Adaptive control method for low-frequency oscillation of power system using variable-speed pumped storage units

Abstract

The present invention relates to a method for adaptively controlling low-frequency oscillations in a variable-speed pumped-storage unit power system, and relates to the field of power control. The method comprises: processing power system operating data and monitored power generation of the pumped-storage unit through a pattern recognizer to output a first low-frequency oscillation pattern; when the low-frequency abnormal oscillation pattern is reached, accessing a turbine mechanical energy prediction network and, in conjunction with the pattern recognizer, executing an optimization algorithm on the unit control parameters to obtain a recommended unit guide vane opening, a recommended unit operating frequency, and a recommended unit speed set point; controlling the turbine based on the recommended unit guide vane opening and the recommended unit operating frequency, and controlling the AC excitation device based on the recommended unit speed set point. This method solves the technical problem of weak adaptability in existing technologies, which relies on the adjustment of physical components to achieve low-frequency oscillation control.

Description

Technical Field

[0001] The present invention relates to the field of power control, and in particular to a method and system for adaptively controlling low-frequency oscillations in a power system of a variable-speed pumped storage unit. Background Art

[0002] Low-frequency oscillations are a common phenomenon in power system operation, particularly in hydropower systems. These oscillations are often associated with dynamic changes in power generation conditions and can negatively impact power system stability and efficiency. Traditional methods for suppressing low-frequency oscillations rely on adjusting the damping properties of the system, but these approaches are often costly and have limited stability and adaptability in different scenarios.

[0003] Existing technologies for addressing low-frequency oscillations typically require adjustments to the physical components of the power system, which is not only costly but can also require complex engineering implementation. Furthermore, these methods may lack adaptability to the oscillation characteristics of different power system operating conditions, resulting in poor results in some cases. Summary of the Invention

[0004] The present invention aims to solve the technical problem that the existing technology relies on the adjustment of physical components to achieve low-frequency oscillation control, resulting in weak adaptability, and provides a variable-speed pumped storage unit power system low-frequency oscillation adaptive control method and system to solve the problem.

[0005] The technical solution of the present invention to solve the above technical problems is as follows:

[0006] In a first aspect, the present invention provides a method for adaptively controlling low-frequency oscillations in a power system of a variable-speed pumped storage unit, comprising:

[0007] Monitor power system operation data and energy storage unit power generation;

[0008] Processing the power system operation data and the energy storage unit power generation monitoring data through a pattern recognizer to output a first low-frequency oscillation pattern, wherein the first low-frequency oscillation pattern has a first oscillation amplitude tag, a first oscillation duration tag, and a first vibration frequency tag;

[0009] When the first oscillation amplitude tag, the first oscillation duration tag, and the first vibration frequency tag belong to a low-frequency abnormal oscillation mode, calling a turbine mechanical energy prediction network, combining the pattern recognizer, and executing an optimization algorithm on the unit control parameters to obtain a recommended unit guide vane opening, a recommended unit operating frequency, and a recommended unit speed set point;

[0010] The turbine is controlled according to the recommended unit guide vane opening and the recommended unit operating frequency, and the AC excitation device is controlled according to the recommended unit speed set point.

[0011] In a second aspect, the present application provides an electronic device, comprising:

[0012] Memory for storing computer software programs;

[0013] The processor is used to read and execute the computer software program, thereby implementing the low-frequency oscillation adaptive control method of the variable-speed pumped storage unit power system described in the first aspect.

[0014] In a third aspect, the present application provides a non-transitory computer-readable storage medium, in which a computer software program is stored. When the computer software program is executed by a processor, the method for adaptively controlling low-frequency oscillations of a variable-speed pumped-storage unit power system described in the first aspect is implemented.

[0015] In a fourth aspect, the present application provides a computer program product, comprising a computer software program, which, when executed by a processor, implements the method for adaptively controlling low-frequency oscillations of a variable-speed pumped-storage unit power system as described in the first aspect.

[0016] The beneficial effects of the present invention are: by providing real-time monitoring of the operation data of the power system and the power generation monitoring power of the energy storage unit, the collected data is processed by a pattern recognizer, and a first low-frequency oscillation pattern with oscillation amplitude, duration and frequency labels is output. When the identified oscillation pattern meets the low-frequency abnormal oscillation pattern, the system will automatically call the turbine mechanical energy prediction network, combine the output of the pattern recognizer and the mechanical energy prediction network, execute the optimization algorithm, and calculate the recommended unit guide vane opening, operating frequency and speed setting point to optimize the operating state of the energy storage unit and effectively suppress the low-frequency oscillation technical solution. By combining artificial intelligence technology, the oscillation state is detected in real time, and the power generation control parameters are adjusted in time to achieve low-frequency oscillation adaptive control. It can be migrated to different hydropower generation scenarios for application, achieving a technical effect with strong applicability. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments of the present application or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 A flow chart of a method for adaptively controlling low-frequency oscillations in a power system of a variable-speed pumped-storage unit provided by the present invention;

[0019] Figure 2A flow chart showing the steps of loading the power generation efficiency coefficient in the low-frequency oscillation adaptive control method for a variable-speed pumped storage unit power system provided by the present invention;

[0020] Figure 3 A schematic structural diagram of the electronic device provided by the present invention;

[0021] Figure 4 A schematic structural diagram of a non-transitory computer-readable storage medium provided by the present invention.

[0022] In the accompanying drawings, the components represented by the reference numerals are as follows:

[0023] Electronic device 500 , memory 510 , processor 520 , computer software program 511 , non-transitory computer-readable storage medium 600 , computer software program 611 . DETAILED DESCRIPTION

[0024] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts shall fall within the scope of protection of the present invention.

[0025] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of the technical features indicated. Therefore, a feature specified as "first" or "second" may explicitly or implicitly include one or more of the specified features. In the description of the present invention, "plurality" means two or more, unless otherwise specifically defined.

[0026] In the description of the present invention, the term "for example" is used to mean "used as an example, illustration or illustration". Any embodiment of the present invention described as "for example" is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any person skilled in the art to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes are not elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed herein.

[0027] Example 1:

[0028] like Figure 1As shown, the embodiment of the present invention provides a method for adaptively controlling low-frequency oscillations in a power system of a variable-speed pumped storage unit, comprising the steps of:

[0029] S10: monitor the power system operation data and the power generation of the energy storage unit;

[0030] In the embodiments of this application, power system operating data refers to parameters related to the power system's operating status, such as voltage, current, power, and frequency. Energy storage unit power monitoring refers to data obtained by real-time monitoring of the power generation (energy generated per unit time) of pumped storage units in the power system during operation. Real-time monitoring enables rapid response to any changes in the power system, providing immediate data support for subsequent control decisions.

[0031] For example, assuming a variable-speed pumped storage unit is in operation, monitoring steps might include installing current and voltage sensors to collect power system operating data. Using a power meter to monitor the unit's real-time power generation. A data acquisition system aggregates this data in real time and transmits it to a central monitoring system.

[0032] S20: Processing the power system operation data and the energy storage unit power generation monitoring data through a pattern recognizer to output a first low-frequency oscillation pattern, wherein the first low-frequency oscillation pattern has a first oscillation amplitude tag, a first oscillation duration tag, and a first vibration frequency tag;

[0033] Furthermore, in step S20, before processing the power system operation data and the energy storage unit power generation monitoring power by a pattern recognizer and outputting the first low-frequency oscillation pattern, the following steps are included to configure a unique pattern recognizer for the power system:

[0034] S21: performing correlation sorting on the power system operation parameter attributes to obtain selected power system operation data attributes;

[0035] S22: Based on the power system topology and the energy storage unit model, perform low-frequency oscillation state mining on the selected power system operation data attributes and energy storage unit power generation attributes to obtain power system operation record data, energy storage unit power generation record data, and low-frequency oscillation pattern record data; wherein the low-frequency oscillation pattern record data includes at least historical oscillation amplitude, historical oscillation duration, and historical oscillation frequency;

[0036] Furthermore, limited to the power system topology and energy storage unit model, low-frequency oscillation state mining is performed on the selected power system operation data attributes and energy storage unit power generation attributes, which means that according to the corresponding power system topology and energy storage unit model are consistent with the power system currently performing low-frequency oscillation adaptive control, low-frequency oscillation state mining is performed on the selected power system operation data attributes and energy storage unit power generation attributes.

[0037] S23: Retrieving the power system operation record data, the energy storage unit power generation record data and the low-frequency oscillation pattern record data, and configuring the pattern identifier uniquely associated with the power system.

[0038] Furthermore, step S23, retrieving the power system operation record data, the energy storage unit power generation record data and the low-frequency oscillation pattern record data, and configuring the pattern identifier uniquely associated with the power system, includes the following steps:

[0039] S231: Construct loss function:

[0040]

[0041] in, Characterize each training The output loss after the second Characterize each The first training The output oscillation amplitude of the training, Characterize each The first training The output oscillation amplitude supervision value of the training, that is, The historical oscillation amplitude in the training time, Characterize each The first training The output oscillation duration of the training, Characterize each The first training The output oscillation duration supervision value of the training, that is, The length of historical oscillations in training, Characterize each The first training The output oscillation frequency of the training, Characterize each The first training The output oscillation frequency supervision value of the training, that is, The historical oscillation frequency in the training time, is a small constant, Respectively represent the weight parameters;

[0042] S232: Retrieving the power system operation record data, the energy storage unit power generation record data and the low-frequency oscillation mode record data through the loss function, and configuring the pattern identifier uniquely associated with the power system; wherein, retrieving the power system operation record data, the energy storage unit power generation record data and the low-frequency oscillation mode record data through the loss function means processing the power system operation record data, the energy storage unit power generation record data and the low-frequency oscillation mode record data through the loss function.

[0043] In an embodiment of the present application, a pattern recognizer is a functional module for analyzing specific state data of low-frequency oscillations. The first low-frequency oscillation pattern is the output data of the pattern recognizer, i.e., state data representing the low-frequency oscillation, and includes at least oscillation amplitude, oscillation duration, and oscillation frequency, specifically a first oscillation amplitude tag, a first oscillation duration tag, and a first oscillation frequency tag. A subsequent step can determine whether the low-frequency oscillation state is abnormal or normal by analyzing the first oscillation amplitude tag, the first oscillation duration tag, and the first oscillation frequency tag, thereby facilitating control and regulation of the variable-speed pumped-storage unit.

[0044] Preferably, the pattern recognizer is a neural network model. Whenever a variable-speed pumped-storage unit is connected to a new power system, a dedicated pattern recognizer needs to be configured for the power system. When the hardware and circuit topology remain unchanged, the influencing parameters for the low-frequency oscillation mode only include the power generation power of the energy storage unit and the power system operating parameters. Therefore, at this time, low-frequency oscillation monitoring data of similar scenarios with the same hardware, circuit topology, and variable-speed pumped-storage unit model can be collected. Each set of monitoring data includes at least power system operation record data, energy storage unit power generation record data, and low-frequency oscillation mode record data. The low-frequency oscillation mode record data includes at least oscillation amplitude record data, oscillation duration record data, and oscillation frequency record data.

[0045] Preferably, the output data is set as a three-element array, i.e., a low-frequency oscillation mode, where the three elements are oscillation amplitude, oscillation duration, and oscillation frequency, respectively. The specific location is not limited here. Two input nodes are configured for the neural network model. The first input node is used to receive the power generation data of the energy storage unit, and the second input node is used to receive the array composed of power system operating parameters. Each attribute of the power system operating parameters has a fixed position in the input array. Furthermore, the topology of the neural network model can adopt the BP neural network topology suitable for regression problems.

[0046] The detailed training process of the pattern recognizer is as follows:

[0047] Step 1: Through correlation analysis, determine the power system operation parameter attributes that are more relevant to the low-frequency oscillation mode and set them as the selected power system operation data attributes. The details are as follows:

[0048] Collect low-frequency oscillation monitoring data with the same power topology and energy storage unit model, including low-frequency oscillation amplitude monitoring data set, low-frequency oscillation duration monitoring data set and low-frequency oscillation frequency monitoring data set, and the first attribute state value monitoring data set, the second attribute state value monitoring data set until the Attribute state value monitoring data set, in the above data, Characterizes the total number of power system operation parameter attributes to be sorted that are pre-configured by the user, and the low-frequency oscillation amplitude monitoring data set, the low-frequency oscillation duration monitoring data set and the low-frequency oscillation frequency monitoring data set, and the first attribute state value monitoring data set, the second attribute state value monitoring data set until the The attribute state value monitoring data sets are all one-to-one corresponding. For example, if the low-frequency oscillation amplitude monitoring data set has 10 data, the first attribute state value monitoring data set will have 10 data corresponding to the 10 data of the low-frequency oscillation amplitude monitoring data set. The same applies to other sets.

[0049] Furthermore, the low frequency oscillation amplitude monitoring data set is normalized to obtain a first benchmark sequence, and the first attribute state value monitoring data set and the second attribute state value monitoring data set are traversed until the first benchmark sequence is obtained. The attribute state value monitoring data sets are normalized to obtain the first comparison sequence, the second comparison sequence, and the third comparison sequence. Comparison sequence; then take the first reference sequence as the first column of data, the first comparison sequence, the second comparison sequence until the The comparison sequence is from the second column to the Column data, construct the grey relational analysis matrix, the form is as follows:

[0050]

[0051] in, data characterizing a first reference sequence, Characterization Compare sequences, Characterize the total number of data sets (that is, the total number of data in each data set), is an integer.

[0052] Furthermore, we can use the grey relational analysis formula 1:

[0053]

[0054] Grey relational analysis formula 2:

[0055]

[0056] in, Characterization Liedi Row value correlation coefficient, Characterization The baseline value of the row, Characterization The first Column values, Characterization resolution coefficient, the default is 0.5, Characterization The column represents the correlation between the attribute and the low frequency oscillation state. If the correlation is greater than or equal to the user-preset correlation threshold, the first The columns represent the attributes added into the selected power system operation data attributes. Through the grey correlation analysis formula 1 and grey correlation analysis formula 2, the first attribute to the second attribute are analyzed. Attributes are obtained by analysing the selected power system operation data attributes. The low-frequency oscillation duration monitoring dataset and the low-frequency oscillation frequency monitoring dataset are then replaced with the low-frequency oscillation amplitude monitoring dataset within the processing flow for grey relational analysis (i.e., the low-frequency oscillation duration monitoring dataset and the low-frequency oscillation frequency monitoring dataset are processed accordingly, referring to the grey relational analysis of the low-frequency oscillation amplitude monitoring dataset). This results in the final selected power system operation data attributes. Through grey relational analysis, the operating parameters with the greatest impact are identified for subsequent processing, avoiding redundant attribute data from participating in the calculation, which could reduce accuracy and efficiency.

[0057] Step 2: Collect historical data for training, as follows: collect historical data limited to the power system topology and energy storage unit model, that is, the power system topology and energy storage unit model of the power system of the collected data must be consistent with the power system topology and energy storage unit model of the power system currently performing low-frequency oscillation adaptive control; perform data mining based on big data to obtain power system operation record data, energy storage unit power generation record data and low-frequency oscillation mode record data.

[0058] Step 3: Construct loss function:

[0059]

[0060] in, Characterize each training The output loss after the second Characterize each The first training The output oscillation amplitude of the training, Characterize each The first training The output oscillation amplitude supervision value of the training, Characterize each The first training The output oscillation duration of the training, Characterize each The first training The output oscillation duration supervision value of the training, Characterize each The first training The output oscillation frequency of the training, Characterize each The first training The output oscillation frequency supervision value of the training, is a small constant, Represent the weight parameters respectively.

[0061] The user configures the loss function threshold. The power system operation record data is converted into an input array, and the low-frequency oscillation pattern record data is converted into an output supervision array. The input array and the energy storage unit power generation record data are used as the input of the BP neural network, and the output supervision array is used as the output data supervision truth value. When the output value of the loss function is less than or equal to the loss function threshold, multiple sets of data that have not participated in training are called for verification. If the output value of the loss function is still less than or equal to the loss function threshold, it is considered converged, and the pattern recognizer uniquely associated with the power system is output.

[0062] S30: When the first oscillation amplitude tag, the first oscillation duration tag, and the first vibration frequency tag belong to a low-frequency abnormal oscillation mode, calling a turbine mechanical energy prediction network, and combining the pattern recognizer, executing an optimization algorithm on the unit control parameters to obtain a recommended unit guide vane opening, a recommended unit operating frequency, and a recommended unit speed set point;

[0063] Furthermore, in step S30, when the first oscillation amplitude tag, the first oscillation duration tag, and the first vibration frequency tag belong to a low-frequency abnormal oscillation mode, the turbine mechanical energy prediction network is retrieved, and in combination with the pattern recognizer, an optimization algorithm is executed on the unit control parameters to obtain a recommended unit guide vane opening, a recommended unit operating frequency, and a recommended unit speed set point, including the following steps:

[0064] S31: Collecting historical power system oscillation accidents based on the power system topology and energy storage unit model, wherein the historical power system oscillation accidents include recorded data on oscillation amplitude, recorded data on oscillation duration, and recorded data on oscillation frequency; wherein the historical power system oscillation accidents refer to low-frequency abnormal oscillation patterns that have occurred;

[0065] S32: performing a centralized value analysis on the oscillation amplitude recorded data to obtain a centralized oscillation amplitude;

[0066] S33: When the concentrated oscillation amplitude is less than the rated oscillation amplitude, setting the concentrated oscillation amplitude to an oscillation amplitude threshold; otherwise, setting the rated oscillation amplitude to the oscillation amplitude threshold;

[0067] S34: performing concentrated value analysis on the oscillation duration record data to obtain concentrated oscillation duration;

[0068] S35: When the concentrated oscillation duration is less than the rated oscillation duration, setting the concentrated oscillation duration as the oscillation duration threshold; otherwise, setting the rated oscillation duration as the oscillation duration threshold;

[0069] S36: performing cluster analysis on the oscillation frequency recording data to obtain a first oscillation frequency characteristic interval and a second oscillation frequency characteristic interval;

[0070] S37: generating a first oscillation frequency threshold interval according to an intersection of a first rated oscillation frequency interval and the first oscillation frequency characteristic interval;

[0071] S38: generating a second oscillation frequency threshold interval according to an intersection of a second rated oscillation frequency interval and the second oscillation frequency characteristic interval of the rated oscillation frequency;

[0072] S39: When the first oscillation amplitude label is greater than or equal to the oscillation amplitude threshold, or / and the first oscillation duration label is greater than or equal to the oscillation duration threshold, or / and the first vibration frequency label belongs to the first oscillation frequency threshold interval, or / and the first vibration frequency label belongs to the second oscillation frequency threshold interval, the first low-frequency oscillation mode belongs to the low-frequency abnormal oscillation mode.

[0073] In an embodiment of the present application, the first oscillation amplitude label, the first oscillation duration label, and the first vibration frequency label can be used to determine whether the current low-frequency oscillation mode belongs to the low-frequency abnormal oscillation mode. Then, if the current low-frequency oscillation mode belongs to the low-frequency abnormal oscillation mode, the unit control parameters are configured through the turbine mechanical energy prediction network and the pattern identifier until the recommended unit guide vane opening, recommended unit operating frequency, and recommended unit speed setting point are obtained when the low-frequency oscillation mode does not belong to the low-frequency abnormal oscillation mode. Among them, the unit guide vane opening refers to a device for controlling water flow. The larger the opening, the greater the water flow, and vice versa; the unit operating frequency represents a parameter of mechanical energy output power; the unit speed setting point represents the expected speed of the turbine; the turbine mechanical energy prediction network is a functional module for predicting the mechanical energy that may be generated by the turbine based on the unit guide vane opening, the unit operating frequency, and the unit speed setting point.

[0074] When the low-frequency abnormal oscillation mode is triggered, the unit guide vane opening, unit operating frequency, and unit speed set point are adjusted. The mechanical energy produced per unit time is then predicted using the turbine mechanical energy prediction network. The power generation per unit time of the energy storage unit is then determined based on the conversion efficiency of mechanical energy and electricity, thereby obtaining the predicted power generation capacity of the energy storage unit. The predicted power generation capacity of the energy storage unit and the power system operating data are then analyzed using a pattern recognizer to output the low-frequency oscillation prediction mode. If the low-frequency oscillation prediction mode does not correspond to the low-frequency abnormal oscillation mode, the updated unit guide vane opening, unit operating frequency, and unit speed set point are set as the recommended unit guide vane opening, recommended unit operating frequency, and recommended unit speed set point for controlling the unit, thereby suppressing low-frequency oscillations at the source and ensuring power grid security.

[0075] Furthermore, in an embodiment of the present application, an exemplary process for analyzing whether the first oscillation amplitude label, the first oscillation duration label, and the first vibration frequency label belong to a low-frequency abnormal oscillation mode is provided:

[0076] First, limited to the power system topology and energy storage unit model, a data set of abnormal low-frequency oscillation modes was collected, namely the historical power system oscillation accidents. Among them, each historical power system oscillation accident includes oscillation amplitude record data, oscillation duration record data and oscillation frequency record data.

[0077] Step 1: Perform centralized value analysis on all oscillation amplitude recording data to obtain centralized oscillation amplitude. The centralized value analysis process is: determine the outlier oscillation amplitude of all oscillation amplitude recording data through the LOF outlier analysis algorithm and delete them, calculate the mean of the retained oscillation amplitude, and set it as the centralized oscillation amplitude. The centralized oscillation amplitude can represent the oscillation amplitude value of most low-frequency oscillation accidents. If the oscillation amplitude values ​​of frequently occurring low-frequency oscillation accidents are all less than the rated oscillation amplitude, it means that the strictness of the rated oscillation amplitude preset by the user is insufficient, and most low-frequency oscillation accidents cannot be sorted out. Since only when the oscillation amplitude is greater than the oscillation amplitude threshold will it be judged as an abnormal low-frequency oscillation amplitude, when the oscillation amplitude values ​​of most low-frequency oscillation accidents cannot be sorted out by the rated oscillation amplitude, it is unreliable based on the rated oscillation amplitude. At this time, it is necessary to set the centralized oscillation amplitude as the oscillation amplitude threshold, otherwise the rated oscillation amplitude is set as the oscillation amplitude threshold.

[0078] Step 2: Perform centralized value analysis on the oscillation duration record data to obtain the centralized oscillation duration; wherein, the process of performing centralized value analysis on the oscillation duration record data is similar to the process of performing centralized value analysis on the oscillation amplitude record data, and will not be repeated here; when the centralized oscillation duration is less than the rated oscillation duration, the centralized oscillation duration is set as the oscillation duration threshold, otherwise the rated oscillation duration is set as the oscillation duration threshold. Based on the same centralized value analysis algorithm, the centralized oscillation duration representing most low-frequency oscillation accidents is calculated. Since the longer the oscillation duration, the more unfavorable it is to the power system, therefore, when the centralized oscillation duration is less than the rated oscillation duration, the centralized oscillation duration is set as the oscillation duration threshold, otherwise the rated oscillation duration is set as the oscillation duration threshold.

[0079] Step 3: Cluster analysis is performed on the oscillation frequency recording data to obtain a first oscillation frequency characteristic interval and a second oscillation frequency characteristic interval. Based on the oscillation frequency threshold preset by the user, cluster analysis is performed on the oscillation frequency recording data. If the frequency deviation of any two oscillation frequency recording data is greater than or equal to the oscillation frequency threshold, they are clustered into two categories. Otherwise, they are clustered into one category, obtaining multiple clusters of oscillation frequency recording data. Clusters with a number less than the number threshold within the category are deleted, as such clusters are not representative. Since oscillation frequencies that are too high or too low are detrimental to the power system, low-frequency oscillation accidents are distributed on both sides of the normal oscillation frequency interval. The clusters analyzed through big data are also distributed on both sides of the normal oscillation frequency interval. Only the two middle clusters need to be extracted to obtain the two boundary values ​​of the normal oscillation frequency interval, thereby determining the oscillation frequency threshold interval. Preferably, the algorithm for determining the two middle clusters is detailed as follows: multiple oscillation frequency means of multiple clusters of oscillation frequency recording data are calculated, and the multiple clusters of oscillation frequency recording data are arranged in descending order according to the multiple oscillation frequency means to obtain the oscillation frequency recording data cluster sorting result.

[0080] When the recorded data of multiple clusters of oscillation frequencies are even, the middle two clusters can be directly extracted;

[0081] When the oscillation frequency record data of multiple clusters is an odd number, it is necessary to calculate the distribution distance between the middle cluster and its two adjacent clusters, and select the adjacent cluster with the larger distribution distance as the other middle cluster. The distribution distance refers to the deviation of the oscillation frequency means of the two clusters.

[0082] The maximum oscillation frequency of the cluster with the smaller oscillation frequency mean of the two middle clusters is extracted and set as the first fitting oscillation frequency. The minimum oscillation frequency of the cluster with the larger oscillation frequency mean of the two middle clusters is extracted and set as the second fitting oscillation frequency. At this time, the first fitting oscillation frequency ~ the second fitting oscillation frequency is the oscillation frequency interval considered normal for fitting. The frequency less than the first fitting oscillation frequency is set as the first oscillation frequency characteristic interval (that is, positive infinity and the first fitting oscillation frequency constitute the first oscillation frequency characteristic interval); the frequency greater than the second fitting oscillation frequency is set as the second oscillation frequency characteristic interval (that is, positive infinity and the second fitting oscillation frequency constitute the second oscillation frequency characteristic interval).

[0083] Furthermore, the rated oscillation frequency set by the user has a first rated oscillation frequency interval and a second rated oscillation frequency interval (wherein the lower limit of the first rated oscillation frequency interval is positive infinity, and the upper limit of the second rated oscillation frequency interval is positive infinity). The maximum value of the first rated oscillation frequency interval represents the minimum value of the normal rated oscillation frequency set by the user, and the minimum value of the second rated oscillation frequency interval represents the maximum value of the normal rated oscillation frequency set by the user. Therefore, the intersection of the first rated oscillation frequency interval and the first oscillation frequency characteristic interval is taken to obtain a more stringent first oscillation frequency threshold interval; and the intersection of the second rated oscillation frequency interval and the second oscillation frequency characteristic interval is taken to generate a more stringent second oscillation frequency threshold interval. By selecting a threshold interval with a higher degree of evaluation strictness, it is avoided that the abnormal low-frequency oscillation state cannot be identified. The detection rate of low-frequency abnormal oscillation mode is improved.

[0084] Step 4: When the first oscillation amplitude tag is greater than or equal to the oscillation amplitude threshold, or / and the first oscillation duration tag is greater than or equal to the oscillation duration threshold, or / and the first vibration frequency tag belongs to the first oscillation frequency threshold interval, or / and the first vibration frequency tag belongs to the second oscillation frequency threshold interval, the first low-frequency oscillation mode belongs to the low-frequency abnormal oscillation mode.

[0085] S40: Controlling the turbine according to the recommended unit guide vane opening and the recommended unit operating frequency, and controlling the AC excitation device according to the recommended unit speed set point.

[0086] In the embodiments of the present application, a turbine is an element in hydroelectric power generation that rotates in contact with the water flow to generate mechanical energy. An AC excitation device is used to control the turbine's speed. The turbine's speed, the turbine's guide vane opening, and the turbine's operating frequency collectively influence the generator's output power, which in turn affects the low-frequency oscillation mode. Therefore, by controlling the turbine using a recommended guide vane opening and a recommended operating frequency, and controlling the AC excitation device based on the recommended speed setpoint, adaptive regulation of the low-frequency oscillation mode can be achieved.

[0087] Furthermore, in step S40, the turbine mechanical energy prediction network is retrieved and, in combination with the pattern recognizer, an optimization algorithm is executed on the unit control parameters to obtain a recommended unit guide vane opening, a recommended unit operating frequency, and a recommended unit speed set point, including the following steps:

[0088] S41: The unit control parameters include the unit guide vane opening, the unit operating frequency, and the unit speed set point;

[0089] S42: Obtaining a guide vane opening range, an operating frequency range, and a speed set point range according to the turbine model, performing random assignment, and obtaining a first guide vane opening, a first operating frequency, and a first speed set point;

[0090] S43: collecting hydraulic scene parameters, wherein the hydraulic scene parameters include guide vane fully open flow and head parameters;

[0091] S44: Analyzing the first guide vane opening, the first operating frequency, the first speed setting point, the guide vane fully-open flow rate, and the water head parameter according to the turbine mechanical energy prediction network to obtain a turbine mechanical energy prediction value;

[0092] S45: Loading power generation efficiency coefficient;

[0093] S46: fitting the predicted value of the hydraulic turbine mechanical energy according to the power generation efficiency coefficient to obtain the predicted power generation power of the energy storage unit;

[0094] S47: Processing the power system operation data and the predicted power generation of the energy storage unit by the pattern recognizer to obtain a second low-frequency oscillation pattern;

[0095] S48: When the second low-frequency oscillation mode does not belong to the low-frequency abnormal oscillation mode, the first guide vane opening, the first operating frequency, and the first speed setting point are set as the recommended unit guide vane opening, the recommended unit operating frequency, and the recommended unit speed setting point.

[0096] Furthermore, the method further comprises the steps of:

[0097] S481: When the second low-frequency oscillation mode belongs to the low-frequency abnormal oscillation mode, update the first guide vane opening, the first operating frequency, and the first speed setting point, and execute an optimization loop, that is, return to step S44 until the recommended unit guide vane opening, the recommended unit operating frequency, and the recommended unit speed setting point are obtained.

[0098] In the embodiment of the present application, when performing the optimization of the unit control parameters, the basic algorithm principle is as follows:

[0099] First, the specific attributes of the unit control parameters are determined. The unit control parameters set in the embodiment of the present application include the unit guide vane opening, the unit operating frequency, and the unit speed set point. Secondly, since the hydraulic environment, that is, the hydraulic scene parameters, is also affected by the power generation power, the embodiment of the present application sets the hydraulic scene parameters to include the guide vane full-open flow and the head parameter. The guide vane full-open flow refers to the water flow when the turbine guide vanes are fully open, and the head parameter refers to the water level difference between the upstream water inlet section and the downstream tailwater outlet section, which characterizes the potential energy of the water flow. Furthermore, the guide vane opening interval, the operating frequency interval, and the speed set point interval refer to the preset rated intervals of the unit control parameters. Random assignment is performed from the guide vane opening interval, the operating frequency interval, and the speed set point interval to obtain the first guide vane opening, the first operating frequency, and the first speed set point.

[0100] Furthermore, the above parameters are the primary factors influencing the mechanical energy generated by the turbine. Through the turbine mechanical energy prediction network, the first guide vane opening, the first operating frequency, the first speed setpoint, the fully opened guide vane flow rate, and the head parameters are analyzed to obtain a predicted value for the turbine mechanical energy, representing the mechanical energy produced per unit time. The power generation efficiency coefficient represents the efficiency of converting the mechanical energy of the energy storage unit into electrical energy. The predicted power generation power of the energy storage unit is obtained by multiplying the power generation efficiency coefficient with the predicted turbine mechanical energy value.

[0101] Finally, a pattern recognizer processes the power system operating data and the second low-frequency oscillation pattern of the predicted power output of the energy storage unit. If the second low-frequency oscillation pattern does not fall within the low-frequency abnormal oscillation pattern, the first guide vane opening, the first operating frequency, and the first speed set point are set as the recommended unit guide vane opening, the recommended unit operating frequency, and the recommended unit speed set point. If the second low-frequency oscillation pattern falls within the low-frequency abnormal oscillation pattern, the first guide vane opening, the first operating frequency, and the first speed set point are updated, and an optimization loop is executed until the second low-frequency oscillation pattern is no longer within the low-frequency abnormal oscillation pattern, thereby obtaining the recommended unit guide vane opening, the recommended unit operating frequency, and the recommended unit speed set point.

[0102] The second low-frequency oscillation pattern is output data from the pattern recognizer, i.e., data representing the state of the low-frequency oscillation, including at least oscillation amplitude, oscillation duration, and oscillation frequency, specifically a second oscillation amplitude tag, a second oscillation duration tag, and a second oscillation frequency tag. A subsequent step can determine whether the low-frequency oscillation state is abnormal or normal by analyzing the second oscillation amplitude tag, the second oscillation duration tag, and the second oscillation frequency tag. The process of determining whether the second low-frequency oscillation pattern belongs to the abnormal low-frequency oscillation pattern is similar to the process of determining whether the first low-frequency oscillation pattern belongs to the abnormal low-frequency oscillation pattern, and will not be further described here.

[0103] The turbine mechanical energy prediction network mentioned above is a recurrent neural network. Users need to collect a guide vane opening record data set, an operating frequency record data set, a speed setting point record data set, a guide vane fully open flow record data set, a head parameter record data set, and a turbine mechanical energy identification data set; the turbine mechanical energy identification data set is used as the supervisory truth value, and the guide vane opening record data set, the operating frequency record data set, the speed setting point record data set, the guide vane fully open flow record data set, and the head parameter record data set are used as input data. When the turbine mechanical energy prediction deviation is less than or equal to the turbine mechanical energy prediction deviation threshold after a preset number of consecutive trainings, the turbine mechanical energy prediction network is considered to have converged. The turbine mechanical energy prediction network can be migrated and applied to different energy storage units. In specific applications, support vector machines, random forests and other models can also be used for training, and there is no restriction here.

[0104] Further, such as Figure 2 As shown, step S45, loading the power generation efficiency coefficient, includes:

[0105] S451: Obtaining historical power generation monitoring data of the energy storage unit in a preset retrospective time zone, wherein the historical power generation monitoring data of the energy storage unit includes recorded values ​​of output mechanical energy and recorded values ​​of output electrical energy;

[0106] S452: Calculating the ratio of the output electrical energy recorded value to the output mechanical energy recorded value to obtain a plurality of initial power generation efficiency coefficients;

[0107] S453: Performing a central trend analysis on the plurality of initial power generation efficiency coefficients to obtain the power generation efficiency coefficients.

[0108] In an embodiment of the present application, since the power generation efficiency of the unit may change dynamically after it is put into service, the power generation efficiency coefficient needs to be dynamically updated. The preset retrospective time zone refers to a historical time zone with a preset time length from the current moment, and the preset time length is preferably 3 to 6 months. The historical power generation monitoring data of the energy storage unit in the preset retrospective time zone is collected, and the historical power generation monitoring data of the energy storage unit includes the output mechanical energy record value and the output electric energy record value. The ratio of the output electric energy record value to the output mechanical energy record value is calculated to obtain several initial power generation efficiency coefficients. A central trend analysis is performed on the several initial power generation efficiency coefficients to obtain the power generation efficiency coefficient. The central trend analysis can adopt any algorithm of conventional central trend analysis. By dynamically updating the power generation efficiency coefficient, the power generation efficiency coefficient is guaranteed to be timely.

[0109] The method for adaptively controlling low-frequency oscillations in a power system of a variable-speed pumped-storage unit provided by the embodiments of the present invention has at least the following technical effects:

[0110] By providing real-time monitoring of the operation data of the power system and the power generation monitoring power of the energy storage unit, the collected data is processed by the pattern recognizer to output the first low-frequency oscillation pattern with oscillation amplitude, duration and frequency labels. When the identified oscillation pattern meets the low-frequency abnormal oscillation pattern, the system will automatically call the turbine mechanical energy prediction network, combine the output of the pattern recognizer and the mechanical energy prediction network, execute the optimization algorithm, and calculate the recommended unit guide vane opening, operating frequency and speed set point to optimize the operating status of the energy storage unit and effectively suppress low-frequency oscillation. By combining artificial intelligence technology, the oscillation status is detected in real time, and the power generation control parameters are adjusted in time to achieve low-frequency oscillation adaptive control. It can be migrated to different hydropower generation scenarios for application, achieving a technical effect with strong applicability.

[0111] Example 2:

[0112] See also Figure 3 , Figure 3 Schematic diagram of an embodiment of an electronic device provided by an embodiment of the present invention. Figure 3 As shown, an embodiment of the present invention provides an electronic device 500, including a memory 510, a processor 520, and a computer software program 511 stored in the memory 510 and executable on the processor 520. When the processor 520 executes the computer software program 511, the following steps are implemented:

[0113] Monitor power system operation data and energy storage unit power generation;

[0114] Processing the power system operation data and the energy storage unit power generation monitoring data through a pattern recognizer to output a first low-frequency oscillation pattern, wherein the first low-frequency oscillation pattern has a first oscillation amplitude tag, a first oscillation duration tag, and a first vibration frequency tag;

[0115] When the first oscillation amplitude tag, the first oscillation duration tag, and the first vibration frequency tag belong to a low-frequency abnormal oscillation mode, calling a turbine mechanical energy prediction network, combining the pattern recognizer, and executing an optimization algorithm on the unit control parameters to obtain a recommended unit guide vane opening, a recommended unit operating frequency, and a recommended unit speed set point;

[0116] The turbine is controlled according to the recommended unit guide vane opening and the recommended unit operating frequency, and the AC excitation device is controlled according to the recommended unit speed set point.

[0117] Example 3:

[0118] See also Figure 4 , Figure 4 Schematic diagram of an embodiment of a non-transitory computer-readable storage medium provided by an embodiment of the present invention. Figure 4As shown, this embodiment provides a non-transitory computer-readable storage medium 600, on which a computer software program 611 is stored. When the computer software program 611 is executed by a processor, the following steps are implemented:

[0119] Monitor power system operation data and energy storage unit power generation;

[0120] Processing the power system operation data and the energy storage unit power generation monitoring data through a pattern recognizer to output a first low-frequency oscillation pattern, wherein the first low-frequency oscillation pattern has a first oscillation amplitude tag, a first oscillation duration tag, and a first vibration frequency tag;

[0121] When the first oscillation amplitude tag, the first oscillation duration tag, and the first vibration frequency tag belong to a low-frequency abnormal oscillation mode, calling a turbine mechanical energy prediction network, combining the pattern recognizer, and executing an optimization algorithm on the unit control parameters to obtain a recommended unit guide vane opening, a recommended unit operating frequency, and a recommended unit speed set point;

[0122] The turbine is controlled according to the recommended unit guide vane opening and the recommended unit operating frequency, and the AC excitation device is controlled according to the recommended unit speed set point.

[0123] Example 4:

[0124] The present application also provides a computer program product, including a computer software program, which, when executed by a processor, implements the following steps:

[0125] Monitor power system operation data and energy storage unit power generation;

[0126] Processing the power system operation data and the energy storage unit power generation monitoring data through a pattern recognizer to output a first low-frequency oscillation pattern, wherein the first low-frequency oscillation pattern has a first oscillation amplitude tag, a first oscillation duration tag, and a first vibration frequency tag;

[0127] When the first oscillation amplitude tag, the first oscillation duration tag, and the first vibration frequency tag belong to a low-frequency abnormal oscillation mode, calling a turbine mechanical energy prediction network, combining the pattern recognizer, and executing an optimization algorithm on the unit control parameters to obtain a recommended unit guide vane opening, a recommended unit operating frequency, and a recommended unit speed set point;

[0128] The turbine is controlled according to the recommended unit guide vane opening and the recommended unit operating frequency, and the AC excitation device is controlled according to the recommended unit speed set point.

[0129] It should be noted that, in the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant description of other embodiments.

[0130] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0131] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0132] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0133] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0134] Although preferred embodiments of the present invention have been described, additional changes and modifications to these embodiments may occur to those skilled in the art once the basic inventive concepts become known.

[0135] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the present invention and its equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for adaptively controlling low-frequency oscillations in a variable-speed pumped storage power system, characterized in that: The method comprises: Monitor power system operation data and energy storage unit power generation; Processing the power system operation data and the energy storage unit power generation monitoring data through a pattern recognizer to output a first low-frequency oscillation pattern, wherein the first low-frequency oscillation pattern has a first oscillation amplitude tag, a first oscillation duration tag, and a first vibration frequency tag; When the first oscillation amplitude tag, the first oscillation duration tag, and the first vibration frequency tag belong to a low-frequency abnormal oscillation mode, obtaining a guide vane opening interval, an operating frequency interval, and a speed set point interval according to the turbine model, performing random assignment to obtain a first guide vane opening, a first operating frequency, and a first speed set point; collecting the guide vane fully-open flow rate and water head parameters, and analyzing the first guide vane opening, the first operating frequency, the first speed set point, the guide vane fully-open flow rate, and the water head parameters according to the turbine mechanical energy prediction network to obtain a turbine mechanical energy prediction value; Loading a power generation efficiency coefficient, fitting the predicted value of the turbine mechanical energy according to the power generation efficiency coefficient, and obtaining the predicted power generation power of the energy storage unit; processing the power system operating data and the predicted power generation of the energy storage unit by the pattern recognizer to obtain a second low-frequency oscillation pattern; and setting the first guide vane opening, the first operating frequency, and the first speed set point as a recommended unit guide vane opening, a recommended unit operating frequency, and a recommended unit speed set point when the second low-frequency oscillation pattern does not belong to the low-frequency abnormal oscillation pattern; Controlling the turbine according to the recommended unit guide vane opening and the recommended unit operating frequency, and controlling the AC excitation device according to the recommended unit speed set point; The method also includes: when the first oscillation amplitude label is greater than or equal to an oscillation amplitude threshold, or / and the first oscillation duration label is greater than or equal to an oscillation duration threshold, or / and the first vibration frequency label belongs to a first oscillation frequency threshold interval, or / and the first vibration frequency label belongs to a second oscillation frequency threshold interval, the first low-frequency oscillation mode belongs to the low-frequency abnormal oscillation mode.

2. The method according to claim 1, wherein Before processing the power system operation data and the energy storage unit power generation monitoring data by the pattern recognizer and outputting the first low-frequency oscillation pattern, the method further includes: Performing correlation sorting on power system operation parameter attributes to obtain selected power system operation data attributes; Based on the power system topology and energy storage unit model, low-frequency oscillation state mining is performed on the selected power system operation data attributes and energy storage unit power generation attributes to obtain power system operation record data, energy storage unit power generation record data, and low-frequency oscillation mode record data; The power system operation record data, the energy storage unit power generation record data and the low-frequency oscillation pattern record data are retrieved, and the pattern identifier uniquely associated with the power system is configured.

3. The method according to claim 2, wherein Retrieving the power system operation record data, the energy storage unit power generation record data, and the low-frequency oscillation pattern record data, and configuring the pattern identifier uniquely associated with the power system, including: Construct the loss function: in, Characterize each training The output loss after the second Characterize each The first training The output oscillation amplitude of the training, Characterize each The first training The output oscillation amplitude supervision value of the training, Characterize each The first training The output oscillation duration of the training, Characterize each The first training The output oscillation duration supervision value of the training, Characterize each The first training The output oscillation frequency of the training, Characterize each The first training The output oscillation frequency supervision value of the training, is a small constant, Respectively represent the weight parameters; The power system operation record data, the energy storage unit power generation record data and the low-frequency oscillation pattern record data are retrieved through the loss function to configure the pattern identifier uniquely associated with the power system.

4. The method according to claim 1, wherein The method further comprises: Based on the power system topology and energy storage unit model, historical power system oscillation accidents are collected, wherein the historical power system oscillation accidents include oscillation amplitude record data, oscillation duration record data, and oscillation frequency record data; Performing a centralized value analysis on the oscillation amplitude recorded data to obtain a centralized oscillation amplitude; When the concentrated oscillation amplitude is less than the rated oscillation amplitude, the concentrated oscillation amplitude is set as the oscillation amplitude threshold; otherwise, the rated oscillation amplitude is set as the oscillation amplitude threshold; Performing concentrated value analysis on the oscillation duration record data to obtain concentrated oscillation duration; When the concentrated oscillation duration is less than the rated oscillation duration, the concentrated oscillation duration is set as the oscillation duration threshold; otherwise, the rated oscillation duration is set as the oscillation duration threshold; Performing cluster analysis on the oscillation frequency recording data to obtain a first oscillation frequency characteristic interval and a second oscillation frequency characteristic interval; generating a first oscillation frequency threshold interval according to an intersection of a first rated oscillation frequency interval and the first oscillation frequency characteristic interval of the rated oscillation frequency; A second oscillation frequency threshold interval is generated according to an intersection of a second rated oscillation frequency interval of the rated oscillation frequency and the second oscillation frequency characteristic interval.

5. The method according to claim 4, wherein Performing cluster analysis on the oscillation frequency recording data to obtain a first oscillation frequency characteristic interval and a second oscillation frequency characteristic interval, including: Performing cluster analysis on the oscillation frequency recording data to obtain multiple clusters; sorting the multiple clusters according to corresponding oscillation frequency means; extracting two middle clusters from the multiple clusters, determining the maximum oscillation frequency of the cluster with the smaller corresponding oscillation frequency mean in the two middle clusters as the first fitting oscillation frequency, and determining the minimum oscillation frequency of the cluster with the larger corresponding oscillation frequency mean in the two middle clusters as the second fitting oscillation frequency; obtaining a first oscillation frequency characteristic interval based on positive infinity and the first fitting oscillation frequency first oscillation frequency characteristic interval, and obtaining a second oscillation frequency characteristic interval based on positive infinity and the second fitting oscillation frequency; The lower limit of the first rated oscillation frequency range is the positive infinity; the upper limit of the second rated oscillation frequency range is the positive infinity.

6. The method according to claim 1, wherein The method further comprises: When the second low-frequency oscillation mode belongs to the low-frequency abnormal oscillation mode, the first guide vane opening, the first operating frequency, and the first speed setting point are updated, and an optimization loop is executed until the recommended unit guide vane opening, the recommended unit operating frequency, and the recommended unit speed setting point are obtained.

7. The method according to claim 1, wherein Load the power generation efficiency coefficient, including: Obtaining historical power generation monitoring data of the energy storage unit in a preset retrospective time zone, wherein the historical power generation monitoring data of the energy storage unit includes recorded values ​​of output mechanical energy and recorded values ​​of output electrical energy; Calculating a ratio of the output electrical energy recorded value to the output mechanical energy recorded value to obtain a plurality of initial power generation efficiency coefficients; A central trend analysis is performed on the several initial power generation efficiency coefficients to obtain the power generation efficiency coefficient.

8. An electronic device, characterized in that: The electronic device comprises: Memory for storing computer software programs; A processor is used to read and execute the computer software program, thereby implementing the low-frequency oscillation adaptive control method of the variable-speed pumped storage unit power system according to any one of claims 1 to 7.

9. A non-transitory computer-readable storage medium, characterized in that The storage medium stores a computer software program, which, when executed by a processor, implements the method for adaptively controlling low-frequency oscillations of a variable-speed pumped-storage unit power system according to any one of claims 1 to 7.

10. A computer program product comprising a computer software program, characterized in that When the computer software program is executed by a processor, the steps of the variable speed pumped storage unit power system low frequency oscillation adaptive control method according to any one of claims 1 to 7 are implemented.

Images