A virtual power plant control terminal with edge regulation function
By analyzing the data change characteristics and frequency of the voltage cycle sequence in the virtual power plant control terminal, distinguishing instantaneous changes from real voltage abnormalities, and building an abnormal prediction model, the accuracy of the isolated forest algorithm in voltage regulation is solved, and the power system is improved stability.
Patent Information
- Application Number
- CN202411204001.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-30
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-08-30
AI Technical Summary
When detecting voltage abnormalities, the existing isolated forest algorithm is difficult to accurately distinguish between instantaneous changes and real voltage abnormalities, resulting in inaccurate voltage regulation and affecting the stable operation of the power system.
The data acquisition module obtains the voltage cycle sequence, and the data processing module uses the isolated forest algorithm to obtain the initial abnormal point, and distinguishes instantaneous change from the truth by analyzing the data change degree, mutation coefficient, abnormal frequency, window size, change trend function and historical law coefficients. For voltage abnormality, build an abnormality prediction model for regulation.
It improves the accuracy of voltage regulation, can more accurately identify the real anomalies in the voltage cycle sequence, predict the abnormal moments of the future electricity consumption cycle, and enhances the stability of the power system.
Smart Images

Figure CN118713310B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of data monitoring, and in particular to a virtual power plant control terminal with edge regulation function. Background Art
[0002] A virtual power plant is a system that uses advanced information and communication technology and software systems to achieve the aggregation and coordinated optimization of distributed energy sources such as distributed power sources and energy storage systems. In order to ensure the stability of the power system and the safe operation of power equipment, the virtual power plant needs to perform marginal voltage regulation in the power system so that its changes do not exceed the prescribed allowable range; when the voltage value is abnormal, the voltage of the power system needs to be inspected and regulated. The existing isolation forest algorithm can detect abnormal voltage values in the voltage sequence, but due to the existence of transient interference such as electromagnetic interference, grid harmonics, and electrical equipment switches, the voltage will cause transient abnormal changes. Such changes will automatically recover after the interference disappears, so there is no need for voltage regulation; the isolation forest will identify all voltage anomalies including such transient voltage changes and abnormal voltage fluctuations. This ultimately leads to inaccurate voltage regulation, affecting the stable operation of the power system. Summary of the invention
[0003] In order to solve the technical problem that the voltage anomaly points obtained by the isolation algorithm include different anomaly types, which makes it difficult to accurately control, the purpose of the present invention is to provide a virtual power plant control terminal with edge control function, and the technical solution adopted is as follows:
[0004] A data acquisition module, used to acquire voltage cycle sequences of different historical power consumption cycles;
[0005] A data processing module is used to obtain an initial abnormal point in the voltage cycle sequence by using an isolation forest algorithm; obtain a data change degree according to a neighborhood data change characteristic of a data point in the voltage cycle sequence; and obtain a mutation coefficient of the initial abnormal point according to a difference characteristic of the data change degree between the initial abnormal point and other data points;
[0006] An abnormality analysis module, used to obtain the abnormal frequency at any moment according to the data type characteristics at any moment in all voltage cycle sequences; obtain the window size according to the abnormal frequency of adjacent moments of the arbitrary moment and the preset window length; obtain the change trend function according to the data distribution within the window size in the voltage cycle sequence; obtain the change similarity coefficient at any moment according to the difference characteristics of the change trend function; obtain the historical regularity coefficient of the initial abnormal point according to the change similarity coefficient and the abnormal frequency;
[0007] The voltage control module is used to obtain the final abnormal point according to the mutation coefficient and historical regularity coefficient of the initial abnormal point; construct an abnormal prediction model according to the distribution characteristics of the final abnormal point in the voltage cycle sequence, and predict the abnormal moment of the future power consumption cycle according to the abnormal prediction model and perform regulation.
[0008] Furthermore, the step of obtaining the data variation degree according to the neighborhood data variation characteristics of the data points in the voltage cycle sequence includes:
[0009] The absolute value of the rate of change of the values of any data point and adjacent data points is calculated to obtain a change characteristic value; the average value of the two change characteristic values of the arbitrary data point is calculated to obtain the data change degree of the arbitrary data point.
[0010] Furthermore, the step of obtaining the mutation coefficient of the initial outlier point according to the difference characteristics of the data change degree between the initial outlier point and other data points includes:
[0011] Calculate the average value of the data change degree of all data points in the voltage cycle sequence to obtain the average change degree; calculate the average value of the absolute value of the difference between the data change degree of all data points and the average change degree to obtain the average change difference value; calculate the difference between the data change degree of the initial abnormal point and the average change degree to obtain the change difference value of the initial abnormal point; calculate the difference between the change difference value and the average change difference value and normalize it to obtain the mutation coefficient of the initial abnormal point.
[0012] Furthermore, the step of obtaining the abnormal frequency at any time according to the data type characteristics at any time in all voltage cycle sequences includes:
[0013] The ratio of the number of initial abnormal points at the arbitrary moment to the number of voltage cycle sequences is calculated to obtain the abnormal frequency at the arbitrary moment.
[0014] Furthermore, the step of obtaining the window size according to the abnormal frequency of the adjacent moments of the arbitrary moment and the preset window length comprises:
[0015] Calculate the difference between the maximum value of the preset window length and the minimum value of the preset window length to obtain the window interval; calculate the product of the window interval and the abnormal frequency at the previous moment and the abnormal frequency at the next moment at the said arbitrary moment to obtain the window reference value; calculate the sum of the window reference value and the minimum value of the preset window length to obtain the window size at the said arbitrary moment.
[0016] Further, the step of obtaining a change trend function according to data distribution within the window size in the voltage cycle sequence includes:
[0017] A univariate linear regression equation is calculated by the least square method according to the data distribution within the window size to obtain the change trend function at any time.
[0018] Furthermore, the step of obtaining the change similarity coefficient at any time according to the difference characteristics of the change trend function includes:
[0019] The slope average of the data points in the change trend function is calculated to obtain the average trend characteristic value; the variance of the average trend characteristic value corresponding to all voltage cycle sequences at any time is calculated and negatively correlated to obtain the change similarity coefficient at any time.
[0020] Furthermore, the step of obtaining the historical regularity coefficient of the initial abnormal point according to the change similarity coefficient and the abnormal frequency includes:
[0021] The average value of the abnormal frequencies of all data points within the window size of the initial abnormal point is calculated to obtain the average abnormal characteristic value; the average value of the average abnormal characteristic value and the change similarity coefficient is calculated to obtain the historical regularity coefficient of the initial abnormal point.
[0022] Furthermore, the step of obtaining the final abnormal point according to the mutation coefficient and the historical regularity coefficient of the initial abnormal point includes:
[0023] The mutation coefficient is negatively correlated and mapped to obtain a mutation characteristic value; the average value of the mutation characteristic value and the historical regularity coefficient is calculated to obtain an abnormality index of the initial abnormal point; when the abnormality index of the initial abnormal point exceeds a preset abnormal threshold, the initial abnormal point is the final abnormal point.
[0024] The present invention has the following beneficial effects:
[0025] In the present invention, obtaining the initial abnormal point can determine the data point with relatively abnormal voltage in the voltage cycle sequence; obtaining the data change degree can reflect the voltage change degree at adjacent moments, and obtaining the mutation coefficient of the initial abnormal point according to the data change degree can make a preliminary distinction through the formation characteristics of different initial abnormal points. Obtaining the abnormal frequency can reflect the frequency of occurrence of the initial abnormal point in different voltage cycle sequences at the same time, and then different initial abnormal points can be distinguished. Obtaining the window size can obtain a more accurate voltage data analysis range according to the difference in abnormal frequency, so that the voltage change trend reflected by the change trend function is more accurate; obtaining the voltage change trend can reflect the voltage change trend in the same period in different voltage cycle sequences. The obtained change similarity coefficient can characterize the similarity of the voltage change trend in the same period in different voltage cycle sequences, and then the obtained historical regularity coefficient can be distinguished according to the formation causes of different initial abnormal points. The final abnormal point obtained can accurately reflect the real abnormal voltage fluctuation in the voltage cycle sequence, and the abnormal prediction model finally obtained can more accurately determine the abnormal moment of the future power consumption cycle, thereby improving the accuracy of regulation. BRIEF DESCRIPTION OF THE DRAWINGS
[0026] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0027] Figure 1 A block diagram of a virtual power plant control terminal with edge regulation function provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0028] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation method, structure, features and effects of a virtual power plant control terminal with edge control function proposed by the present invention are described in detail below in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.
[0029] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
[0030] The following describes in detail a specific scheme of a virtual power plant control terminal with edge regulation function provided by the present invention in conjunction with the accompanying drawings.
[0031] See also Figure 1 , which shows a block diagram of a virtual power plant control terminal with edge regulation function provided by an embodiment of the present invention, the terminal includes the following modules:
[0032] The data acquisition module S1 is used to acquire voltage cycle sequences of different historical power consumption cycles.
[0033] In the embodiment of the present invention, the implementation scenario is to analyze the law of abnormal voltage fluctuations in the power system and improve the accuracy of voltage regulation. First, the voltage cycle sequence of different historical power consumption cycles is obtained. Since the power consumption behavior of power-consuming equipment has periodic laws, the change law of voltage data in the same period in different cycles is relatively close, so the voltage anomaly analysis can be performed based on the voltage cycle sequence of different historical power consumption cycles; in the embodiment of the present invention, each day is regarded as a power consumption cycle, and a voltage cycle sequence is obtained every day. The implementer can determine the time period of the power consumption cycle and the frequency of collecting voltage data in the voltage cycle sequence according to the implementation scenario.
[0034] The data processing module S2 is used to obtain the initial abnormal point in the voltage cycle sequence through the isolation forest algorithm; obtain the data change degree according to the neighborhood data change characteristics of the data point in the voltage cycle sequence; and obtain the mutation coefficient of the initial abnormal point according to the difference characteristics of the data change degrees between the initial abnormal point and other data points.
[0035] When the voltage value is not within the normal range, it means that the voltage is relatively abnormal. Therefore, the initial abnormal point in the voltage cycle sequence can be obtained through the isolation forest algorithm. It should be noted that the isolation forest algorithm belongs to the existing technology, and the specific steps are not repeated; the isolation forest algorithm can quickly obtain the initial abnormal point with abnormal voltage value. Due to electromagnetic interference, grid harmonics and electrical equipment switching in the power system, the voltage changes instantaneously and then recovers automatically; at the moment of instantaneous change, the voltage value will exceed the normal voltage range, resulting in detection as an initial abnormal point, but this type of instantaneous voltage change will automatically recover, so there is no need to regulate the voltage, otherwise the regulated voltage will fluctuate. For real voltage anomalies, regulation is required, so it is necessary to analyze the obtained initial abnormal points to distinguish the initial abnormal points of instantaneous changes from the initial abnormal points of real voltage fluctuations.
[0036] For the initial abnormal point of instantaneous change, due to the large instantaneous change degree of voltage, its adjacent moments will show the characteristics of voltage mutation in the voltage cycle sequence; while the real voltage anomaly will show volatility and small mutation characteristics; therefore, the data change degree can be obtained according to the neighborhood data change characteristics of the data point in the voltage cycle sequence, and the initial abnormal point of instantaneous change can be distinguished by the difference in data change degree; preferably, in the embodiment of the present invention, the step of obtaining the data change degree includes: calculating the absolute value of the change rate of the value of any data point and the adjacent data point to obtain the change characteristic value; the arbitrary data point is any voltage data point in the voltage cycle sequence, the adjacent data point is two data points at the adjacent moments before and after the arbitrary data point, the change rate is the ratio of the difference between the voltage values at the two moments and the interval time, when the change characteristic value is larger, it means that the voltage change is more drastic and the mutation characteristics are more obvious. Calculate the average value of the two change characteristic values of any data point to obtain the data change degree of any data point; when the data change degree is larger, the voltage difference between the arbitrary data point and the adjacent moment is larger, the mutation degree is more obvious, and it is more in line with the characteristics of instantaneous change.
[0037] Then, the mutation coefficient of the initial abnormal point can be obtained according to the difference characteristics of the data change degree between the initial abnormal point and other data points; preferably, in the embodiment of the present invention, the step of obtaining the mutation coefficient includes: calculating the average value of the data change degree of all data points in the voltage cycle sequence to obtain the average change degree; the average change degree reflects the average level of the data change degree of all data points in the voltage cycle sequence. Calculate the average value of the absolute value of the difference between the data change degree of all data points and the average change degree to obtain the average change difference value; the average change difference value reflects the average level of the difference between the data change degree of different data points and the average change degree. Calculate the difference between the data change degree of the initial abnormal point and the average change degree to obtain the change difference value of the initial abnormal point; calculate the difference between the change difference value and the average change difference value and normalize it to obtain the mutation coefficient of the initial abnormal point; when the mutation coefficient is larger, it means that the change difference value of the initial abnormal point is larger, the more it exceeds the average level, and the initial abnormal point is more likely to be a voltage abnormal point with instantaneous change. The formula for obtaining the mutation coefficient includes:
[0038]
[0039] In the formula, represents the mutation coefficient of the i-th initial abnormal point in the voltage cycle sequence, represents the data change degree of the i-th initial abnormal point, W represents the average change degree, represents the change difference value of the i-th initial abnormal point, N represents the n-th data point in the voltage cycle sequence, Indicates the data change degree of the nth data point, represents the average change difference value, Represents the normalization function.
[0040] The abnormal analysis module S3 is used to obtain the abnormal frequency at any time according to the data type characteristics at any time in all voltage cycle sequences; obtain the window size according to the abnormal frequency of adjacent moments at any time and the preset window length; obtain the change trend function according to the data distribution within the window size in the voltage cycle sequence; obtain the change similarity coefficient at any time according to the difference characteristics of the change trend function; obtain the historical regularity coefficient of the initial abnormal point according to the change similarity coefficient and the abnormal frequency.
[0041] The instantaneous voltage change caused by electromagnetic interference, grid harmonics, and electrical equipment switching is accidental, and the time when it occurs in different voltage cycle sequences is not fixed, and the frequency of occurrence is low; while the real abnormal voltage fluctuation is caused by human electricity consumption behavior and the operation of large-scale electrical equipment, and the time periods when it occurs in different electricity cycles are relatively similar. Therefore, the frequency of initial abnormal points appearing at the same time in different electricity cycles can be analyzed to distinguish the situations that cause the initial abnormal points to appear, so the abnormal frequency at any time is obtained according to the data type characteristics at any time in all voltage cycle sequences; preferably, in an embodiment of the present invention, the step of obtaining the abnormal frequency includes: calculating the ratio of the number of initial abnormal points at any time to the number of voltage cycle sequences, and obtaining the abnormal frequency at any time; any time is any time point of voltage collection in the voltage cycle sequence, when the number of initial abnormal points at the same arbitrary time in different voltage cycle sequences is more, the abnormal frequency is greater, and the more likely it is a real abnormal voltage fluctuation, when the abnormal frequency is smaller, the more likely it is an accidental voltage instantaneous change.
[0042] Furthermore, the initial abnormal points caused by the real abnormal voltage fluctuation will appear densely within a period of time, and their changing rules are similar in the same period of time in different voltage cycle sequences, while the initial abnormal points caused by instantaneous changes have too large differences in the same period of time in different voltage cycle sequences. Therefore, the voltage data change characteristics of the same period in different cycle sequences can be analyzed to distinguish the initial abnormal points caused by different reasons; but the selection of the period length will affect the judgment result. The shorter period will more obviously reflect the voltage change trend, but it is easily affected by noise data. The longer period can reflect the long-term voltage change trend, but it is not sensitive to short-term fluctuations. Therefore, it is necessary to adaptively adjust the selection of the period length, so the window size is obtained according to the abnormal frequency of the adjacent moments at any time and the preset window length.
[0043] Preferably, in an embodiment of the present invention, the step of obtaining the window size includes: calculating the difference between the maximum value of the preset window length and the minimum value of the preset window length to obtain the window interval; in an embodiment of the present invention, the maximum value of the preset window length is half of the length of the voltage cycle sequence, and the minimum value of the preset window length is 3 acquisition moments. The implementer can determine it by himself according to the implementation scenario, and use the window interval as the initial value of the window size adjustment. Calculate the product of the window interval and the abnormal frequency of the previous moment and the abnormal frequency of the next moment at any moment to obtain the window reference value; when the abnormal frequencies of the previous moment and the next moment at any moment are higher, it means that the real voltage abnormal fluctuation is more likely to occur at any moment, and the longer the duration is, so the larger the window for analysis needs to be, the larger the window reference value is, so that the complete voltage change trend can be obtained; when the abnormal frequencies are lower, it means that the voltage data at any moment may be normal or the instantaneous voltage change situation, and no longer period of time is required for analysis, and the smaller the window reference value is, so as to ensure that the voltage change trend of instantaneous change can be reflected. The sum of the window reference value and the minimum value of the preset window length is calculated to obtain the window size at any time; the window size is the time period length in the voltage cycle sequence, and the larger the window size value, the longer the time period length involved in the calculation.
[0044] After determining the window size required for calculation at different moments, the change trend function can be obtained in the voltage cycle sequence according to the data distribution within the window size. Preferably, in an embodiment of the present invention, the step of obtaining the change trend function includes: calculating a univariate linear regression equation by the least square method according to the data distribution within the window size to obtain the change trend function at any moment. In the voltage cycle sequence, any moment is at the center of the window size range, that is, the change trend function is obtained by fitting the data within the length range of the window size with any moment as the center. When any moment does not meet the central position of the window size, the data within the window size closest to the center position is used for calculation. The change trend function characterizes the voltage change trend within the window size range corresponding to the arbitrary moment, and then the type of the initial abnormal point can be analyzed according to the difference characteristics of the change trend function at the same moment in different voltage cycle sequences; therefore, the change similarity coefficient at any moment is obtained according to the difference characteristics of the change trend function.
[0045] Preferably, in an embodiment of the present invention, the step of obtaining the change similarity coefficient includes: calculating the average slope of the data points in the change trend function to obtain the average trend characteristic value; it should be noted that the slope of the data points in the change trend function is the slope of the data point and the previous data point obtained by the existing point slope method; the average trend characteristic value reflects the voltage change trend of the time period corresponding to the arbitrary moment in a voltage cycle sequence. Calculate the variance of the average trend characteristic value corresponding to the arbitrary moment in all voltage cycle sequences and negatively map them to obtain the change similarity coefficient at any moment; when the variance is smaller, it means that the voltage change trends at the arbitrary moment in different voltage cycle sequences are more similar, and the change similarity coefficient is larger; when the variance is larger, it means that the voltage change trends at the arbitrary moment in different voltage cycle sequences are more dissimilar.
[0046] Further, after obtaining the abnormal frequency and the change similarity coefficient, the historical regularity coefficient of the initial abnormal point can be obtained according to the change similarity coefficient and the abnormal frequency; preferably, in the embodiment of the present invention, the step of obtaining the historical regularity coefficient includes: calculating the average value of the abnormal frequency of all data points within the window size of the initial abnormal point to obtain the average abnormal characteristic value; when the average abnormal characteristic value is larger, it means that the frequency of voltage abnormalities in different voltage cycle sequences within the window size range is higher, and the initial abnormal point is more likely to be caused by real voltage fluctuations. Calculate the average value of the average abnormal characteristic value and the change similarity coefficient to obtain the historical regularity coefficient of the initial abnormal point; when the change similarity coefficient is higher, it means that the voltage change trend at the time of the initial abnormal point in different historical voltage cycle sequences is the same; therefore, when the historical regularity coefficient is larger, the initial abnormal point is more likely to be caused by real voltage fluctuations.
[0047] The voltage control module S4 is used to obtain the final abnormal point according to the mutation coefficient and historical regularity coefficient of the initial abnormal point; construct an abnormal prediction model according to the distribution characteristics of the final abnormal point in the voltage cycle sequence, and predict the abnormal moment of the future power consumption cycle according to the abnormal prediction model and perform regulation.
[0048] After obtaining the mutation coefficient and historical regularity coefficient of the initial abnormal point, the final abnormal point can be obtained according to the mutation coefficient and historical regularity coefficient of the initial abnormal point. Preferably, in the embodiment of the present invention, the step of obtaining the final abnormal point includes: mapping the mutation coefficient in a negative correlation to obtain the mutation characteristic value; when the mutation coefficient is smaller, the mutation characteristic value is larger, and the initial abnormal point is more likely to be caused by real voltage fluctuations. Calculate the average value of the mutation characteristic value and the historical regularity coefficient to obtain the abnormality index of the initial abnormal point; when the abnormality index is larger, it means that the initial abnormal point is more likely to be caused by real voltage fluctuations, and less likely to be caused by instantaneous voltage changes. When the abnormality index of the initial abnormal point exceeds the preset abnormality threshold, the initial abnormal point is the final abnormal point; the final abnormal point reflects the real abnormal voltage fluctuation, and then the voltage needs to be regulated at the moment of the final abnormal point; in the embodiment of the present invention, the preset abnormality threshold is 0.8, which can be determined by the implementer according to the implementation scenario. The formula for obtaining the abnormality index includes:
[0049]
[0050] In the formula, represents the abnormal index of the i-th initial abnormal point in the voltage cycle sequence, represents the mutation coefficient of the i-th initial abnormal point, represents an exponential function with a natural constant as base, represents the mutation feature value, Represents the historical regularity coefficient of the i-th initial abnormal point.
[0051] Further, after determining the time of the final abnormal point in the voltage cycle sequence, an abnormal prediction model can be constructed according to the distribution characteristics of the final abnormal point in the voltage cycle sequence, and the abnormal time of the future power consumption cycle can be predicted and regulated according to the abnormal prediction model; it should be noted that the construction of an abnormal prediction model through a neural network belongs to the prior art, and the specific steps are not repeated. The abnormal prediction model can determine the time of the real abnormal voltage fluctuation in the future voltage cycle sequence, so as to assist technical personnel in the field to perform high-frequency detection and timely regulation at this moment, improve the accuracy of regulation, and thus ensure the stable operation of the power system. The regulation process is determined by the implementer according to the implementation scenario. So far, by analyzing all the initial abnormal points obtained by the isolation forest algorithm, the characteristic differences of the two types of initial abnormal points caused by instantaneous changes and abnormal voltage fluctuations are distinguished, so as to obtain a more accurate final abnormal point, and the neural network model is trained and constructed according to the distribution characteristics of the final abnormal point at the moment, so as to predict the future abnormal voltage fluctuation moment that is prone to high frequency, so as to facilitate timely detection and regulation.
[0052] In summary, the embodiment of the present invention provides a virtual power plant control terminal with edge control function; obtains the initial abnormal point in the voltage cycle sequence through the isolation forest algorithm; obtains the mutation coefficient according to the difference characteristics of the data change degree between the initial abnormal point and other data points; obtains the abnormal frequency according to the data type characteristics at any time in all voltage cycle sequences; obtains the window size according to the abnormal frequency and the preset window length; obtains the change trend function according to the data distribution within the window size; obtains the change similarity coefficient at any time according to the difference characteristics of the change trend function. The present invention obtains the final abnormal point according to the change similarity coefficient, abnormal frequency and mutation coefficient; constructs an abnormal prediction model according to the distribution characteristics of the final abnormal point in the voltage cycle sequence and regulates the abnormal moments of the future power consumption cycle, thereby improving the accuracy of voltage regulation.
[0053] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0054] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.
Claims
1. A virtual power plant control terminal with voltage regulation function, characterized in that: The terminal includes the following modules: A data acquisition module, used to acquire voltage cycle sequences of different historical power consumption cycles; A data processing module is used to obtain an initial abnormal point in the voltage cycle sequence by using an isolation forest algorithm; obtain a data change degree according to a neighborhood data change characteristic of a data point in the voltage cycle sequence; and obtain a mutation coefficient of the initial abnormal point according to a difference characteristic of the data change degree between the initial abnormal point and other data points; An abnormality analysis module, used to obtain the abnormal frequency at any moment according to the data type characteristics at any moment in all voltage cycle sequences; obtain the window size according to the abnormal frequency of adjacent moments of the arbitrary moment and the preset window length; obtain the change trend function according to the data distribution within the window size in the voltage cycle sequence; obtain the change similarity coefficient at any moment according to the difference characteristics of the change trend function; obtain the historical regularity coefficient of the initial abnormal point according to the change similarity coefficient and the abnormal frequency; A voltage control module is used to obtain a final abnormal point according to the mutation coefficient and the historical regularity coefficient of the initial abnormal point; construct an abnormal prediction model according to the distribution characteristics of the final abnormal point in the voltage cycle sequence, and predict the abnormal moment of the future power consumption cycle according to the abnormal prediction model and perform regulation; The step of obtaining the mutation coefficient of the initial outlier point according to the difference characteristics of the data change degree between the initial outlier point and other data points comprises: Calculate the average value of the data change degree of all data points in the voltage cycle sequence to obtain the average change degree; calculate the average value of the absolute value of the difference between the data change degree of all data points and the average change degree to obtain the average change difference value; calculate the difference between the data change degree of the initial abnormal point and the average change degree to obtain the change difference value of the initial abnormal point; calculate the difference between the change difference value and the average change difference value and normalize it to obtain the mutation coefficient of the initial abnormal point; The step of obtaining the change similarity coefficient at any time according to the difference characteristics of the change trend function comprises: Calculate the average slope of the data points in the change trend function to obtain the average trend characteristic value; calculate the variance of the average trend characteristic value corresponding to all voltage cycle sequences at any time and negatively map them to obtain the change similarity coefficient at any time; The step of obtaining the historical regularity coefficient of the initial abnormal point according to the change similarity coefficient and the abnormal frequency comprises: The average value of the abnormal frequencies of all data points within the window size of the initial abnormal point is calculated to obtain the average abnormal characteristic value; the average value of the average abnormal characteristic value and the change similarity coefficient is calculated to obtain the historical regularity coefficient of the initial abnormal point.
2. A virtual power plant control terminal with voltage regulation function according to claim 1, characterized in that: The step of obtaining the data variation degree according to the neighborhood data variation characteristics of the data points in the voltage cycle sequence comprises: The absolute value of the rate of change of the values of any data point and adjacent data points is calculated to obtain a change characteristic value; the average value of the two change characteristic values of the arbitrary data point is calculated to obtain the data change degree of the arbitrary data point.
3. The virtual power plant control terminal with voltage regulation function according to claim 1, characterized in that: The step of obtaining the abnormal frequency at any time according to the data type characteristics at any time in all voltage cycle sequences comprises: The ratio of the number of initial abnormal points at the arbitrary moment to the number of voltage cycle sequences is calculated to obtain the abnormal frequency at the arbitrary moment.
4. The virtual power plant control terminal with voltage regulation function according to claim 1, characterized in that: The step of obtaining the window size according to the abnormal frequency of the adjacent moments of the arbitrary moment and the preset window length comprises: Calculate the difference between the maximum value of the preset window length and the minimum value of the preset window length to obtain the window interval; calculate the product of the window interval and the abnormal frequency at the previous moment and the abnormal frequency at the next moment at the said arbitrary moment to obtain the window reference value; calculate the sum of the window reference value and the minimum value of the preset window length to obtain the window size at the said arbitrary moment.
5. The virtual power plant control terminal with voltage regulation function according to claim 1, characterized in that: The step of obtaining a change trend function according to data distribution within the window size in the voltage cycle sequence comprises: A univariate linear regression equation is calculated by the least square method according to the data distribution within the window size to obtain the change trend function at any time.
6. The virtual power plant control terminal with voltage regulation function according to claim 1, characterized in that: The step of obtaining the final abnormal point according to the mutation coefficient and the historical regularity coefficient of the initial abnormal point comprises: The mutation coefficient is negatively correlated and mapped to obtain a mutation characteristic value; the average value of the mutation characteristic value and the historical regularity coefficient is calculated to obtain an abnormality index of the initial abnormal point; when the abnormality index of the initial abnormal point exceeds a preset abnormal threshold, the initial abnormal point is the final abnormal point.
Citation Information
Patent Citations
Intelligent inspection method for power distribution equipment based on artificial intelligence
CN116702081A
Power supply output power failure abnormity early warning method based on data driving
CN117007979A
Power data monitoring method and system based on big data
CN117391481A
Operation control method of virtual power plant
CN117394346A
Coordination control method of power grid system
CN117595391A
