Safety control method and system for new energy power system

Abstract

The invention discloses a safety control method and system for a new energy power system, and relates to the technical field of power control, and the method comprises the steps: constructing a behavior deviation distance value Dis with time correlation through the combined extraction and normalization analysis of a voltage deviation rate parameter, a frequency disturbance rate parameter and a power imbalance rate parameter; and the cooperative unbalance degree of the new energy power system in different energy dimensions is reflected. Compared with a traditional single-dimensional monitoring mode which only depends on voltage out-of-limit or frequency drift, the system hidden disturbance caused by wind power output fluctuation, photovoltaic array shadow shielding or load abrupt change and the like can be recognized earlier. By comparing the behavior deviation distance value Dis with the safety triggering threshold value St obtained through dynamic calculation, the judgment sensitivity can be automatically adjusted according to different operation conditions, and therefore the false triggering problem caused by new energy fluctuation characteristics is effectively avoided. And a corresponding control instruction set Cmd is output, and the instructions are issued to a control end for execution in real time through the communication module.

Description

Technical Field

[0001] This invention relates to the field of power control technology, specifically to a safety control method and system for a new energy power system. Background Technology

[0002] With the rapid growth of the proportion of new energy power generation, wind power systems, photovoltaic power systems, and energy storage units constitute a complex dynamic energy network in modern power systems. In this multi-source power environment, power quality and system security no longer depend solely on the robustness of hardware, but rather on the intelligence and flexibility of control methods. Especially in new energy scenarios, the fluctuation and intermittency of power output can easily trigger dynamic changes in system parameters, making it difficult for traditional single-dimensional safety control modes to respond accurately to complex multi-parameter disturbances.

[0003] Current new energy power systems mostly employ single-parameter monitoring methods for anomaly detection, such as using voltage exceeding limits or frequency deviations as the basis for judgment. However, in a multi-source grid-connected new energy environment, monitoring only a single parameter often fails to accurately reflect the true state of the system. For example, voltage fluctuations may be related to instantaneous power imbalances, while frequency disturbances may originate from energy storage response lags or communication delays. Existing methods ignore the coupling relationships between these multi-dimensional factors, leading to false alarms or missed alarms under complex operating conditions. When control logic triggers an error, the system may perform unnecessary load shedding, frequent energy storage start-ups and shutdowns, or even misjudge it as a system fault, resulting in decreased power utilization efficiency and system dispatch instability. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a safety control method and system for new energy power systems, solving the problems mentioned in the background section.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a safety control method for a new energy power system, comprising the following steps:

[0006] S1. Collect voltage, frequency, and power data during the operation of the new energy power system to construct a set of system operation characteristics;

[0007] S2. Normalize the system operation feature set to obtain a preprocessed feature set that has been standardized and extracted with time series features;

[0008] S3. Perform internal dynamic feature analysis on the preprocessed feature set, calculate the instantaneous volatility, mutual coupling degree and covariance relationship of each feature in the time series, extract the comprehensive deviation features reflecting the self-stability of the new energy power system, and calculate the behavioral deviation distance value.

[0009] S4. Compare the obtained behavior deviation distance value with the preset safety trigger threshold, establish an abnormal trigger set based on the comparison result, and then perform matching to generate a safety control command, which is then sent to the new energy power system control terminal for execution.

[0010] Preferably, S1 includes S11;

[0011] S11. During the operation of the new energy power system, real-time monitoring signals of the main control node and distributed nodes of the system are collected, including voltage signal sequence, frequency signal sequence and power output signal sequence. The relative deviation of each signal is calculated by signal analysis and time slicing processing, and the following parameters are extracted: voltage deviation rate parameter Vra, frequency disturbance rate parameter Fra and power imbalance rate parameter Pra.

[0012] The voltage deviation rate parameter Vra is obtained by segmenting the acquired voltage signal into time windows, calculating the difference between the average voltage within the window and the rated voltage of the system, and defining the proportion of the difference to the rated voltage as the voltage deviation rate parameter Vra.

[0013] The frequency disturbance rate parameter Fra is calculated by taking the sampling time interval as a basis, calculating the absolute value of the frequency change at adjacent time intervals, and then calculating the average change. The ratio of the average change to the rated frequency of the new energy power system is defined as the frequency disturbance rate parameter Fra, which is used to reflect the dynamic fluctuation of the system frequency in a short period.

[0014] The power imbalance rate parameter Pra is obtained by calculating the difference between the power generation and the load power within the same time period, obtaining the absolute value of the power deviation, and then comparing it with the maximum value of the power generation and the load power. The resulting ratio is defined as the power imbalance rate parameter Pra.

[0015] Preferably, S1 further includes S12;

[0016] S12. Based on the voltage deviation rate parameter Vra, frequency disturbance rate parameter Fra, and power imbalance rate parameter Pra obtained from multiple nodes, perform data verification and structured integration to obtain the system operation feature set Beh.

[0017] The data verification and structured integration uses multi-node parameters to check for outliers. When the voltage deviation rate parameter Vra, frequency disturbance rate parameter Fra, or power imbalance rate parameter Pra exceeds the preset constraint range, it is marked as an abnormal sample.

[0018] The verified parameter data is structured and integrated using time tags as the primary index and node identifiers as the secondary index to form a system operation feature set Beh with temporal continuity and spatial correlation.

[0019] Each record in the system operation feature set Beh contains three feature parameters and their source node information at the same time, realizing a global quantitative expression of the system operation status.

[0020] Preferably, S2 includes S21;

[0021] S21. Based on the acquired system operation feature set Beh, the voltage deviation rate parameter Vra, frequency disturbance rate parameter Fra, and power imbalance rate parameter Pra are extracted using the node identifier. Then, the extreme value normalization method is used to scale and map them to the [0,1] interval to eliminate the dimensional differences between different parameters. The normalized voltage normalized parameter Vrn, frequency normalized parameter Frn, and power normalized parameter Prn are obtained and formed into a normalized standard feature group, which is then bound to the node identifier.

[0022] Preferably, S2 further includes S22;

[0023] S22. Obtain the normalized standard feature group of each node identifier and perform time-series sliding analysis to extract dynamic mean and fluctuation amplitude features, forming a structured preprocessed feature set Bcl. The extraction is completed through steps S221, S222 and S223.

[0024] S221. Sliding segmentation: Set the sliding window length Win and step size Hop, and continuously segment each node identifier along the time label to form a series of time windows. The data segment in each time window contains three normalized parameters of multiple sampling times, which constitute a subset of the running features of a local time period.

[0025] S222, Statistical calculation within the window: For each segmented time window, calculate the mean and standard deviation of the three features, and obtain the voltage mean parameter Vrm, frequency mean parameter Frm, power mean parameter Prm, voltage fluctuation parameter Vrs, frequency fluctuation parameter Frs, and power fluctuation parameter Prs;

[0026] S223. The calculation results of the time window sequence are integrated according to the time window order and node identifier to form a structured preprocessing feature set Bcl.

[0027] Preferably, S3 includes S31;

[0028] S31. Read the average voltage parameter Vrm, average frequency parameter Frm, and average power parameter Prm of each node in the continuous time window from the preprocessed feature set Bcl, and analyze the characteristic change rate of adjacent time windows.

[0029] Based on the time window order, compare the same type of feature values ​​in adjacent windows one by one:

[0030] By comparing the average voltage parameter Vrm between the current time window and the previous time window, the difference in change is calculated, reflecting the instantaneous change in voltage on the time axis.

[0031] The variation difference between the frequency mean parameter Frm and the power mean parameter Prm is calculated simultaneously to obtain the variation amplitude within a continuous time window;

[0032] The differences in the changes of the obtained average voltage parameter Vrm, average frequency parameter Frm, and average power parameter Prm are divided by the time span between adjacent windows and then standardized to obtain standardized voltage dynamic fluctuation parameters Vrd, frequency dynamic fluctuation parameters Frd, and power dynamic fluctuation parameters Prd.

[0033] The voltage dynamic fluctuation parameter Vrd, frequency dynamic fluctuation parameter Frd, and power dynamic fluctuation parameter Prd calculated for each time window are bound to the corresponding node identifier to form a node-level dynamic fluctuation set.

[0034] Preferably, S3 further includes S32;

[0035] S32. Extract the voltage dynamic fluctuation parameter Vrd, frequency dynamic fluctuation parameter Frd, and power dynamic fluctuation parameter Prd, perform weighted fusion calculation, obtain the index reflecting the disturbance intensity of the new energy power system, and mark it as the behavior deviation distance value Dis;

[0036] The deviation distance value Dis is obtained by the following formula:

[0037] ;

[0038] In the formula, d1, d2 and d3 represent the weighting coefficients of the voltage dynamic fluctuation parameter Vrd, the frequency dynamic fluctuation parameter Frd and the power dynamic fluctuation parameter Prd, respectively, and d1+d2+d3=1, the specific values ​​of which are set by the user.

[0039] Preferably, S4 includes S41;

[0040] S41. Compare the obtained behavior deviation distance value Dis with the preset safety trigger threshold St, determine the operating status of the new energy power system based on the comparison result, and construct an abnormal trigger set Trig to identify and store trigger conditions and corresponding response instructions.

[0041] Among them, the safety trigger threshold St is calculated based on the sequence of behavioral deviation distance values ​​Dis within a fixed time window. It reflects the average deviation and fluctuation range of the operation of the new energy power system, and dynamically reflects the self-stabilizing state of the new energy power system.

[0042] The security trigger threshold St is obtained by the following calculation formula:

[0043] ;

[0044] In the formula, μDis represents the moving average of the behavioral deviation distance Dis within a fixed time window; σDis represents the moving standard deviation of the behavioral deviation distance Dis within a fixed time window; K represents the sensitivity coefficient;

[0045] The exception trigger set Trig is obtained through the following exception state identification method:

[0046] When the safety trigger threshold St is less than the safety trigger threshold St, the new energy power system is determined to be in a stable state under the current time window.

[0047] When the safety trigger threshold St is greater than or equal to the safety trigger threshold St, the new energy power system is determined to be in an abnormal state under the current time window.

[0048] The results of abnormal state identification, time window information, node identifier, voltage dynamic fluctuation parameter Vrd, frequency dynamic fluctuation parameter Frd, and power dynamic fluctuation parameter Prd are stored in the constructed abnormal trigger set Trig.

[0049] Preferably, S4 further includes S42;

[0050] S42. Extract the voltage dynamic fluctuation parameter Vrd, frequency dynamic fluctuation parameter Frd, and power dynamic fluctuation parameter Prd from the current abnormal time window; sort the real-time values ​​of the three parameters in ascending order using the bubble sort algorithm to obtain a fluctuation intensity sequence from low to high, and select the largest fluctuation intensity sequence as the dominant abnormal feature of the current time window, which is marked as the dominant fluctuation parameter Ddom.

[0051] Based on the acquired dominant fluctuation parameter Ddom, a set of control instructions Cmd is generated and sent to the control terminal of the new energy power system for execution.

[0052] The control command set Cmd is generated using the following matching method:

[0053] When the dominant fluctuation parameter Ddom is the voltage dynamic fluctuation parameter Vrd, a voltage steady-state adjustment instruction is generated, and the node identifier binding instruction of the voltage dynamic fluctuation parameter Vrd is extracted.

[0054] When the dominant fluctuation parameter Ddom is the frequency dynamic fluctuation parameter Frd, a frequency compensation control instruction is generated, and the node identifier binding instruction of the frequency dynamic fluctuation parameter Frd is extracted.

[0055] When the dominant fluctuation parameter Ddom is the power dynamic fluctuation parameter Prd, a power balance scheduling instruction is generated, and the node identifier binding instruction of the power dynamic fluctuation parameter Prd is extracted.

[0056] A safety control system for a new energy power system includes a system operation data acquisition module, a system data processing module, a data feature extraction module, and a control distribution module;

[0057] The system operation data acquisition module collects voltage data, frequency data, and power data during the operation of the new energy power system to construct a set of system operation characteristics;

[0058] The system data processing module performs normalization preprocessing on the system operation feature set to obtain a preprocessed feature set that has been standardized and extracted with time-series features;

[0059] The data feature extraction module performs internal dynamic feature analysis on the preprocessed feature set, calculates the instantaneous volatility, mutual coupling degree and covariance relationship of each feature in the time series, extracts the comprehensive deviation features reflecting the self-stability of the new energy power system, and calculates the behavioral deviation distance value.

[0060] The control delivery module compares the acquired behavior deviation distance value with the preset safety trigger threshold, establishes an abnormal trigger set based on the comparison result, and then generates a safety control command by matching it, which is then sent to the new energy power system control terminal for execution.

[0061] This invention provides a safety control method and system for a new energy power system, which has the following beneficial effects:

[0062] (1) By jointly extracting and normalizing the voltage deviation rate parameter, frequency disturbance rate parameter, and power imbalance rate parameter, a time-correlated behavioral deviation distance value Dis is constructed, which can simultaneously reflect the degree of coordinated imbalance of the new energy power system in different energy dimensions. Compared with the traditional single-dimensional monitoring method that only relies on voltage over-limit or frequency drift, this method can identify the system's hidden disturbances caused by wind power output fluctuations, photovoltaic array shading, or load changes earlier. By comparing the behavioral deviation distance value Dis with the dynamically calculated safety trigger threshold St, the judgment sensitivity can be automatically adjusted according to different operating conditions, thereby effectively avoiding false triggering problems caused by the fluctuation characteristics of new energy. The corresponding control command set Cmd is output, and the commands are sent to the control terminal for execution in real time through the communication module to realize rapid regulation such as voltage correction, frequency compensation, and power redistribution.

[0063] (2) By performing rate-of-change analysis on the average voltage parameter Vrm, average frequency parameter Frm, and average power parameter Prm within a continuous time window, standardized voltage dynamic fluctuation parameters Vrd, frequency dynamic fluctuation parameters Frd, and power dynamic fluctuation parameters Prd are extracted, enabling the system to identify the intensity of dynamic changes in voltage, frequency, and power on the time axis in real time. Subsequently, the three dynamic fluctuation characteristics are weighted and fused to calculate the behavioral deviation distance value Dis, which reflects the overall disturbance intensity, thereby compressing the complex multi-parameter time-varying behavior into a single quantifiable stability index. At this point, the traditional system may be misjudged as "stable" due to the mutual cancellation of local parameters, but by calculating the behavioral deviation distance value Dis, the overall disturbance trend of the system can still be correctly identified even when the fluctuation directions of each parameter are inconsistent, and it can be determined that the system has entered the dynamic imbalance stage. This realizes the quantitative evaluation of the internal dynamic coordination characteristics of the new energy power system, and can integrate the dispersed voltage, frequency, and power fluctuation information into a comprehensive operational stability index, providing a high-precision and traceable quantitative basis for subsequent anomaly triggering judgment and safety control.

[0064] (3) By comparing the real-time calculated behavioral deviation distance value Dis with the dynamically generated safety trigger threshold St, an abnormal trigger set Trig is automatically generated based on the comparison result. This enables the system to automatically adjust the judgment boundary under different fluctuating operating conditions. When the output of new energy changes frequently, the safety trigger threshold St can be adjusted in real time according to the moving average and standard deviation of the behavioral deviation distance value Dis through the dynamic threshold mechanism of this invention. This allows the system to identify the imbalance trend in the early stage of rapid wind speed changes and generate the abnormal trigger set Trig. Subsequently, the control command set Cmd is automatically generated based on the judgment result of the abnormal trigger set Trig and is sent to the distributed nodes for execution in real time. This includes multi-dimensional actions such as voltage steady-state adjustment, frequency compensation control, and power redistribution, so that the new energy power system can start the self-recovery process as soon as it detects an anomaly. Attached Figure Description

[0065] Figure 1 This is a schematic diagram illustrating the steps of a safety control method for a new energy power system according to the present invention;

[0066] Figure 2 This is a schematic diagram of a safety control system for a new energy power system according to the present invention;

[0067] Figure 3 This is a schematic diagram illustrating the dynamic changes in the operational deviation and safety trigger threshold of a new energy power system. Detailed Implementation

[0068] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0069] Example 1

[0070] This invention provides a safety control method for a new energy power system. Please refer to [link / reference]. Figure 1 This includes the following steps:

[0071] S1. Collect voltage, frequency, and power data during the operation of the new energy power system to construct a set of system operation characteristics;

[0072] S2. Normalize the system operation feature set to obtain a preprocessed feature set that has been standardized and extracted with time series features;

[0073] S3. Perform internal dynamic feature analysis on the preprocessed feature set, calculate the instantaneous volatility, mutual coupling degree and covariance relationship of each feature in the time series, extract the comprehensive deviation features reflecting the self-stability of the new energy power system, and calculate the behavioral deviation distance value.

[0074] S4. Compare the acquired behavioral deviation distance value with the preset safety trigger threshold, establish an abnormal trigger set based on the comparison result, then perform matching to generate a safety control command, and send it to the new energy power system control terminal for execution.

[0075] In this embodiment, a time-correlated behavioral deviation distance value Dis is constructed by jointly extracting and normalizing voltage deviation rate parameters, frequency disturbance rate parameters, and power imbalance rate parameters. This value can synchronously reflect the degree of coordinated imbalance of the new energy power system in different energy dimensions. Compared with the traditional single-dimensional monitoring method that only relies on voltage over-limit or frequency drift, this method can identify the system's hidden disturbances caused by wind power output fluctuations, photovoltaic array shading, or load changes earlier. In step S4, by comparing the behavioral deviation distance value Dis with the dynamically calculated safety trigger threshold St, the judgment sensitivity can be automatically adjusted according to different operating conditions, thereby effectively avoiding false triggering problems caused by the fluctuation characteristics of new energy. When the behavioral deviation distance value Dis exceeds the safety trigger threshold St, the system immediately generates an abnormal trigger set Trig and outputs a corresponding control command set Cmd. The commands are sent to the control terminal in real time via the communication module for execution, realizing rapid regulation such as voltage correction, frequency compensation, and power redistribution. By enabling adaptive correction within millisecond-level response periods when complex disturbances such as power imbalance, frequency drift, or voltage transients occur in the new energy power system, the system significantly improves its stability and security under conditions of high volatility and high grid connection ratios. For example, in the event of a transient increase in wind turbine power caused by a sudden increase in wind speed, the system can detect the deviation trend before the deviation distance value Dis exceeds the threshold, triggering power balance control commands in advance. This avoids problems such as bus voltage exceeding limits and frequency drops, ensuring the safe and stable operation of the power grid.

[0076] Example 2

[0077] Specifically: S1 includes S11;

[0078] S11. During the operation of the new energy power system, real-time monitoring signals of the main control node and distributed nodes of the system are collected, including voltage signal sequence, frequency signal sequence and power output signal sequence. The relative deviation of each signal is calculated by signal analysis and time slicing processing, and the following parameters are extracted: voltage deviation rate parameter Vra, frequency disturbance rate parameter Fra and power imbalance rate parameter Pra.

[0079] The voltage deviation rate parameter Vra is obtained by segmenting the acquired voltage signal into time windows, calculating the difference between the average voltage within the window and the rated voltage of the system, and defining the proportion of the difference to the rated voltage as the voltage deviation rate parameter Vra, so as to reflect the degree of deviation of the system voltage from the rated value.

[0080] The frequency disturbance rate parameter Fra is calculated by taking the sampling time interval as a basis, calculating the absolute value of the frequency change at adjacent time intervals, and then calculating the average change. The ratio of the average change to the rated frequency of the new energy power system is defined as the frequency disturbance rate parameter Fra, which is used to reflect the dynamic fluctuation of the system frequency in a short period.

[0081] The power imbalance rate parameter Pra is obtained by calculating the difference between the power generation and the load power in the same time period, obtaining the absolute value of the power deviation, and then comparing it with the maximum value of the power generation and the load power. The resulting ratio is defined as the power imbalance rate parameter Pra, which is used to measure the real-time balance of energy supply and demand in the system.

[0082] It should be noted that:

[0083] The main control node refers to the core node in a new energy power system that undertakes centralized control, power dispatch and grid connection management. It is usually located in the main control center or substation control unit and is used to collect system voltage, frequency, power and communication status information.

[0084] Distributed nodes refer to the local monitoring terminals or microgrid control terminals of distributed generation units (such as photovoltaic inverters, wind turbine control units, and energy storage converters). They directly collect local voltage, frequency, and power information and upload it to the main control node.

[0085] The voltage deviation rate parameter Vra is calculated as follows: Vra = |Vavg − Vref| ÷ Vref; where Vavg represents the average voltage value within the time window, and Vref represents the system rated voltage.

[0086] The frequency disturbance rate parameter Fra is calculated as follows: Fra = (1 / n) × Σ|fi−fi−1| ÷ Fref; where fi represents the i-th sampling frequency, Fref represents the system rated frequency, and n represents the number of sampling points.

[0087] The power imbalance rate parameter Pra is calculated as follows: Pra = |Pg − Pl| ÷ max(Pg, Pl); where Pg represents the generating power and Pl represents the load power.

[0088] S1 further includes S12;

[0089] S12. Based on the voltage deviation rate parameter Vra, frequency disturbance rate parameter Fra, and power imbalance rate parameter Pra obtained from multiple nodes, perform data verification and structured integration to obtain the system operation feature set Beh.

[0090] The data verification and structured integration uses multi-node parameters to check for outliers. When the voltage deviation rate parameter Vra, frequency disturbance rate parameter Fra, or power imbalance rate parameter Pra exceeds the preset constraint range, it is marked as an abnormal sample.

[0091] The verified parameter data is structured and integrated using time tags as the primary index and node identifiers as the secondary index to form a system operation feature set Beh with temporal continuity and spatial correlation.

[0092] Each record in the system operation feature set Beh contains three feature parameters and their source node information at the same time, realizing a global quantitative expression of the system operation status;

[0093] It should be noted that:

[0094] The system operation feature set Beh consists of several records, each of which is composed of the following fields: {time tag, node identifier, voltage deviation rate parameter Vra, frequency disturbance rate parameter Fra, power imbalance rate parameter Pra}. The primary key of the set is a combination of time tag and node identifier, which enables fast indexing and tracing. After the data is structured and integrated, it still maintains the consistency of physical measurement points and time resolution, ensuring the continuity of subsequent sliding window processing and anomaly identification.

[0095] The constraint ranges for the voltage deviation rate parameter Vra, the frequency disturbance rate parameter Fra, and the power imbalance rate parameter Pra are obtained from system design standards or historical operating characteristics statistics.

[0096] In this embodiment, by synchronously acquiring signal sequences at the main control node and distributed nodes, the voltage deviation rate parameter Vra, the frequency disturbance rate parameter Fra, and the power imbalance rate parameter Pra are extracted. A constraint interval-based verification mechanism is employed during data integration to ensure the consistency of the resulting system operation feature set Beh in both temporal and spatial dimensions. For example, when a node in a wind farm experiences a sudden drop in local voltage due to a wind speed change, traditional systems often only record the voltage anomaly without simultaneously capturing power and frequency changes. However, this invention, through a multi-node feature synchronous acquisition and verification integration mechanism, accurately reflects the correspondence between the voltage deviation rate parameter Vra, the frequency disturbance rate parameter Fra, and the power imbalance rate parameter Pra when forming the system operation feature set Beh. This allows for the identification that the anomaly is a typical transient coupling imbalance phenomenon rather than a single-point measurement error. Therefore, this stage not only significantly improves the reliability and correlation of the original feature data but also effectively suppresses the interference of communication delays, sampling noise, or isolated node errors on the overall data quality, providing a high-quality input foundation for the accurate calculation of the behavior deviation distance value Dis in subsequent stages.

[0097] Example 3

[0098] Specifically: S2 includes S21;

[0099] S21. Based on the acquired system operation feature set Beh, the voltage deviation rate parameter Vra, frequency disturbance rate parameter Fra, and power imbalance rate parameter Pra are extracted using the node identifier. Then, the extreme value normalization method is used to scale and map them to the [0,1] interval to eliminate the dimensional differences between different parameters. The normalized voltage normalized parameter Vrn, frequency normalized parameter Frn, and power normalized parameter Prn are obtained and formed into a normalized standard feature group, which is then bound to the node identifier.

[0100] It should be noted that:

[0101] Voltage normalization parameter Vrn, frequency normalization parameter Frn, and power normalization parameter Prn: correspond to the normalized voltage, frequency, and power characteristics, respectively. The three parameters retain the original parameter variation trend and proportional relationship, and their numerical range is unified to [0,1] to eliminate the dimensional differences between different characteristics.

[0102] S2 further includes S22;

[0103] S22. Obtain the normalized standard feature group of each node identifier and perform time-series sliding analysis to extract dynamic mean and fluctuation amplitude features, forming a structured preprocessed feature set Bcl. The extraction is completed through steps S221, S222 and S223.

[0104] S221. Sliding segmentation: Set the sliding window length Win and step size Hop, and continuously segment each node identifier along the time label to form a series of time windows. The data segment in each time window contains three normalized parameters of multiple sampling times, which constitute a subset of the running features of a local time period.

[0105] S222, Statistical calculation within the window: For each segmented time window, calculate the mean and standard deviation of the three features, and obtain the voltage mean parameter Vrm, frequency mean parameter Frm, power mean parameter Prm, voltage fluctuation parameter Vrs, frequency fluctuation parameter Frs, and power fluctuation parameter Prs;

[0106] S223. The calculation results of the time window sequence are integrated according to the time window order and node identifier to form a structured preprocessing feature set Bcl. Each record in the preprocessing feature set Bcl contains six statistical feature fields of a time window and its time and node index information, realizing the temporal description of the node running features.

[0107] It should be noted that:

[0108] The average operating level within the time period is obtained by summing all the sampled values ​​within the sliding window and dividing by the number of sampling points. These are defined as the average voltage parameter Vrm, the average frequency parameter Frm, and the average power parameter Prm, respectively.

[0109] By analyzing the deviation of each sampled value from the average level within the sliding window, indicators reflecting the fluctuation amplitude are obtained, which are defined as voltage fluctuation parameter Vrs, frequency fluctuation parameter Frs, and power fluctuation parameter Prs, respectively; voltage mean parameter Vrm, frequency mean parameter Frm, and power mean parameter Prm are used to represent the stable operating trend of the node within the sliding window.

[0110] Voltage fluctuation parameter Vrs, frequency fluctuation parameter Frs, and power fluctuation parameter Prs are used to represent the instantaneous fluctuation of operating characteristics.

[0111] In this embodiment, in traditional monitoring systems, the voltage, frequency, and power signals from different acquisition nodes exhibit significant dimensional differences and amplitude imbalances in their original data due to variations in measurement point voltage levels, sampling sensitivity, and equipment types. This makes it difficult to directly compare the fluctuation levels of different features in subsequent analysis, thus affecting the uniformity of overall anomaly detection. This invention maps the voltage deviation rate parameter Vra, frequency disturbance rate parameter Fra, and power imbalance rate parameter Pra to the [0,1] interval using the extreme value normalization method, obtaining the voltage normalized parameter Vrn, frequency normalized parameter Frn, and power normalized parameter Prn. This constructs a normalized standard feature set with consistent dimensions and accurate trends, thereby eliminating the influence of physical units between features and ensuring the direct comparability of multi-node data in the system. Based on this, a processing method combining sliding partitioning and in-window statistical analysis is used to further calculate the voltage mean parameter Vrm, frequency mean parameter Frm, power mean parameter Prm, and voltage fluctuation parameters Vrs, frequency fluctuation parameters Frs, and power fluctuation parameters Prs, forming a structured preprocessed feature set Bcl. This set not only describes the average operating level of the system but also simultaneously captures short-term fluctuation characteristics, enabling the system to more sensitively identify potential signs of dynamic disturbances. For example, in a photovoltaic grid-connected scenario, when rapid cloud movement causes periodic fluctuations in the output power of the photovoltaic array, traditional monitoring systems often fail to accurately identify stable trends due to large changes in the power reference. However, through the normalization and standardization processing and sliding window calculation mechanism of this invention, the numerical changes of the voltage normalization parameter Vrn, frequency normalization parameter Frn, and power normalization parameter Prn are unified to the same scale. The corresponding voltage fluctuation parameter Vrs, frequency fluctuation parameter Frs, and power fluctuation parameter Prs can simultaneously reflect the coupling characteristics of photovoltaic output disturbances and frequency response delays, enabling the system to determine whether there is a mismatch trend between power generation and load in the early stages. Therefore, by constructing a structured preprocessed feature set Bcl in the S2 stage, cross-node standardization, time continuity, and quantifiable dynamic fluctuations of the operating status of the new energy power system are achieved, providing a unified and highly reliable feature input basis for subsequent calculation of the behavioral deviation distance value Dis.

[0112] Example 4

[0113] Specifically: S3 includes S31;

[0114] S31. Read the average voltage parameter Vrm, average frequency parameter Frm, and average power parameter Prm of each node in the continuous time window from the preprocessed feature set Bcl, and analyze the characteristic change rate of adjacent time windows.

[0115] Based on the time window order, compare the same type of feature values ​​in adjacent windows one by one:

[0116] By comparing the average voltage parameter Vrm between the current time window and the previous time window, the difference in change is calculated, reflecting the instantaneous change in voltage on the time axis.

[0117] The variation difference between the frequency mean parameter Frm and the power mean parameter Prm is calculated simultaneously to obtain the variation amplitude within a continuous time window;

[0118] The differences in the changes of the obtained average voltage parameter Vrm, average frequency parameter Frm, and average power parameter Prm are divided by the time span between adjacent windows and then standardized to obtain standardized voltage dynamic fluctuation parameters Vrd, frequency dynamic fluctuation parameters Frd, and power dynamic fluctuation parameters Prd, which respectively represent the intensity of the change in the average voltage level, the intensity of the change in the frequency stability center, and the intensity of the change in the power balance state within a unit standardized time.

[0119] The voltage dynamic fluctuation parameter Vrd, frequency dynamic fluctuation parameter Frd, and power dynamic fluctuation parameter Prd calculated for each time window are bound to the corresponding node identifier to form a node-level dynamic fluctuation set.

[0120] It should be noted that:

[0121] Voltage dynamic fluctuation parameter Vrd: It is obtained by calculating the rate of change of the average voltage parameter Vrm in adjacent time windows. It represents the intensity of the change in the average voltage level per unit time. If this parameter continues to rise, it indicates that the system voltage has an unstable trend.

[0122] Frequency dynamic fluctuation parameter Frd: obtained by calculating the rate of change of the frequency mean parameter Frm in adjacent time windows, used to characterize the dynamic sensitivity of the system frequency response. A larger value usually means that the system has a frequency adjustment delay when the load fluctuates.

[0123] Power dynamic fluctuation parameter Prd: obtained by calculating the rate of change of the power mean parameter Prm in adjacent time windows. It is used to quantify the dynamic balance of power matching between the generation side and the load side. Its increase often indicates that the adjustment of energy storage unit or distributed power source has failed to respond to power deviation in a timely manner.

[0124] S3 further includes S32;

[0125] S32. Extract the voltage dynamic fluctuation parameter Vrd, frequency dynamic fluctuation parameter Frd, and power dynamic fluctuation parameter Prd, perform weighted fusion calculation, obtain the index reflecting the disturbance intensity of the new energy power system, and mark it as the behavior deviation distance value Dis;

[0126] The deviation distance value Dis is obtained by the following formula:

[0127] ;

[0128] In the formula, d1, d2 and d3 represent the weighting coefficients of voltage dynamic fluctuation parameter Vrd, frequency dynamic fluctuation parameter Frd and power dynamic fluctuation parameter Prd, respectively, and d1+d2+d3=1, the specific values ​​of which are set by the user.

[0129] It should be noted that:

[0130] When the deviation distance value Dis is small and fluctuates stably, it indicates that the voltage, frequency, and power are in dynamic equilibrium, and the new energy power system is operating stably.

[0131] When the deviation distance value Dis increases rapidly or remains above the threshold, it indicates that the system characteristics have synchronous disturbances, signifying a decline in the overall stability of the new energy power system.

[0132] In this embodiment, traditional operation monitoring methods typically rely on the over-limit judgment of a single parameter, such as issuing alarms based solely on voltage deviation or frequency offset. This fails to reflect the synergistic relationship between internal system features, leading to delayed or misjudged anomaly identification under multi-source fluctuation conditions. By performing rate-of-change analysis on the average voltage parameter Vrm, average frequency parameter Frm, and average power parameter Prm within a continuous time window, standardized voltage dynamic fluctuation parameters Vrd, frequency dynamic fluctuation parameter Frd, and power dynamic fluctuation parameter Prd are extracted. This enables the system to identify the dynamic intensity of voltage, frequency, and power changes over time in real time. Subsequently, the three dynamic fluctuation features are weighted and fused to calculate the behavioral deviation distance value Dis, which reflects the overall disturbance intensity. This compresses the complex multi-parameter time-varying behavior into a single quantifiable stability index. This mechanism allows the system not only to determine "whether an anomaly has occurred," but also to identify "which type of energy feature dominates the anomaly" and "the evolution trend of the anomaly intensity over time." For example, in the grid-connected operation scenario of a wind-solar hybrid power station, when a sudden increase in wind speed leads to an increase in wind turbine output while the output of the photovoltaic array decreases due to cloud shading, the voltage dynamic fluctuation parameter Vrd, the frequency dynamic fluctuation parameter Frd, and the power dynamic fluctuation parameter Prd will all show opposite changes in a short period of time. At this time, a traditional system may be misjudged as "stable" due to the mutual cancellation of local parameters. However, by calculating the behavioral deviation distance value Dis, the overall disturbance trend of the system can be correctly identified even when the fluctuation directions of various parameters are inconsistent, and it can be determined that the system has entered a dynamic imbalance stage. Therefore, this stage realizes the quantitative assessment of the dynamic coordination characteristics within the new energy power system, and can integrate the dispersed voltage, frequency, and power fluctuation information into a comprehensive operational stability index, providing a high-precision and traceable quantitative basis for subsequent anomaly trigger judgment and safety control.

[0133] Example 5

[0134] Please see Figure 1 and Figure 3 Specifically: S4 includes S41;

[0135] S41. Compare the obtained behavior deviation distance value Dis with the preset safety trigger threshold St, determine the operating status of the new energy power system based on the comparison result, and construct an abnormal trigger set Trig to identify and store trigger conditions and corresponding response instructions.

[0136] Among them, the safety trigger threshold St is calculated based on the sequence of behavioral deviation distance values ​​Dis within a fixed time window. It reflects the average deviation and fluctuation range of the operation of the new energy power system, and dynamically reflects the self-stabilizing state of the new energy power system.

[0137] The security trigger threshold St is obtained by the following calculation formula:

[0138] ;

[0139] In the formula, μDis represents the moving average of the behavioral deviation distance Dis within a fixed time window, used to reflect the overall deviation level; σDis represents the moving standard deviation of the behavioral deviation distance Dis within a fixed time window, used to reflect the fluctuation intensity; K represents the sensitivity coefficient, used to adjust the sensitivity of the safety trigger threshold St. When the system is running smoothly, σDis is small and the St value is high, thus avoiding false triggers; when the fluctuation intensifies, σDis increases and St automatically tightens to improve the warning sensitivity.

[0140] The exception trigger set Trig is obtained through the following exception state identification method:

[0141] When the safety trigger threshold St is less than the safety trigger threshold St, the new energy power system is determined to be in a stable state under the current time window.

[0142] When the safety trigger threshold St is greater than or equal to the safety trigger threshold St, the new energy power system is determined to be in an abnormal state under the current time window.

[0143] The results of abnormal state identification, time window information, node identifier, voltage dynamic fluctuation parameter Vrd, frequency dynamic fluctuation parameter Frd, and power dynamic fluctuation parameter Prd are stored in the constructed abnormal trigger set Trig.

[0144] S4 also includes S42;

[0145] S42. Extract the voltage dynamic fluctuation parameter Vrd, frequency dynamic fluctuation parameter Frd, and power dynamic fluctuation parameter Prd from the current abnormal time window; sort the real-time values ​​of the three parameters in ascending order using the bubble sort algorithm to obtain a fluctuation intensity sequence from low to high, and select the largest fluctuation intensity sequence as the dominant abnormal feature of the current time window, which is marked as the dominant fluctuation parameter Ddom.

[0146] Based on the acquired dominant fluctuation parameter Ddom, a set of control instructions Cmd is generated and sent to the control terminal of the new energy power system for execution.

[0147] The control command set Cmd is generated using the following matching method:

[0148] When the dominant fluctuation parameter Ddom is the voltage dynamic fluctuation parameter Vrd, a voltage steady-state adjustment instruction is generated, and the node identifier binding instruction of the voltage dynamic fluctuation parameter Vrd is extracted.

[0149] When the dominant fluctuation parameter Ddom is the frequency dynamic fluctuation parameter Frd, a frequency compensation control instruction is generated, and the node identifier binding instruction of the frequency dynamic fluctuation parameter Frd is extracted.

[0150] When the dominant fluctuation parameter Ddom is the power dynamic fluctuation parameter Prd, a power balance scheduling instruction is generated, and the node identifier binding instruction of the power dynamic fluctuation parameter Prd is extracted.

[0151] In this embodiment, in traditional power systems, safety triggering typically relies on fixed threshold settings. Alarms or controls are only triggered when voltage, frequency, or power deviates beyond the threshold. However, this static setting is difficult to adapt to the characteristics of rapid fluctuations in renewable energy generation output and large variations in grid-connected power, often leading to "delayed response" or "false triggering" problems. By comparing the real-time calculated behavioral deviation distance value Dis with the dynamically generated safety trigger threshold St, an abnormal trigger set Trig is automatically generated based on the comparison result, enabling the system to automatically adjust the judgment boundary under different fluctuating operating conditions. When renewable energy output changes frequently, such as a sudden power surge in a large-scale wind farm under gust conditions, relying solely on static thresholds may result in delayed detection of anomalies. However, through the dynamic threshold mechanism of this invention, the safety trigger threshold St can be adjusted in real time based on the moving average and standard deviation of the behavioral deviation distance value Dis, thereby identifying system imbalance trends in the early stages of rapid wind speed changes and generating an abnormal trigger set Trig. Subsequently, based on the judgment result of the anomaly trigger set Trig, a control command set Cmd is automatically generated and sent to the distributed nodes for execution in real time. This includes multi-dimensional actions such as voltage steady-state adjustment, frequency compensation control, and power redistribution, enabling the new energy power system to initiate a self-recovery process as soon as an anomaly is detected. Through the linkage of dynamic threshold judgment and distributed control commands, the new energy power system achieves rapid response, safe regulation, and continuous self-stabilization, significantly improving the system's stability and operational reliability in complex fluctuating environments.

[0152] Example 6

[0153] A safety control system for a new energy power system, please refer to Figure 2 Specifically, it includes a system operation data acquisition module, a system data processing module, a data feature extraction module, and a control distribution module;

[0154] The system operation data acquisition module collects voltage data, frequency data, and power data during the operation of the new energy power system to construct a set of system operation characteristics;

[0155] The system data processing module performs normalization preprocessing on the system operation feature set to obtain a preprocessed feature set that has been standardized and extracted with time-series features;

[0156] The data feature extraction module performs internal dynamic feature analysis on the preprocessed feature set, calculates the instantaneous volatility, mutual coupling degree and covariance relationship of each feature in the time series, extracts the comprehensive deviation features reflecting the self-stability of the new energy power system, and calculates the behavioral deviation distance value.

[0157] The control delivery module compares the acquired behavior deviation distance value with the preset safety trigger threshold, establishes an abnormal trigger set based on the comparison result, and then generates a safety control command by matching it, which is then sent to the new energy power system control terminal for execution.

[0158] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A safety control method for a new energy power system, characterized in that: Includes the following steps: S1. Collect voltage, frequency, and power data during the operation of the new energy power system to construct a set of system operation characteristics; S2. Normalize the system operation feature set to obtain a preprocessed feature set that has been standardized and extracted with time series features; S3. Perform internal dynamic feature analysis on the preprocessed feature set, calculate the instantaneous volatility, mutual coupling degree and covariance relationship of each feature in the time series, extract the comprehensive deviation features reflecting the self-stability of the new energy power system, and calculate the behavioral deviation distance value. S4. Compare the obtained behavior deviation distance value with the preset safety trigger threshold, establish an abnormal trigger set based on the comparison result, and then perform matching to generate a safety control command, which is then sent to the new energy power system control terminal for execution.

2. The safety control method for a new energy power system according to claim 1, characterized in that: S1 includes S11; S11. During the operation of the new energy power system, real-time monitoring signals of the main control node and distributed nodes of the system are collected, including voltage signal sequence, frequency signal sequence and power output signal sequence. The relative deviation of each signal is calculated by signal analysis and time slicing processing, and the following parameters are extracted: voltage deviation rate parameter Vra, frequency disturbance rate parameter Fra and power imbalance rate parameter Pra. The voltage deviation rate parameter Vra is obtained by segmenting the acquired voltage signal into time windows, calculating the difference between the average voltage within the window and the rated voltage of the system, and defining the proportion of the difference to the rated voltage as the voltage deviation rate parameter Vra. The frequency disturbance rate parameter Fra is calculated by taking the sampling time interval as a basis, calculating the absolute value of the frequency change at adjacent time intervals, and then calculating the average change. The ratio of the average change to the rated frequency of the new energy power system is defined as the frequency disturbance rate parameter Fra, which is used to reflect the dynamic fluctuation of the system frequency in a short period. The power imbalance rate parameter Pra is obtained by calculating the difference between the power generation and the load power within the same time period, obtaining the absolute value of the power deviation, and then comparing it with the maximum value of the power generation and the load power. The resulting ratio is defined as the power imbalance rate parameter Pra.

3. The safety control method for a new energy power system according to claim 2, characterized in that: S1 further includes S12; S12. Based on the voltage deviation rate parameter Vra, frequency disturbance rate parameter Fra, and power imbalance rate parameter Pra obtained from multiple nodes, perform data verification and structured integration to obtain the system operation feature set Beh. The data verification and structured integration uses multi-node parameters to check for outliers. When the voltage deviation rate parameter Vra, frequency disturbance rate parameter Fra, or power imbalance rate parameter Pra exceeds the preset constraint range, it is marked as an abnormal sample. The verified parameter data is structured and integrated using time tags as the primary index and node identifiers as the secondary index to form a system operation feature set Beh with temporal continuity and spatial correlation. Each record in the system operation feature set Beh contains three feature parameters and their source node information at the same time, realizing a global quantitative expression of the system operation status.

4. The safety control method for a new energy power system according to claim 3, characterized in that: S2 includes S21; S21. Based on the acquired system operation feature set Beh, the voltage deviation rate parameter Vra, frequency disturbance rate parameter Fra, and power imbalance rate parameter Pra are extracted using the node identifier. Then, the extreme value normalization method is used to scale and map them to the [0,1] interval to eliminate the dimensional differences between different parameters. The normalized voltage normalized parameter Vrn, frequency normalized parameter Frn, and power normalized parameter Prn are obtained and formed into a normalized standard feature group, which is then bound to the node identifier.

5. The safety control method for a new energy power system according to claim 4, characterized in that: S2 further includes S22; S22. Obtain the normalized standard feature group of each node identifier and perform time-series sliding analysis to extract dynamic mean and fluctuation amplitude features, forming a structured preprocessed feature set Bcl. The extraction is completed through steps S221, S222 and S223. S221. Sliding segmentation: Set the sliding window length Win and step size Hop, and continuously segment each node identifier along the time label to form a series of time windows. The data segment in each time window contains three normalized parameters of multiple sampling times, which constitute a subset of the running features of a local time period. S222, Statistical calculation within the window: For each segmented time window, calculate the mean and standard deviation of the three features, and obtain the voltage mean parameter Vrm, frequency mean parameter Frm, power mean parameter Prm, voltage fluctuation parameter Vrs, frequency fluctuation parameter Frs, and power fluctuation parameter Prs; S223. The calculation results of the time window sequence are integrated according to the time window order and node identifier to form a structured preprocessing feature set Bcl.

6. The safety control method for a new energy power system according to claim 5, characterized in that: S3 includes S31; S31. Read the average voltage parameter Vrm, average frequency parameter Frm, and average power parameter Prm of each node in the continuous time window from the preprocessed feature set Bcl, and analyze the characteristic change rate of adjacent time windows. Based on the time window order, compare the same type of feature values ​​in adjacent windows one by one: By comparing the average voltage parameter Vrm between the current time window and the previous time window, the difference in change is calculated, reflecting the instantaneous change in voltage on the time axis. The variation difference between the frequency mean parameter Frm and the power mean parameter Prm is calculated simultaneously to obtain the variation amplitude within a continuous time window; The differences in the changes of the obtained average voltage parameter Vrm, average frequency parameter Frm, and average power parameter Prm are divided by the time span between adjacent windows and then standardized to obtain standardized voltage dynamic fluctuation parameters Vrd, frequency dynamic fluctuation parameters Frd, and power dynamic fluctuation parameters Prd. The voltage dynamic fluctuation parameter Vrd, frequency dynamic fluctuation parameter Frd, and power dynamic fluctuation parameter Prd calculated for each time window are bound to the corresponding node identifier to form a node-level dynamic fluctuation set.

7. A safety control method for a new energy power system according to claim 6, characterized in that: S3 further includes S32; S32. Extract the voltage dynamic fluctuation parameter Vrd, frequency dynamic fluctuation parameter Frd, and power dynamic fluctuation parameter Prd, perform weighted fusion calculation, obtain the index reflecting the disturbance intensity of the new energy power system, and mark it as the behavior deviation distance value Dis; The deviation distance value Dis is obtained by the following formula: ； In the formula, d1, d2 and d3 represent the weighting coefficients of the voltage dynamic fluctuation parameter Vrd, the frequency dynamic fluctuation parameter Frd and the power dynamic fluctuation parameter Prd, respectively, and d1+d2+d3=1, the specific values ​​of which are set by the user.

8. The safety control method for a new energy power system according to claim 7, characterized in that: S4 includes S41; S41. Compare the obtained behavior deviation distance value Dis with the preset safety trigger threshold St, determine the operating status of the new energy power system based on the comparison result, and construct an abnormal trigger set Trig to identify and store trigger conditions and corresponding response instructions. Among them, the safety trigger threshold St is calculated based on the sequence of behavioral deviation distance values ​​Dis within a fixed time window. It reflects the average deviation and fluctuation range of the operation of the new energy power system, and dynamically reflects the self-stabilizing state of the new energy power system. The security trigger threshold St is obtained by the following calculation formula: ； In the formula, μDis represents the moving average of the behavioral deviation distance Dis within a fixed time window; σDis represents the moving standard deviation of the behavioral deviation distance Dis within a fixed time window; K represents the sensitivity coefficient; The exception trigger set Trig is obtained through the following exception state identification method: When the safety trigger threshold St is less than the safety trigger threshold St, the new energy power system is determined to be in a stable state under the current time window. When the safety trigger threshold St is greater than or equal to the safety trigger threshold St, the new energy power system is determined to be in an abnormal state under the current time window. The results of abnormal state identification, time window information, node identifier, voltage dynamic fluctuation parameter Vrd, frequency dynamic fluctuation parameter Frd, and power dynamic fluctuation parameter Prd are stored in the constructed abnormal trigger set Trig.

9. A safety control method for a new energy power system according to claim 8, characterized in that: S4 also includes S42; S42. Extract the voltage dynamic fluctuation parameter Vrd, frequency dynamic fluctuation parameter Frd, and power dynamic fluctuation parameter Prd from the current abnormal time window; sort the real-time values ​​of the three parameters in ascending order using the bubble sort algorithm to obtain a fluctuation intensity sequence from low to high, and select the largest fluctuation intensity sequence as the dominant abnormal feature of the current time window, which is marked as the dominant fluctuation parameter Ddom. Based on the acquired dominant fluctuation parameter Ddom, a set of control instructions Cmd is generated and sent to the control terminal of the new energy power system for execution. The control command set Cmd is generated using the following matching method: When the dominant fluctuation parameter Ddom is the voltage dynamic fluctuation parameter Vrd, a voltage steady-state adjustment instruction is generated, and the node identifier binding instruction of the voltage dynamic fluctuation parameter Vrd is extracted. When the dominant fluctuation parameter Ddom is the frequency dynamic fluctuation parameter Frd, a frequency compensation control instruction is generated, and the node identifier binding instruction of the frequency dynamic fluctuation parameter Frd is extracted. When the dominant fluctuation parameter Ddom is the power dynamic fluctuation parameter Prd, a power balance scheduling instruction is generated, and the node identifier binding instruction of the power dynamic fluctuation parameter Prd is extracted.

10. A safety control system for a new energy power system, applied to the safety control method for a new energy power system as described in any one of claims 1 to 9, characterized in that: It includes a system operation data acquisition module, a system data processing module, a data feature extraction module, and a control distribution module; The system operation data acquisition module collects voltage data, frequency data, and power data during the operation of the new energy power system to construct a set of system operation characteristics; The system data processing module performs normalization preprocessing on the system operation feature set to obtain a preprocessed feature set that has been standardized and extracted with time-series features; The data feature extraction module performs internal dynamic feature analysis on the preprocessed feature set, calculates the instantaneous volatility, mutual coupling degree and covariance relationship of each feature in the time series, extracts the comprehensive deviation features reflecting the self-stability of the new energy power system, and calculates the behavioral deviation distance value. The control delivery module compares the acquired behavior deviation distance value with the preset safety trigger threshold, establishes an abnormal trigger set based on the comparison result, and then generates a safety control command by matching it, which is then sent to the new energy power system control terminal for execution.

Images