Power system frequency prediction method and device, equipment and storage medium

Abstract

The invention discloses a power system frequency prediction method and device, equipment and a storage medium. The method comprises the following steps: acquiring key parameters influencing the frequency supporting capability in a power system containing new energy equipment; inputting the key parameters into a pre-constructed frequency prediction model to obtain a frequency prediction result of the power system; wherein the frequency prediction model is a model obtained based on regression decision tree training. According to the method, the key parameters influencing the frequency supporting capacity in the power system are obtained, the key parameters are input into the frequency prediction model obtained based on regression decision tree training, the frequency prediction result of the power system can be accurately obtained, and the obtained frequency prediction result has better interpretability.

Description

Technical Field

[0001] Embodiments of the present invention relate to the technical field of power systems, and in particular to a method, device, equipment and storage medium for predicting power system frequency. Background Art

[0002] Frequency is one of the three major indicators of power quality and an important indicator reflecting the safe and stable operation of the power system. Frequency stability is a global issue to ensure the safe and stable operation of the power system. When the power system is disturbed and the power supply is insufficient, the frequency of the power grid will drop. In severe cases, it may cause frequency collapse and trigger a major blackout. As the world's environmental crisis continues to intensify, it has become a consensus to accelerate the development of a new power system based on renewable energy power generation. The construction of the new power system has led to the application of a large number of power electronic products and the continuous increase in the proportion of new units, the gradual decrease in the proportion of conventional units, the change in the ratio of voltage and current sources in the power system, the reduction of system inertia, the weakening of frequency regulation capability, and the increasingly prominent problem of frequency regulation stability.

[0003] With the continuous development of artificial intelligence technology, there has been research on using the powerful feature mining capabilities of machine learning methods to analyze the frequency response characteristics of power systems after disturbances. For example, the method based on the Deep Belief Network (DBN) can predict the frequency extremes and their occurrence times, and the method using the Extreme Learning Machine (ELM) can evaluate the safety margin of the system frequency. However, the results of the machine learning method lack interpretability and cannot accurately predict the frequency of the power system. Summary of the invention

[0004] The present invention provides a method, device, equipment and storage medium for predicting the frequency of an electric power system, so as to solve the problem that the frequency of an electric power system cannot be accurately predicted in the prior art.

[0005] According to one aspect of the present invention, a method for predicting frequency of a power system is provided, the method comprising:

[0006] Obtain key parameters that affect frequency support capabilities in power systems containing new energy equipment;

[0007] The key parameters are input into a pre-built frequency prediction model to obtain a frequency prediction result of the power system; wherein the frequency prediction model is a model obtained by training based on a regression decision tree.

[0008] Furthermore, the training of the frequency prediction model includes:

[0009] Determine a training sample data set based on key parameters affecting frequency support capability in a power system including new energy equipment;

[0010] The model constructed based on the regression decision tree is trained based on the training sample data set to obtain a frequency prediction model.

[0011] Furthermore, the determining of the training sample data set based on the key parameters affecting the frequency support capability in the power system including the new energy equipment includes:

[0012] Acquire key parameters in a power system including new energy equipment, wherein the key parameters include at least a power disturbance value, a new energy ratio, a new energy ratio participating in frequency support, a sampling time constant, a negative frequency difference dead zone, a corresponding negative frequency difference change rate, and a maximum output power deviation;

[0013] determining a maximum frequency deviation of the power system based on key parameters of the power system;

[0014] Determine the frequency steady-state deviation of the power system based on the maximum value of the output power deviation of the power system, the proportion of new energy sources participating in frequency support, the frequency droop coefficient, the proportion of non-synchronous power sources, and the system equivalent regulation coefficient when the proportion of non-synchronous power sources is not considered;

[0015] The power disturbance value of the power system, the proportion of new energy, the proportion of new energy participating in frequency support, the sampling time constant, the negative frequency difference dead zone, the corresponding negative frequency difference change rate and the maximum output power deviation are used as input features in the training sample data set;

[0016] The maximum frequency deviation and the frequency steady-state deviation are used as output features in a training sample data set.

[0017] Furthermore, the frequency steady-state deviation is specifically determined by the following formula:

[0018]

[0019] Where Δf s is the steady-state frequency deviation of the system frequency, ΔP max K is the maximum output power deviation of the power system, sup is the proportion of renewable energy participating in frequency support in the power system, K re is the proportion of non-synchronous power in the power system, K d is the frequency droop coefficient, R sys is the system equivalent regulation coefficient when the proportion of non-synchronous power source is not considered, and D is the equivalent generator damping coefficient;

[0020] Among them, K sup It is expressed by the following formula:

[0021]

[0022] Among them, S f(NSG) is the power capacity of the non-synchronous generator involved in frequency control in the power system, S B(NSG) The power supply capacity of all non-synchronous machines in the power system;

[0023] Among them, R sys It is expressed by the following formula:

[0024]

[0025] Among them, R′ sys It is the system equivalent regulation coefficient when the proportion of non-synchronous power supply is considered.

[0026] Furthermore, the model constructed based on the regression decision tree is trained based on the training sample data set to obtain a frequency prediction model, including:

[0027] Determining a segmentation criterion through input features in the training sample data set;

[0028] Segmenting the samples in the training sample data set according to the segmentation standard to obtain a post-segmentation regression decision tree;

[0029] When the sample purity loss of the segmented regression decision tree reaches a loss threshold, using the segmented regression decision tree as a frequency prediction model;

[0030] Otherwise, continue to determine a new segmentation criterion, and segment the segmented regression decision tree according to the new segmentation criterion to obtain a new segmented regression decision tree.

[0031] Furthermore, the determining of the segmentation criteria by using the input features in the training sample data set includes:

[0032] Among all the input features in the training sample data set, the input feature that can reduce the variance of the regression decision tree the most and the value corresponding to the input feature are used as the segmentation criterion.

[0033] Furthermore, the method further comprises:

[0034] The importance I of the input feature to the frequency prediction result is determined by the following formula:

[0035]

[0036] Among them, ΔR(γ,t i ) represents the reduction in the prescription error of the regression decision tree at the i-th decision node, n node represents the number of all decision nodes in the initial regression decision tree, and γ is the segmentation criterion;

[0037] When it is determined according to the frequency prediction result that the frequency needs to be adjusted, a parameter to be adjusted is determined according to the importance.

[0038] According to another aspect of the present invention, there is provided a device for predicting frequency of a power system, the device comprising:

[0039] An acquisition module, used to acquire key parameters affecting frequency support capability in a power system including new energy equipment;

[0040] An input module is used to input the key parameters into a pre-built frequency prediction model to obtain a frequency prediction result of the power system; wherein the frequency prediction model is a model obtained by training based on a regression decision tree.

[0041] Furthermore, the training of the frequency prediction model includes:

[0042] A determination module, for determining a training sample data set based on key parameters affecting frequency support capability in a power system including new energy equipment;

[0043] The training module is used to train the model constructed based on the regression decision tree based on the training sample data set to obtain a frequency prediction model.

[0044] Furthermore, the determination module includes:

[0045] A first acquisition unit is used to acquire key parameters in a power system including new energy equipment, wherein the key parameters include at least a power disturbance value, a new energy ratio, a new energy ratio participating in frequency support, a sampling time constant, a negative frequency difference dead zone, a corresponding negative frequency difference change rate, and a maximum output power deviation value;

[0046] A first determining unit, configured to determine a maximum frequency deviation of the power system based on key parameters of the power system;

[0047] A second determination unit is used to determine the frequency steady-state deviation of the power system based on the maximum value of the output power deviation of the power system, the proportion of new energy sources participating in frequency support, the frequency droop coefficient, the proportion of non-synchronous power sources, and the system equivalent regulation coefficient when the proportion of non-synchronous power sources is not considered;

[0048] An input unit, used to use the power disturbance value of the power system, the proportion of new energy, the proportion of new energy participating in frequency support, the sampling time constant, the negative frequency difference dead zone, the change rate of the corresponding negative frequency difference and the maximum value of the output power deviation as input features in the training sample data set;

[0049] An output unit is used to use the maximum frequency deviation and the frequency steady-state deviation as output features in a training sample data set.

[0050] Furthermore, the frequency steady-state deviation is specifically determined by the following formula:

[0051]

[0052] Where Δf s is the steady-state frequency deviation of the system frequency, ΔP max K is the maximum output power deviation of the power system, sup is the proportion of renewable energy participating in frequency support in the power system, K re is the proportion of non-synchronous power in the power system, K d is the frequency droop coefficient, R sys is the system equivalent regulation coefficient when the proportion of non-synchronous power source is not considered, and D is the equivalent generator damping coefficient;

[0053] Among them, K sup It is expressed by the following formula:

[0054]

[0055] Among them, S f(NSG) is the power capacity of the non-synchronous generator involved in frequency control in the power system, S B(NSG) The power supply capacity of all non-synchronous machines in the power system;

[0056] Among them, R sys It is expressed by the following formula:

[0057]

[0058] Among them, R′ sys It is the system equivalent regulation coefficient when the proportion of non-synchronous power supply is considered.

[0059] Furthermore, the training module includes:

[0060] A third determining unit, configured to determine a segmentation criterion through input features in the training sample data set;

[0061] A segmentation unit, used to segment the samples in the training sample data set according to the segmentation standard to obtain a regression decision tree after segmentation;

[0062] A judgment unit is used to use the segmented regression decision tree as a frequency prediction model when the sample purity loss of the segmented regression decision tree reaches a loss threshold; otherwise, a new segmentation standard is continued to be determined, and the segmented regression decision tree is segmented according to the new segmentation standard to obtain a new segmented regression decision tree.

[0063] Furthermore, the third determining unit is specifically configured to:

[0064] Among all the input features in the training sample data set, the input feature that can reduce the variance of the regression decision tree the most and the value corresponding to the input feature are used as the segmentation criterion.

[0065] Furthermore, the device is also used for:

[0066] The importance I of the input feature to the frequency prediction result is determined by the following formula:

[0067]

[0068] Among them, ΔR(γ,t i ) represents the reduction in the prescription error of the regression decision tree at the i-th decision node, n node represents the number of all decision nodes in the initial regression decision tree, and γ is the segmentation criterion;

[0069] When it is determined according to the frequency prediction result that the frequency needs to be adjusted, a parameter to be adjusted is determined according to the importance.

[0070] According to another aspect of the present invention, there is provided an electronic device, the electronic device comprising: at least one processor; and

[0071] a memory communicatively connected to the at least one processor; wherein,

[0072] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the power system frequency prediction method described in any embodiment of the present invention.

[0073] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the power system frequency prediction method described in any embodiment of the present invention when executed.

[0074] An embodiment of the present invention provides a method, device, equipment and storage medium for predicting the frequency of an electric power system, the method comprising: obtaining key parameters that affect the frequency support capability of an electric power system including new energy equipment; inputting the key parameters into a pre-built frequency prediction model to obtain a frequency prediction result of the electric power system; wherein the frequency prediction model is a model obtained through training based on a regression decision tree. By obtaining key parameters that affect the frequency support capability of an electric power system and inputting the key parameters into a frequency prediction model obtained through training based on a regression decision tree, the method can accurately obtain a frequency prediction result of the electric power system, and the obtained frequency prediction result has better interpretability, thus solving the problem in the prior art that the frequency of an electric power system cannot be accurately predicted.

[0075] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. 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 creative work.

[0077] Figure 1 A schematic diagram of a flow chart of a method for predicting frequency of a power system provided in Embodiment 1 of the present invention;

[0078] Figure 2 A schematic diagram of a dynamic response process of a system frequency provided by an embodiment of the present invention;

[0079] Figure 3 A schematic diagram of a flow chart of a method for predicting frequency of a power system provided in Embodiment 2 of the present invention;

[0080] Figure 4 A schematic diagram of the structure of a frequency prediction model provided by an embodiment of the present invention;

[0081] Figure 5 A schematic diagram of a frequency variation curve provided by an embodiment of the present invention;

[0082] Figure 6 A network topology diagram of a power system provided by an embodiment of the present invention;

[0083] Figure 7 A schematic diagram of the structure of a power system frequency prediction device provided in Embodiment 3 of the present invention;

[0084] Figure 8 Schematic diagram of the structure of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0085] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only an embodiment of a part of the present invention, rather than all the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work should belong to the scope of protection of the present invention. It should be understood that the various steps recorded in the method implementation of the present invention can be performed in different orders and / or in parallel. In addition, the method implementation may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this respect.

[0086] The term "including" and its variations used herein are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0087] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, any variations of the terms "including" and "having" are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0088] It should be noted that the modifications of "one" and "plurality" mentioned in the present invention are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0089] The names of the messages or information exchanged between multiple devices in the embodiments of the present invention are only used for illustrative purposes, and are not used to limit the scope of these messages or information.

[0090] Embodiment 1

[0091] Figure 1A flow chart of a method for predicting the frequency of a power system provided in Embodiment 1 of the present invention is applicable to predicting the frequency of a power system in a power system including new energy equipment. The method can be executed by a power system frequency prediction device, wherein the device can be implemented by software and / or hardware and is generally integrated on an electronic device. In this embodiment, the electronic device includes but is not limited to: computers and other devices.

[0092] like Figure 1 As shown, a method for predicting power system frequency provided by Embodiment 1 of the present invention includes the following steps:

[0093] S110. Obtain key parameters that affect frequency support capability in a power system including new energy equipment.

[0094] Among them, the new energy equipment can be equipment that generates electricity through new energy (such as hydropower, solar energy, wind energy, biomass energy and geothermal energy, etc.). The frequency support capability of the power system refers to the ability of the power system to maintain the system frequency within a specified range when facing load changes or fluctuations in power generation capacity. This capability is crucial to maintaining the stability and reliability of the power system, because the frequency of the power system usually needs to be maintained at a relatively stable level (such as 50Hz or 60Hz) to ensure the normal operation of power equipment. Key parameters can be parameters in the power system that can affect the frequency support capability. In the context of new power systems, the proportion of traditional synchronous generator power supplies in the power grid is getting lower and lower, while the proportion of new energy power supplies is getting higher and higher. For renewable energy power supplies dominated by converters such as wind power and solar energy, they can be called "asynchronous machine power supplies". The new energy equipment in this embodiment may include asynchronous machine power supplies.

[0095] In this embodiment, when new energy equipment is involved in the power system, with the increase in the proportion of renewable energy, especially the access of wind and solar energy, frequency prediction becomes more important due to the instability and intermittency of their power generation. Therefore, this embodiment can obtain key parameters that affect the frequency support capability of the power system when the power system is operating normally or abnormally.

[0096] S120. Input the key parameters into a pre-built frequency prediction model to obtain a frequency prediction result of the power system; wherein the frequency prediction model is a model obtained by training based on a regression decision tree.

[0097] The frequency prediction model may be a model obtained by training based on a regression decision tree. The frequency prediction result may include a maximum frequency deviation and a frequency steady-state deviation.

[0098] In this embodiment, key parameters may be input into a pre-built frequency prediction model, and a frequency prediction result of the power system may be directly obtained through the frequency prediction model.

[0099] A method for frequency prediction of a power system provided in a first embodiment of the present invention includes: obtaining key parameters that affect frequency support capability in a power system including new energy equipment; inputting the key parameters into a pre-built frequency prediction model to obtain a frequency prediction result of the power system; wherein the frequency prediction model is a model obtained through training based on a regression decision tree. By obtaining key parameters that affect frequency support capability in a power system and inputting the key parameters into a frequency prediction model obtained through training based on a regression decision tree, the method can accurately obtain a frequency prediction result of the power system, and the obtained frequency prediction result has better interpretability, thus solving the problem in the prior art that the frequency of the power system cannot be accurately predicted.

[0100] Based on the above embodiment, a variant embodiment of the above embodiment is proposed. It should be noted that in order to make the description concise, only the differences from the above embodiment are described in the variant embodiment.

[0101] In one embodiment, the training of the frequency prediction model includes: determining a training sample data set based on key parameters affecting frequency support capability in a power system including new energy equipment; and training a model constructed based on a regression decision tree based on the training sample data set to obtain a frequency prediction model.

[0102] The training sample data set may include multiple samples.

[0103] In this embodiment, a training sample data set can be constructed based on key parameters that affect the frequency support capability in the power system. The specific construction method can directly generate multiple key parameter data through simulation software, use these data as training sample data sets, and use the training sample data sets to train the model constructed based on the regression decision tree to obtain a frequency prediction model.

[0104] In one embodiment, the method of determining a training sample data set based on key parameters affecting frequency support capability in an electric power system including new energy equipment comprises: obtaining key parameters in an electric power system including new energy equipment, wherein the key parameters include at least a power disturbance value, a new energy ratio, a new energy ratio participating in frequency support, a sampling time constant, a negative frequency difference dead zone, a corresponding negative frequency difference change rate, and a maximum output power deviation; determining a maximum frequency deviation of the electric power system based on the key parameters of the electric power system; determining a frequency steady-state deviation of the electric power system based on a maximum output power deviation of the electric power system, a new energy ratio participating in frequency support, a frequency droop coefficient, a non-synchronous power source ratio, and a system equivalent adjustment coefficient when the non-synchronous power source ratio is not considered; using the power disturbance value, the new energy ratio, the new energy ratio participating in frequency support, the sampling time constant, the negative frequency difference dead zone, the corresponding negative frequency difference change rate, and the maximum output power deviation of the electric power system as input features in the training sample data set; and using the maximum frequency deviation and the frequency steady-state deviation as output features in the training sample data set.

[0105] Among them, the power disturbance value can be the power change caused by load change, power generation capacity fluctuation, fault or other reasons in the power system. The proportion of new energy can be the proportion of new energy equipment to all equipment in the power system, and the proportion of new energy participating in frequency support can be the proportion of equipment participating in frequency support to all new energy equipment in the new energy equipment. The sampling time constant can refer to the response time of the sampling device used for frequency monitoring and control in the power system, and the sampling time constant represents the time required for the sampling device to output the corresponding control signal from receiving the frequency change signal. The negative frequency difference dead zone can refer to the interval in which the control system will not make adjustments or responses to the control of frequency deviation in the power system when the frequency deviation (negative frequency difference) is within a certain range. The rate of change of negative frequency difference can refer to the speed at which the system frequency decreases, indicating the amplitude of the frequency decrease in unit time. The maximum output power deviation can refer to the maximum deviation of the generator output power relative to the set value during the frequency control process, which can reflect the response capability and adjustment range of the generator in frequency control. The maximum frequency deviation can refer to the maximum value of the system frequency deviating from its normal value in the power system. After a disturbance (such as a sudden load increase or generator failure), the system frequency will change, and the maximum frequency deviation indicates the extreme value reached by the frequency during this change. The frequency steady-state deviation can refer to the deviation of the system frequency from its nominal value when the power system reaches a steady state after a certain period of time. After the system has experienced a disturbance and has been adjusted for a period of time, the frequency will eventually stabilize at a certain value. The frequency steady-state deviation is the difference between this stable value and the normal frequency.

[0106] In this embodiment, the key parameters may include a power disturbance value, a new energy proportion, a new energy proportion participating in frequency support, a sampling time constant, a negative frequency difference dead zone, a corresponding negative frequency difference change rate and a maximum output power deviation. These key parameters may be generated by simulation software. The maximum frequency deviation of the power system may be determined through the key parameters. The frequency steady-state deviation of the power system may be determined through the maximum output power deviation, the new energy proportion participating in frequency support, the frequency droop coefficient, the proportion of non-synchronous power sources and the system equivalent adjustment coefficient when the proportion of non-synchronous power sources is not considered. Thereafter, the power disturbance value, the new energy proportion, the new energy proportion participating in frequency support, the sampling time constant, the negative frequency difference dead zone, the corresponding negative frequency difference change rate and the maximum output power deviation may be used as input features in a training sample data set, and the maximum frequency deviation and the frequency steady-state deviation may be used as output features in a training sample data set to obtain a constructed training sample data set.

[0107] The maximum frequency deviation in this embodiment can be determined comprehensively by the power disturbance value, the proportion of new energy, the proportion of new energy participating in frequency support, the sampling time constant, the negative frequency difference dead zone, the corresponding negative frequency difference change rate and the maximum output power deviation. The specific acquisition method is obtained through existing simulation software, which will not be elaborated in this embodiment.

[0108] In one embodiment, the frequency steady-state deviation is specifically determined by the following formula:

[0109]

[0110] Where Δf s is the steady-state frequency deviation of the system frequency, ΔP max K is the maximum output power deviation of the power system, sup is the proportion of renewable energy participating in frequency support in the power system, K re is the proportion of non-synchronous power in the power system, K d is the frequency droop coefficient, R sys is the system equivalent regulation coefficient when the proportion of non-synchronous power source is not considered, and D is the equivalent generator damping coefficient;

[0111] Among them, K sup It is expressed by the following formula:

[0112]

[0113] Among them, S f(NSG) is the power capacity of the non-synchronous generator involved in frequency control in the power system, S B(NSG) The power supply capacity of all non-synchronous machines in the power system;

[0114] Among them, R sysIt is expressed by the following formula:

[0115]

[0116] Among them, R′ sys It is the system equivalent regulation coefficient when the proportion of non-synchronous power supply is considered.

[0117] In this embodiment, as the access capacity of the asynchronous machine power supply continues to increase, the proportion of the asynchronous machine power supply K re Will continue to increase, and its value range is 0≤K re ≤1. Since the retirement growth rate of conventional synchronous units is greater than their growth rate, the total inertia of the system is decreasing. In order to reflect the impact of the frequency response characteristics of the non-synchronous power supply access, this embodiment achieves the change in the proportion of non-synchronous power supply by replacing conventional synchronous generator sets with equal capacity.

[0118] When the proportion of the asynchronous machine power is K re When the system's equivalent inertia time constant H s ' ys It can be expressed as:

[0119] H′ sys =(1-K re )H sys ;

[0120] This embodiment can determine how to select key parameters through the following process analysis: the kinetic energy of the synchronous generator rotor speed directly responds to the frequency change of the system. When active power disturbance occurs, its rotational inertia has the function of damping frequency change. By introducing frequency droop control, the non-synchronous generator power can respond to the frequency change of the system. The active power output of the converter can be quickly adjusted through the frequency droop coefficient, thereby simulating the static power-frequency characteristics of primary frequency modulation. The principle can be expressed as follows:

[0121] ΔP d =-K d Δf;

[0122] Among them, ΔP d K is the power change of frequency droop control; d is the frequency droop coefficient. Since there is no prime mover-speed regulator link, the asynchronous machine power supply has a faster response speed based on inverter control and is an ideal source for primary frequency regulation services.

[0123] The synchronous machine speed regulation system participates in primary frequency regulation by adjusting the active power output, and its static power-frequency characteristics reflect the ability of primary frequency regulation. By introducing virtual inertia control, the non-synchronous machine power supply can respond to the rate of change of the system frequency, and the output of the converter active power can be quickly adjusted through the virtual inertia coefficient, thereby simulating the effect of system inertia on the frequency response characteristics, thereby equivalently improving the system inertia level. The principle can be expressed as:

[0124]

[0125] Among them, ΔP i K is the virtual inertia control power change; i is the virtual inertia coefficient.

[0126] From the perspective of power source, wind turbines can provide the power required for frequency regulation by releasing the rotational kinetic energy of the rotor or the spare capacity of load-reduced operation. Since photovoltaic systems do not have mechanical rotating parts, they mainly provide the spare power required for frequency regulation through load-reduced operation. The energy storage device supporting the new energy power generation system can also provide a certain amount of frequency regulation power support. When considering the participation of non-synchronous power sources in frequency regulation, the virtual inertia control and frequency droop control of the non-synchronous power sources bring new power-frequency action pathways, resulting in changes in the structure of the system frequency prediction model. The frequency regulation control of the non-synchronous power source can change the system inertia level and primary frequency regulation capability, and is reflected in the changes in the corresponding parameter sizes.

[0127] When considering that the non-synchronous power supply adopts virtual inertia control to provide virtual inertia, after the non-synchronous power supply replaces the synchronous unit with equal capacity, the system equivalent inertia time constant H based on the total rated capacity of the system s ' ys It can be further expressed as:

[0128]

[0129] Among them, H sys is the inertia time constant of the equivalent unit when the non-synchronous power supply is not considered, n is the number of synchronous units in the power system, S i is the rated capacity of the synchronous unit, H i is the inertia time constant of the synchronous unit, S B(NSG) is the total rated capacity of the asynchronous power supply, m is the number of asynchronous units, ΔH is the equivalent inertia time constant of the synchronous unit replaced by the asynchronous power supply, K reis the equivalent power gain coefficient of the non-synchronous power supply. It can be seen that the non-synchronous power supply can provide virtual inertia for the system through virtual inertia control, which is equivalent to improving the inertia level of the system. The size of the virtual inertia is determined by the capacity ratio of the non-synchronous power supply and the virtual inertia coefficient. By setting the virtual inertia coefficient, the non-synchronous power supply can provide virtual inertia of different sizes, and in theory can simulate a larger virtual inertia than the synchronous unit of the same capacity.

[0130] When considering the use of frequency droop control by non-synchronous power sources to participate in primary frequency regulation, after the non-synchronous power sources replace the synchronous units with equal capacity, the system equivalent regulation coefficient R based on the total rated capacity of the system is s ' ys It can be further expressed as:

[0131]

[0132] Among them, K' mh is the turbine mechanical power gain coefficient, K' ms is the turbine mechanical power gain coefficient, R s is the turbine regulation coefficient, R w is the turbine regulation coefficient. It can be seen that the asynchronous power supply can provide the system with a primary frequency regulation function through frequency droop control, which is equivalent to improving the primary frequency regulation capability of the system. The size of the asynchronous power supply's primary frequency regulation capability is determined by the capacity ratio of the asynchronous power supply and the frequency droop coefficient. By setting the frequency droop coefficient, the asynchronous power supply can provide different sizes of primary frequency regulation capabilities, and in theory, it can also simulate a larger primary frequency regulation capability than the synchronous unit of the same capacity.

[0133] Different from the inertia response mechanism of the synchronous unit, which responds almost instantaneously, the primary frequency regulation function of the synchronous unit needs to involve the action of the prime mover-speed regulator system, and a long response time is required from the action of the speed regulator to the completion of the adjustment of the output mechanical power. The frequency droop control is an active control link acting on the power electronic converter. It simulates the function of the primary frequency regulation of the synchronous generator by adjusting the output active power. Compared with the prime mover-speed regulation system, the response speed is faster. Since the two have certain differences in response time, the power characteristics of the frequency droop control of the non-synchronous power supply and the primary frequency regulation of the synchronous unit have certain differences in the frequency dynamic process, and cannot be completely equivalent, and the impact on the frequency response characteristics is not exactly the same. Under the same differential coefficient, the frequency droop control has a fast response speed and a more obvious improvement on the frequency dynamic characteristics. However, when the frequency enters the steady state period, since the frequency no longer changes dynamically, the two are exactly the same in the power frequency static characteristics. Therefore, the equivalent differential coefficient mentioned in this embodiment refers more to the equivalence in the power frequency static characteristics.

[0134] The two indicators of maximum frequency deviation and frequency steady-state deviation correspond to the frequency response characteristics of the frequency dynamic initial inertia response stage, the frequency dynamic mid-stage and the steady-state stage respectively. In order to ensure the safe and stable operation of the power grid, it is necessary to ensure that the system frequency meets the operation requirements in each time period of the dynamic process after being disturbed, that is, it is necessary to meet the maximum frequency deviation constraint and the frequency steady-state deviation constraint at the same time. For example, Figure 2 A schematic diagram of a dynamic response process of a system frequency provided by an embodiment of the present invention, such as Figure 2 As shown in the figure, affected by the inertia response and the different action time of the primary frequency modulation, after the active disturbance occurs, the dynamic response process of the system frequency is mainly divided into the following three stages:

[0135] The first stage: from the occurrence of active disturbance to the first frequency modulation response (about 0 to 2 seconds), the system first frequency modulation has not yet started, the synchronous machine releases the rotor rotational kinetic energy to maintain the system active power balance, which causes the system frequency to drop. This process is also called the inertia response stage. The rate-of-change-of-frequency (RoCoF) is the main characteristic indicator of this stage, in which the maximum frequency drop speed occurs at the moment of disturbance, and its size is only related to the disturbance power size and the system inertia level.

[0136] The second stage: from the time when the primary frequency modulation responds to the time when the system enters the steady state (about 2 seconds to tens of seconds), the inertia response of the synchronous generator and the primary frequency modulation work together to support the power balance of the system. The synchronous generator speed control system begins to respond to the changes in the system frequency, adjusts the output mechanical power of the prime mover, reduces the unbalanced power of the system, and the inertia support power decreases accordingly. In this stage, the system frequency reaches the lowest value and the maximum frequency deviation occurs, and the maximum frequency deviation is determined by the system inertia level and the primary frequency modulation capability. After that, the system frequency gradually transitions to a new steady state under the action of the synchronous machine speed control system.

[0137] The third stage: After tens of seconds of dynamic changes, the prime mover continues to increase the output mechanical power under the action of the synchronous machine speed control system until the generator mechanical power and the system load are completely balanced. The disappearance of unbalanced power makes the rotor speed no longer change, and the system frequency enters a new steady state. This process is also called the frequency steady-state stage. Since only the synchronous machine primary frequency regulation plays a role at this time, the frequency steady-state deviation is only determined by the speed regulator characteristics and the rotating reserve capacity, and has nothing to do with the system inertia level.

[0138] After a power disturbance accident occurs in the power system, the system frequency drops rapidly to the minimum value under the combined effect of inertia response and primary frequency regulation, and then gradually recovers to a new steady state. During this process, the system frequency reaches its lowest value. The maximum frequency deviation is the focus of attention in the frequency dynamic process. The maximum frequency deviation reflects the disturbance power that the current inertia level of the system and the primary frequency regulation capacity can withstand, and is an important indicator for evaluating the overall frequency stability of the system. At present, the power system protection device mainly sets and operates low-frequency load shedding based on the frequency deviation. When a power accident occurs, it is necessary to try to avoid touching the action threshold of low-frequency load shedding. The frequency steady-state deviation is affected by the system disturbance power size and the primary frequency regulation capacity, and represents the system's ability to recover from power disturbance faults.

[0139] Embodiment 2

[0140] Figure 3 This is a flow chart of a method for predicting the frequency of a power system provided by Embodiment 2 of the present invention. Embodiment 2 is optimized on the basis of the above embodiments. For details not yet fully described in this embodiment, please refer to Embodiment 1. Figure 3 As shown, a method for predicting the frequency of a power system provided by Embodiment 2 of the present invention, wherein a model constructed based on a regression decision tree is trained based on the training sample data set to obtain a frequency prediction model, comprising the following steps:

[0141] S210: Determine a segmentation criterion based on input features in the training sample data set.

[0142] The segmentation criteria may include input features and values ​​corresponding to the input features.

[0143] In this embodiment, the input features in the training sample data set may be analyzed to find the optimal segmentation criterion.

[0144] S220 , segmenting the samples in the training sample data set according to the segmentation standard to obtain a segmented regression decision tree.

[0145] Among them, the regression decision tree is a supervised machine learning technology that uses a tree structure to show the relationship between features and output labels. In a regression decision tree, each non-leaf node represents a decision point based on a certain feature, that is, the samples are divided according to the feature, and each branch corresponds to the result of the decision. The regression decision tree is recursively split into multiple subtrees according to the different values ​​of the sample set on a specific feature until it reaches a leaf node, at which point each leaf node contains the value of the regression prediction. Compared with other machine learning algorithms, the generation process of the regression decision tree has the advantages of intuitive visualization and easy understanding.

[0146] In this embodiment, all samples in the training sample data set may be segmented according to the segmentation standard to obtain a segmented regression decision tree.

[0147] S230: When the sample purity loss of the segmented regression decision tree reaches a loss threshold, the segmented regression decision tree is used as a frequency prediction model.

[0148] The sample purity loss may be a loss value of the sample purity. The loss threshold may be set according to actual conditions, and this embodiment does not limit this.

[0149] In this embodiment, when the sample purity loss of the segmented regression decision tree reaches the loss threshold, it can be considered that the segmented regression decision tree has met the requirements, and the segmented regression decision tree can be used as a frequency prediction model. The sample purity loss ΔR in this embodiment can be calculated as follows:

[0150]

[0151] Among them, R(t R ) is the variance of the right subtree after the split regression decision tree is split, R(t L ) is the variance of the left subtree after the split regression decision tree is split, N R is the number of samples in the right subtree of the regression decision tree after segmentation, N L is the number of samples in the left subtree of the regression decision tree after segmentation.

[0152] S240 , otherwise continue to determine a new segmentation criterion, and segment the segmented regression decision tree according to the new segmentation criterion to obtain a new segmented regression decision tree.

[0153] In this embodiment, when the sample purity loss of the segmented regression decision tree does not reach the loss threshold, a new segmentation criterion can be further determined, and the segmented regression decision tree can be segmented according to the new segmentation criterion to obtain a new segmented regression decision tree.

[0154] A method for predicting the frequency of a power system provided in the second embodiment of the present invention specifically trains a model constructed based on a regression decision tree based on the training sample data set to obtain a frequency prediction model, including: determining a segmentation criterion through input features in the training sample data set; segmenting samples in the training sample data set through the segmentation criterion to obtain a segmented regression decision tree; when the sample purity loss of the segmented regression decision tree reaches a loss threshold, using the segmented regression decision tree as a frequency prediction model; otherwise, continue to determine a new segmentation criterion, and segment the segmented regression decision tree through the new segmentation criterion to obtain a new segmented regression decision tree. This method constructs a frequency prediction model based on a regression decision tree, and can train a frequency prediction model with a high accuracy in predicting frequency.

[0155] In one embodiment, determining the segmentation criterion through the input features in the training sample data set includes: using the input feature that can reduce the variance of the regression decision tree the most among all the input features in the training sample data set and the value corresponding to the input feature as the segmentation criterion.

[0156] In this embodiment, among all the input features in the training sample data set, the input feature that can reduce the variance of the regression decision tree the most and the value corresponding to the input feature can be used as the segmentation criterion.

[0157] In this embodiment, the reduction in the variance of the regression decision tree can be calculated by the variance of the regression decision tree before and after segmentation. The calculation process of the variance of the regression decision tree is as follows:

[0158]

[0159] Among them, R(t) represents the variance of the regression decision tree at the tth node, N represents the number of samples corresponding to the tth node in the regression decision tree, and y i (t) is the value of the i-th sample, Represents the average value of N samples.

[0160] In this embodiment, in order to achieve the best separation effect of high-purity samples, the segmentation standard γ is selected when determining the segmentation standard. max The variance reduction ΔR(γ,t) should be maximized, that is, satisfying the following equation:

[0161] γ max =argmaxΔR(γ,t);

[0162] In one embodiment, the method further comprises:

[0163] The importance I of the input feature to the frequency prediction result is determined by the following formula:

[0164]

[0165] Among them, ΔR(γ,t i ) represents the reduction in the prescription error of the regression decision tree at the i-th decision node, n node represents the number of all decision nodes in the initial regression decision tree, and γ is the segmentation criterion;

[0166] When it is determined according to the frequency prediction result that the frequency needs to be adjusted, a parameter to be adjusted is determined according to the importance.

[0167] In this embodiment, the importance of the feature can be measured by accumulating the variance reduction of each feature appearing in the regression decision tree. When it is determined that the frequency needs to be adjusted based on the frequency prediction result, the parameters to be adjusted can be determined based on the importance of all key parameters.

[0168] To enhance the comparison between feature importances, the importance can also be normalized by scaling all values ​​to the range (0, 1) using the following formula:

[0169]

[0170] in, represents the importance of the normalized i-th input feature, I i Indicates the importance of the i-th input feature, I max Indicates the maximum importance of all input features.

[0171] The embodiment of the present invention provides several specific implementation methods based on the technical solutions of the above embodiments. As a specific implementation method of this embodiment, Figure 4 A schematic diagram of a frequency prediction model provided by an embodiment of the present invention is shown in FIG. Figure 4 As shown in Figure 2, the input features of the frequency prediction model include the power disturbance value ΔP, the proportion of new energy K new 、The proportion of new energy sources involved in frequency support K sup , sampling time constant T pe , negative frequency difference dead zone Δf db_neg , corresponding to the rate of change of negative frequency difference Δf min The maximum output power deviation ΔP max As input features, the frequency stability level can be characterized by the maximum frequency deviation and the frequency steady-state deviation as model outputs.

[0172] Figure 5 A schematic diagram of a frequency variation curve provided by an embodiment of the present invention, such as Figure 5As shown in the figure, in the test of this simulation example, the importance of the frequency control parameters of the renewable energy generator set is analyzed by using the characteristics of the regression decision tree model, and the regression decision tree generation process is visualized, using the low-frequency engineering test system as the application of the simulation system. In the operation scenario, the system fails and causes a power loss of 350MW. The maximum frequency deviation of the system is 1.182Hz, and the frequency steady-state deviation is 0.207Hz. The system has a low-frequency problem. Figure 6 A network topology diagram of a power system provided by an embodiment of the present invention, such as Figure 6 As shown in the figure, this is a DC transmission system that includes renewable energy. The example of this system is based on a 500kV grid structure and contains 102 nodes. Among them, the installed capacity of the wind farm and conventional units are 240MW and 540MW respectively. The system contains three DC lines, and the rated transmission power of the DC system is 800MW.

[0173] The training sample data set can be obtained through simulation software and used to train the regression decision tree model. The selection range of each parameter is shown in Table 1.

[0174] Table 1 Parameter selection range

[0175]

[0176]

[0177] The trained regression decision tree model can be used to predict the maximum frequency deviation and the frequency steady-state deviation. At the same time, in order to verify the effectiveness of the selected decision tree-based method in accurately predicting the power system frequency, five different regression models were trained using the same data set for comparison.

[0178] Table 2 Calculation results of different methods

[0179]

[0180] As shown in Table 2, other selected methods include K-nearest neighbor regression, random forest regression, AdaBoost regression, gradient boosting regression and bootstrap aggregating (Bagging) regression. The decision tree-based frequency prediction model selected in this paper has higher efficiency in predicting power grid frequency than other data-driven regression methods used for comparison.

[0181] The influence of each control parameter of the new energy unit on the maximum frequency deviation and the frequency steady-state deviation are shown in Table 3 and Table 4 respectively:

[0182] Table 3 The influence of maximum frequency deviation

[0183] parameter Impact ratio <![CDATA[Measuring the time constant T pe > 0.29110684 <![CDATA[Frequency deviation negative dead zone Δf db_neg > 0.45037653 <![CDATA[Rate of change Δf corresponding to negative frequency deviation min > 0.02009775 <![CDATA[Upper limit amplitude of power deviation output ΔP max > 0.23841887

[0184] Table 4 Impact ratio of frequency steady-state deviation

[0185] parameter Impact ratio <![CDATA[Measuring the time constant T pe > 0.24826213 <![CDATA[Frequency deviation negative dead zone Δf db_neg > 0.06983806 <![CDATA[Rate of change Δf corresponding to negative frequency deviation min > 0.41024606 <![CDATA[Upper limit amplitude of power deviation output ΔP max > 0.27165375

[0186] It can be seen from the simulation results that the factor with the greatest impact on frequency problems is power disturbance. The increase in the penetration rate of new energy has a greater impact on the maximum frequency deviation, but has a relatively small impact on the frequency steady-state deviation.

[0187] In terms of the influence of the control parameters of the new energy unit, the time constant T is measured pe and the upper limit peak amplitude output power deviation ΔP max It has a significant impact on both the maximum frequency deviation and the steady-state frequency deviation. Frequency deviation Δf db_neg The negative dead zone has a significant effect on the maximum frequency deviation, but has almost no effect on the steady-state frequency deviation, and corresponds to the negative frequency deviation Δf min The rate of change is just the opposite.

[0188] The frequency prediction method of this embodiment can have better interpretability and visualization effects through the frequency prediction model based on the regression decision tree. Through the frequency prediction model, the key factors affecting the frequency stability can be automatically screened out, thereby further improving the prediction accuracy.

[0189] Embodiment 3

[0190] Figure 7 This is a structural schematic diagram of a power system frequency prediction device provided in Example 3 of the present invention. The device can be used to predict the frequency of a power system in a power system including new energy equipment, wherein the device can be implemented by software and / or hardware and is generally integrated on an electronic device.

[0191] like Figure 7 As shown, the device comprises:

[0192] An acquisition module 310 is used to acquire key parameters affecting frequency support capability in a power system including new energy equipment;

[0193] The input module 320 is used to input the key parameters into a pre-built frequency prediction model to obtain a frequency prediction result of the power system; wherein the frequency prediction model is a model obtained based on regression decision tree training.

[0194] The present embodiment provides a power system frequency prediction device, including: an acquisition module, used to acquire key parameters that affect the frequency support capability in a power system including new energy equipment; an input module, used to input the key parameters into a pre-built frequency prediction model to obtain a frequency prediction result of the power system; wherein the frequency prediction model is a model obtained based on regression decision tree training. By acquiring the key parameters that affect the frequency support capability in the power system, and inputting the key parameters into the frequency prediction model obtained based on regression decision tree training, the frequency prediction result of the power system can be accurately obtained, and the obtained frequency prediction result has better interpretability, which solves the problem that the frequency of the power system cannot be accurately predicted in the prior art.

[0195] Furthermore, the training of the frequency prediction model includes: a determination module, which is used to determine a training sample data set based on key parameters affecting the frequency support capability in a power system including new energy equipment; and a training module, which is used to train a model constructed based on a regression decision tree based on the training sample data set to obtain a frequency prediction model.

[0196] Furthermore, the determination module includes: a first acquisition unit, used to acquire key parameters in the power system including new energy equipment, the key parameters at least including power disturbance value, new energy proportion, new energy proportion participating in frequency support, sampling time constant, negative frequency difference dead zone, corresponding negative frequency difference change rate and output power deviation maximum amplitude; a first determination unit, used to determine the maximum frequency deviation of the power system based on the key parameters of the power system; a second determination unit, used to determine the frequency steady-state deviation of the power system based on the maximum output power deviation of the power system, the proportion of new energy participating in frequency support, frequency droop coefficient, non-synchronous power supply proportion and system equivalent adjustment coefficient when the non-synchronous power supply proportion is not considered; an input unit, used to use the power disturbance value of the power system, new energy proportion, new energy proportion participating in frequency support, sampling time constant, negative frequency difference dead zone, corresponding negative frequency difference change rate and output power deviation maximum amplitude as input features in the training sample data set; an output unit, used to use the maximum frequency deviation and the frequency steady-state deviation as output features in the training sample data set.

[0197] Furthermore, the frequency steady-state deviation is specifically determined by the following formula:

[0198]

[0199] Where Δf s is the steady-state frequency deviation of the system frequency, ΔP max K is the maximum output power deviation of the power system, sup is the proportion of renewable energy participating in frequency support in the power system, Kre is the proportion of non-synchronous power in the power system, K d is the frequency droop coefficient, R sys is the system equivalent regulation coefficient when the proportion of non-synchronous power source is not considered, and D is the equivalent generator damping coefficient;

[0200] Among them, K sup It is expressed by the following formula:

[0201]

[0202] Among them, S f(NSG) is the power capacity of the non-synchronous generator involved in frequency control in the power system, S B(NSG) The power supply capacity of all non-synchronous machines in the power system;

[0203] Among them, R sys It is expressed by the following formula:

[0204]

[0205] Among them, R′ sys It is the system equivalent regulation coefficient when the proportion of non-synchronous power supply is considered.

[0206] Furthermore, the training module includes: a third determination unit, used to determine the segmentation standard through the input features in the training sample data set; a segmentation unit, used to segment the samples in the training sample data set according to the segmentation standard to obtain a segmented regression decision tree; a judgment unit, used to use the segmented regression decision tree as a frequency prediction model when the sample purity loss of the segmented regression decision tree reaches a loss threshold; otherwise, continue to determine a new segmentation standard, and segment the segmented regression decision tree according to the new segmentation standard to obtain a new segmented regression decision tree.

[0207] Furthermore, the third determination unit is specifically used to: use the input feature that can reduce the variance of the regression decision tree the most among all the input features in the training sample data set and the value corresponding to the input feature as the segmentation criterion.

[0208] Furthermore, the device is also used for:

[0209] The importance I of the input feature to the frequency prediction result is determined by the following formula:

[0210]

[0211] Among them, ΔR(γ,t i ) represents the reduction in the prescription error of the regression decision tree at the i-th decision node, n noderepresents the number of all decision nodes in the initial regression decision tree, and γ is the segmentation criterion;

[0212] When it is determined according to the frequency prediction result that the frequency needs to be adjusted, a parameter to be adjusted is determined according to the importance.

[0213] The above-mentioned power system frequency prediction device can execute the power system frequency prediction method provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0214] Embodiment 4

[0215] Figure 8 A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0216] like Figure 8 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0217] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0218] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as a power system frequency prediction method.

[0219] In some embodiments, the power system frequency prediction method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the power system frequency prediction method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to execute the power system frequency prediction method in any other appropriate manner (e.g., by means of firmware).

[0220] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0221] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0222] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0223] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0224] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0225] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0226] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0227] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for predicting power system frequency, characterized in that: The method comprises: Obtain key parameters that affect frequency support capabilities in power systems containing new energy equipment; The key parameters are input into a pre-built frequency prediction model to obtain a frequency prediction result of the power system; wherein the frequency prediction model is a model obtained by training based on a regression decision tree.

2. The method according to claim 1, characterized in that The training of the frequency prediction model includes: Determine a training sample data set based on key parameters affecting frequency support capability in a power system including new energy equipment; The model constructed based on the regression decision tree is trained based on the training sample data set to obtain a frequency prediction model.

3. The method according to claim 2, characterized in that The determining of the training sample data set based on the key parameters affecting the frequency support capability in the power system including the new energy equipment comprises: Acquire key parameters in a power system including new energy equipment, wherein the key parameters include at least a power disturbance value, a new energy ratio, a new energy ratio participating in frequency support, a sampling time constant, a negative frequency difference dead zone, a corresponding negative frequency difference change rate, and a maximum output power deviation; determining a maximum frequency deviation of the power system based on key parameters of the power system; Determine the frequency steady-state deviation of the power system based on the maximum value of the output power deviation of the power system, the proportion of new energy sources participating in frequency support, the frequency droop coefficient, the proportion of non-synchronous power sources, and the system equivalent regulation coefficient when the proportion of non-synchronous power sources is not considered; The power disturbance value of the power system, the proportion of new energy, the proportion of new energy participating in frequency support, the sampling time constant, the negative frequency difference dead zone, the corresponding negative frequency difference change rate and the maximum output power deviation are used as input features in the training sample data set; The maximum frequency deviation and the frequency steady-state deviation are used as output features in a training sample data set.

4. The method according to claim 3, characterized in that The frequency steady-state deviation is specifically determined by the following formula: Where Δf s is the steady-state frequency deviation of the system frequency, ΔP max K is the maximum output power deviation of the power system, sup is the proportion of renewable energy participating in frequency support in the power system, K re is the proportion of non-synchronous power in the power system, K d is the frequency droop coefficient, R sys is the system equivalent regulation coefficient when the proportion of non-synchronous power source is not considered, and D is the equivalent generator damping coefficient; Among them, K sup It is expressed by the following formula: Among them, S f(NSG) is the power capacity of the non-synchronous generator involved in frequency control in the power system, S B(NSG) The power supply capacity of all non-synchronous machines in the power system; Among them, R sys It is expressed by the following formula: Among them, R′ sys It is the system equivalent regulation coefficient when the proportion of non-synchronous power supply is considered.

5. The method according to claim 2, characterized in that: The step of training a model constructed based on a regression decision tree based on the training sample data set to obtain a frequency prediction model includes: Determining a segmentation criterion through input features in the training sample data set; Segmenting the samples in the training sample data set according to the segmentation standard to obtain a post-segmentation regression decision tree; When the sample purity loss of the segmented regression decision tree reaches a loss threshold, using the segmented regression decision tree as a frequency prediction model; Otherwise, continue to determine a new segmentation criterion, and segment the segmented regression decision tree according to the new segmentation criterion to obtain a new segmented regression decision tree.

6. The method according to claim 5, characterized in that Determining the segmentation criteria by using the input features in the training sample data set includes: Among all the input features in the training sample data set, the input feature that can reduce the variance of the regression decision tree the most and the value corresponding to the input feature are used as the segmentation criterion.

7. The method according to claim 5, characterized in that The method further comprises: The importance I of the input feature to the frequency prediction result is determined by the following formula: Among them, ΔR(γ,t i ) represents the reduction in the prescription error of the regression decision tree at the i-th decision node, n node represents the number of all decision nodes in the initial regression decision tree, and γ is the segmentation criterion; When it is determined according to the frequency prediction result that the frequency needs to be adjusted, a parameter to be adjusted is determined according to the importance.

8. A power system frequency prediction device, characterized in that: The device comprises: An acquisition module, used to acquire key parameters affecting frequency support capability in a power system including new energy equipment; An input module is used to input the key parameters into a pre-built frequency prediction model to obtain a frequency prediction result of the power system; wherein the frequency prediction model is a model obtained by training based on a regression decision tree.

9. The device according to claim 8, characterized in that The training of the frequency prediction model includes: A determination module, for determining a training sample data set based on key parameters affecting frequency support capability in a power system including new energy equipment; The training module is used to train the model constructed based on the regression decision tree based on the training sample data set to obtain a frequency prediction model.

10. The device according to claim 9, characterized in that The determining module comprises: A first acquisition unit is used to acquire key parameters in a power system including new energy equipment, wherein the key parameters include at least a power disturbance value, a new energy ratio, a new energy ratio participating in frequency support, a sampling time constant, a negative frequency difference dead zone, a corresponding negative frequency difference change rate, and a maximum output power deviation value; A first determining unit, configured to determine a maximum frequency deviation of the power system based on key parameters of the power system; A second determination unit is used to determine the frequency steady-state deviation of the power system based on the maximum value of the output power deviation of the power system, the proportion of new energy sources participating in frequency support, the frequency droop coefficient, the proportion of non-synchronous power sources, and the system equivalent regulation coefficient when the proportion of non-synchronous power sources is not considered; An input unit, used to use the power disturbance value of the power system, the proportion of new energy, the proportion of new energy participating in frequency support, the sampling time constant, the negative frequency difference dead zone, the change rate of the corresponding negative frequency difference and the maximum value of the output power deviation as input features in the training sample data set; An output unit is used to use the maximum frequency deviation and the frequency steady-state deviation as output features in a training sample data set.

11. The device according to claim 10, characterized in that The frequency steady-state deviation is specifically determined by the following formula: Where Δf s is the steady-state frequency deviation of the system frequency, ΔP max K is the maximum output power deviation of the power system, sup is the proportion of renewable energy participating in frequency support in the power system, K re is the proportion of non-synchronous power in the power system, K d is the frequency droop coefficient, R sys is the system equivalent regulation coefficient when the proportion of non-synchronous power source is not considered, and D is the equivalent generator damping coefficient; Among them, K sup It is expressed by the following formula: Among them, S f(NSG) is the power capacity of the non-synchronous generator involved in frequency control in the power system, S B(NSG) The power supply capacity of all non-synchronous machines in the power system; Among them, R sys It is expressed by the following formula: Among them, R′ sys It is the system equivalent regulation coefficient when the proportion of non-synchronous power supply is considered.

12. The device according to claim 9, characterized in that The training module comprises: A third determining unit, configured to determine a segmentation criterion through input features in the training sample data set; A segmentation unit, used to segment the samples in the training sample data set according to the segmentation standard to obtain a regression decision tree after segmentation; A judgment unit is used to use the segmented regression decision tree as a frequency prediction model when the sample purity loss of the segmented regression decision tree reaches a loss threshold; otherwise, a new segmentation standard is continued to be determined, and the segmented regression decision tree is segmented according to the new segmentation standard to obtain a new segmented regression decision tree.

13. The device according to claim 12, characterized in that The third determining unit is specifically configured to: Among all the input features in the training sample data set, the input feature that can reduce the variance of the regression decision tree the most and the value corresponding to the input feature are used as the segmentation criterion.

14. The device according to claim 12, characterized in that The device is also used for: The importance I of the input feature to the frequency prediction result is determined by the following formula: Among them, ΔR(γ,t i ) represents the reduction in the prescription error of the regression decision tree at the i-th decision node, n node represents the number of all decision nodes in the initial regression decision tree, and γ is the segmentation criterion; When it is determined according to the frequency prediction result that the frequency needs to be adjusted, a parameter to be adjusted is determined according to the importance.

15. An electronic device, characterized in that: The device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the power system frequency prediction method according to any one of claims 1 to 7.

16. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the power system frequency prediction method according to any one of claims 1 to 7 when executed.

Images