated current, the alternating current circuit breaker action time is delayed to 30 milliseconds, and the direct current bus capacitor voltage oscillation amplitude reaches ±8%, it is determined that the second fault feature is triggered, at this time, the system immediately sets the third cooperative control action corresponding to the filter, such as switching the filter to the reverse flow suppression mode to block the reverse current path, and at the same time, sets the fourth cooperative control action corresponding to the reactive power compensation device, such as switching the reactive power compensation device from the constant voltage mode to the damping mode according to the direct current bus capacitor voltage oscillation to suppress oscillation, in this process, the high-speed transmission and real-time control capability of the industrial Ethernet is used to realize the real-time linkage of the filter and the reactive power compensation device, preferably, the transmission rate of the industrial Ethernet is as high as 100 Mbps, which can ensure that the control signal is transmitted to the target device within 1 millisecond to realize precise control; in the above steps, by monitoring the key fault features in real time and responding quickly, the development of the fault is effectively suppressed, the influence of the fault on the system is reduced, and the stable operation of the photovoltaic power station is ensured.

[0052] Further, the third cooperative control action corresponding to the filter and the fourth cooperative control action corresponding to the reactive power compensation device are set, and the method described in the application includes: After the first fault feature is triggered, the filter is switched to the reverse flow suppression mode to block the reverse current path as the third cooperative control action; according to the direct current bus capacitor voltage oscillation, the reactive power compensation device is switched from the constant voltage mode to the damping mode for oscillation suppression as the fourth cooperative control action.

[0053] Specifically, the reverse flow suppression mode is a filter operation mode for preventing reverse flow of current. When a fault of reverse flow of current is detected, the filter blocks the path of reverse current by adjusting its internal circuit structure or control strategy, such as changing the input amount of filter capacitance, adjusting the current direction of filter inductance, etc., so as to protect the system from damage caused by reverse current; blocking the path of reverse current refers to cutting off the circulation loop of reverse current by physical or electrical means to ensure that the current can only flow in the normal direction; the constant voltage mode refers to that the reactive power compensation device automatically adjusts the output of reactive power according to the preset voltage target value to maintain the stability of the DC bus voltage; the damping mode is a control mode for suppressing system oscillation. The reactive power compensation device adjusts the output of reactive power quickly to follow the oscillation change of the DC bus capacitor voltage, and injects reactive power in the opposite direction of the oscillation direction into the system, thereby effectively suppressing voltage oscillation.

[0054] The execution steps are as follows: after the first fault feature is triggered, the system first switches the filter to the reverse flow suppression mode to block the reverse current path. For example, when the input current of the rectifier is detected to flow in the reverse direction to 10A, the filter changes its internal circuit parameters to switch the originally used capacitor for filtering to a reverse current blocking module, which uses the unidirectional conductivity of a diode to block the reverse current path, so that the reverse current is reduced to below 1A, effectively protecting the system from the impact of reverse current. At the same time, according to the oscillation of the DC bus capacitor voltage, the reactive power compensation device is switched from the constant voltage mode to the damping mode for oscillation suppression. For example, when the oscillation amplitude of the DC bus capacitor voltage is ±8V and the frequency is 10Hz, the reactive power compensation device adjusts the output of reactive power with a response speed of 1ms to inject reactive power in the opposite direction of the oscillation direction into the system, so that the voltage oscillation amplitude is reduced to below ±2V within 5 seconds. In the above steps, by quickly switching the operation modes of the filter and the reactive power compensation device, the reverse flow fault and voltage oscillation are effectively suppressed, the stability and reliability of the system are improved, and the impact of the fault on the normal operation of the photovoltaic power station is reduced.

[0055] Further, the filter and the reactive power compensation device are cooperatively controlled and optimized. The method described in the application comprises the following steps: The filter adopts an LC filter topology, and the filter parameters are adjusted according to the Uth harmonic content and the Vth harmonic content in the frequency domain harmonic characteristics; the compensation capacity of the reactive power compensation device is dynamically configured according to the power factor target value, and the filter and the reactive power compensation device are cooperatively controlled by using a pulse synchronization mechanism in combination with the filter parameters.

[0056] Specifically, the LC filter topology refers to a filter circuit structure composed of inductance (L) and capacitance (C). Through the synergistic effect of inductance and capacitance, the LC filter topology can effectively filter out harmonics of specific frequencies. Among them, the Uth harmonic content and the Vth harmonic content represent the proportion of specific harmonic components in the signal. By monitoring the changes of these harmonic contents, the parameters of the filter can be adjusted in real time to enhance the filtering effect of the main harmonics. The compensation capacity of the reactive power compensation device refers to the size of the reactive power it can provide, which is used to adjust the voltage and power factor of the system. The power factor target value is the power factor value that the system expects to achieve, which is usually set to a value close to 1 to improve power quality and system efficiency. The pulse synchronization mechanism is a control strategy that coordinates the actions of the filter and the reactive power compensation device through a synchronous pulse signal, ensuring that the two are synchronized in time to achieve more precise collaborative control.

[0057] Execution steps: The filter adopts an LC filter topology, and the filter parameters are adjusted according to the Uth harmonic content and the Vth harmonic content in the frequency domain harmonic characteristics. For example, real-time monitoring finds that the 5th (Uth) and 7th (Vth) harmonic contents of the inverter output current are high, at 12% and 8% respectively. At this time, the system will automatically adjust the inductance and capacitance parameters of the filter according to these harmonic contents, specifically increasing the inductance value to 0.5H and increasing the capacitance value to 100μF, to enhance the filtering effect of the 5th and 7th harmonics, and reduce the total harmonic distortion rate of the filtered current to below 5%.

[0058] At the same time, the compensation capacity of the reactive power compensation device is dynamically configured according to the power factor target value. If the current system power factor is 0.85 and the power factor target is 0.95, the required reactive power compensation capacity is determined, and the compensation capacity is increased from 80kvar to 120kvar to increase the power factor to the target value. Combined with the change of the filter parameters, the pulse synchronization mechanism is used for collaborative control of the filter and the reactive power compensation device. The synchronous pulse signal is transmitted through industrial Ethernet to ensure that the actions of the filter and the reactive power compensation device are synchronized. The filter adjusts the parameters upon receiving the pulse signal, and the reactive power compensation device updates the capacity configuration at the same time. The synergistic effect of the two makes the system recover to normal operation in a short time, ensuring the stability and power quality of the photovoltaic power station, achieving efficient cooperation of filtering and reactive power compensation, and effectively improving the stability and reliability of the system.

[0059] In summary, the beneficial effects of the embodiments of the present application are: Due to the characteristic processing of the direct current side electrical parameters, the time domain fluctuation characteristics and the frequency domain harmonic characteristics are extracted; based on the first one-way fault feature mapping set corresponding to the conversion of direct current into alternating current and the second one-way fault feature mapping set corresponding to the conversion of alternating current into direct current, in combination with the leakage action value of the direct current circuit breaker and the overcurrent protection threshold of the alternating current circuit breaker, the filter and the reactive power compensation device are cooperatively controlled and optimized; meanwhile, the time domain fluctuation characteristics and the frequency domain harmonic characteristics are used as local electrical characteristics, the device state characteristics corresponding to the intermediate equipment are used as global electrical characteristic combinations, the fusion feature data is obtained by using feature selection downsampling, the fusion feature data is compared and analyzed with the normal operation feature data in the data storage unit, the fault risk index is evaluated and the fault risk reminding is performed. The application provides a photovoltaic power station direct current side fault prediction method and system. The time domain fluctuation characteristics and the frequency domain harmonic characteristics of the direct current side electrical parameters are extracted as local electrical characteristics, the device state characteristics of the intermediate equipment are combined as global electrical characteristic combinations and the fusion feature data is obtained by using feature selection downsampling, the deep fusion of local and global characteristics is performed, the filter and the reactive power compensation device are cooperatively controlled and optimized according to the protection parameters of the direct current circuit breaker and the alternating current circuit breaker, the cross-side fault feature linkage mechanism is established, and the technical effect of improving the fault prevention effectiveness is achieved.

[0060] In the embodiment two, based on the same inventive concept as the photovoltaic power station direct current side fault prediction method in the foregoing embodiment, as shown in the embodiment two, the application provides a photovoltaic power station direct current side fault prediction system, wherein the system comprises: Figure 2 The feature processing module M100 is configured to perform the following method:

[0061] The cooperative control module M200 is configured to perform the following method:

[0062] The feature selection module M300 is configured to perform the following method:

[0063] The fault risk reminding module M400 is configured to perform the following method:

[0064] Further, the feature processing module M100 is configured to perform the following method: ​Based on the photovoltaic array group string, the DC side raw data including group string voltage, branch current, insulation resistance are obtained; the DC side raw data are time sequence aligned to form the DC side electrical parameters including group string voltage sequence, branch current sequence and insulation resistance sequence.

[0065] Further, the cooperative control module M200 is used to execute the following method: According to the inverters, DC circuit breakers and AC circuit breakers in the intermediate equipment between the DC side of the photovoltaic power station and the AC side of the photovoltaic power station, a first one-way fault feature mapping set is set; the tripping coil current of the DC circuit breaker is associated with the circuit breaking fault.

[0066] Further, the cooperative control module M200 is also used to execute the following method: The harmonic distortion rate of the inverter output current, the tripping coil current mutation of the DC circuit breaker and the abnormal contact temperature of the AC circuit breaker are taken as the first fault feature; after the first fault feature is triggered, the filter corresponding first cooperative control action and the second cooperative control action corresponding to the reactive power compensation device are set; according to the first cooperative control action and the second cooperative control action, rolling optimization is performed to balance the filtering effect and the reactive power compensation efficiency.

[0067] Further, the cooperative control module M200 is also used to execute the following method: After the first fault feature is triggered, the main harmonic frequency is identified and the resonant frequency of the filter is adjusted as the first cooperative control action; based on the DC bus voltage fluctuation amplitude, the compensation capacity of the dynamic adjustment of the reactive power compensation device is taken as the second cooperative control action.

[0068] Further, the cooperative control module M200 is also used to execute the following method: According to the rectifiers, DC circuit breakers and AC circuit breakers in the intermediate equipment between the DC side of the photovoltaic power station and the AC side of the photovoltaic power station, a second one-way fault feature mapping set is set; the action time of the AC circuit breaker is associated with the reverse flow fault.

[0069] Further, the cooperative control module M200 is also used to execute the following method: The reverse flow of the rectifier input current, the action time delay of the AC circuit breaker and the oscillation of the DC bus capacitor voltage are taken as the second fault feature; after the second fault feature is triggered, the filter corresponding third cooperative control action and the fourth cooperative control action corresponding to the reactive power compensation device are set; according to the third cooperative control action and the fourth cooperative control action, the real-time linkage of the filter and the reactive power compensation device is performed using industrial Ethernet.

[0070] Further, the cooperative control module M200 is also used to execute the following method: After the first fault feature triggers, the filter is switched to a reverse flow suppression mode to block the reverse current path as a third coordinated control action; and the reactive power compensation device is switched from a constant voltage mode to a damping mode for oscillation suppression according to the DC bus capacitor voltage oscillation as a fourth coordinated control action.

[0071] Further, the coordinated control module M200 is further configured to perform the following method: The filter adopts an LC filter topology, and the filter parameters are adjusted according to the Uth harmonic content and the Vth harmonic content in the frequency domain harmonic characteristics; the compensation capacity of the reactive power compensation device is dynamically configured according to the power factor target value, and the filter and the reactive power compensation device are coordinated controlled by using a pulse synchronization mechanism in combination with the filter parameters.

[0072] In summary, any step can be stored as computer instructions or programs in an unrestricted computer memory and can be called and recognized by an unrestricted computer processor, and no additional limitation is made here.

[0073] Further, the above technical solution only embodies the preferred technical solution of the technical solution of the embodiments of the present application, and some variations of certain parts made by the person skilled in the art also embody the principles of the novel embodiments of the present application. Obviously, the person skilled in the art can make various modifications and variations to the present application without departing from the scope of the present application.

Claims

1. A method for predicting direct current side fault of a photovoltaic power station, characterized in that, The method comprises: characteristic processing of the direct current side electrical parameters, extracting time domain fluctuation characteristics and frequency domain harmonic characteristics; based on the first one-way fault feature mapping set corresponding to the conversion of direct current to alternating current and the second one-way fault feature mapping set corresponding to the conversion of alternating current to direct current, combined with the leakage action value of the direct current circuit breaker and the overcurrent protection threshold of the alternating current circuit breaker, the filter and the reactive power compensation device are cooperatively controlled and optimized; Meanwhile, the time domain fluctuation characteristics and the frequency domain harmonic characteristics are used as local electrical characteristics, and the device state characteristics corresponding to the intermediate device are used as global electrical characteristics for combination, and a fusion feature data is obtained by using feature selection downsampling; The fusion feature data is compared and analyzed with the normal operation feature data in the data storage unit to evaluate the fault risk index and give a fault risk reminder.

2. The method of claim 1, wherein the method further comprises: The method comprises: Based on the photovoltaic array group string, the direct current side original data including group string voltage, branch current and insulation resistance are obtained; The direct current side original data are time series aligned to form direct current side electrical parameters including group string voltage sequence, branch current sequence and insulation resistance sequence.

3. The method of claim 1, wherein the method further comprises: Based on the first one-way fault feature mapping set corresponding to the conversion of direct current to alternating current, the method comprises: According to the inverter, the direct current circuit breaker and the alternating current circuit breaker in the intermediate device between the direct current side of the photovoltaic power station and the alternating current side of the photovoltaic power station, a first one-way fault feature mapping set is set; The tripping coil current of the direct current circuit breaker is related to the circuit breaking fault.

4. The method of claim 3, wherein the method further comprises: Combined with the leakage action value of the direct current circuit breaker and the overcurrent protection threshold of the alternating current circuit breaker, the filter and the reactive power compensation device are cooperatively controlled and optimized, the method comprises: The inverter output current harmonic distortion rate, the direct current circuit breaker tripping coil current mutation and the alternating current circuit breaker contact temperature anomaly are taken as the first fault feature; After the first fault feature is triggered, the first cooperative control action corresponding to the filter and the second cooperative control action corresponding to the reactive power compensation device are set; According to the first cooperative control action and the second cooperative control action, rolling optimization is performed to balance the filtering effect and the reactive power compensation efficiency.

5. The method of claim 4, wherein the method further comprises: The first cooperative control action corresponding to the filter and the second cooperative control action corresponding to the reactive power compensation device are set, the method comprises: After the first fault feature is triggered, the main harmonic frequency is identified and the resonant frequency of the filter is adjusted as the first cooperative control action; Based on the direct current bus voltage fluctuation amplitude, the compensation capacity of the reactive power compensation device is dynamically adjusted as the second cooperative control action.

6. The method of claim 4, wherein the method further comprises: Based on the second one-way fault feature mapping set corresponding to the conversion of electrical energy to direct current, the method comprises: According to the rectifier, the direct current circuit breaker and the alternating current circuit breaker in the intermediate device between the direct current side of the photovoltaic power station and the alternating current side of the photovoltaic power station, a second one-way fault feature mapping set is set; The action time of the alternating current circuit breaker is related to the reverse flow fault.

7. The method of claim 6, wherein the method further comprises: Combined with the leakage action value of the direct current circuit breaker and the overcurrent protection threshold of the alternating current circuit breaker, the filter and the reactive power compensation device are cooperatively controlled and optimized, the method comprises: The rectifier input current reverse flow, the alternating current circuit breaker action time delay and the direct current bus capacitor voltage oscillation are taken as the second fault feature; After the second fault feature triggers, a third cooperative control action corresponding to the filter and a fourth cooperative control action corresponding to the reactive power compensation device are set; According to the third cooperative control action and the fourth cooperative control action, real-time linkage of the filter and the reactive power compensation device is performed using an industrial Ethernet.

8. The method of claim 7, wherein the method further comprises: The third cooperative control action corresponding to the filter and the fourth cooperative control action corresponding to the reactive power compensation device are set, and the method comprises: After the first fault feature triggers, the filter is switched to a reverse current suppression mode to block a reverse current path as a third cooperative control action; According to DC bus capacitor voltage oscillation, the reactive power compensation device is switched from a constant voltage mode to a damping mode for oscillation suppression as a fourth cooperative control action.

9. The method of claim 7, wherein the method further comprises: The filter and the reactive power compensation device are cooperatively controlled and optimized, and the method comprises: The filter adopts an LC filtering topology, and filtering parameters are adjusted according to Uth harmonic content and Vth harmonic content in frequency domain harmonic features; The compensation capacity of the reactive power compensation device is dynamically configured according to a power factor target value, and a pulse synchronization mechanism is used to cooperatively control the filter and the reactive power compensation device in combination with the filtering parameters.

10. A photovoltaic power plant DC side fault prediction system, characterized by, A step of a photovoltaic power station DC side fault prediction method according to any one of claims 1-9, the system comprises: a feature processing module: processing DC side electrical parameters to extract time domain fluctuation features and frequency domain harmonic features; a cooperative control module: based on a first one-way fault feature mapping set corresponding to DC-to-AC conversion and a second one-way fault feature mapping set corresponding to AC-to-DC conversion, in combination with a leakage action value of a DC circuit breaker and an overcurrent protection threshold of an AC circuit breaker, cooperatively controlling and optimizing the filter and the reactive power compensation device; a feature selection module: simultaneously, using the time domain fluctuation features and the frequency domain harmonic features as local electrical features, using device state features corresponding to intermediate devices as global electrical features, using feature selection downsampling to obtain fusion feature data; a fault risk reminding module: comparing and analyzing the fusion feature data with normal operation feature data in a data storage unit, evaluating a fault risk index, and reminding of a fault risk.

Images