New energy and energy storage combined power generation system black start space-time support capability evaluation method

Abstract

The application provides a new energy and energy storage combined power generation system black start space-time support capability evaluation method, first analyzes the key technical factors restricting the safety problem of new energy participating in the initial black start, defines the black start space-time support capability from the safety margin level and the continuous support time; secondly, a simulation model of the initial black start of the light storage is built and fast time domain simulation is carried out, and a neural network training sample set is constructed; finally, the long short-term memory neural network is used to establish the evaluation model of the black start space-time support capability, and based on the black start space-time support capability evaluation model and the constructed neural network training sample set, the black start space-time support capability of the defined new energy and energy storage combined power generation system is evaluated. The application can consider the randomness of new energy output, improve the use effect of starting power in the initial black start stage, and provide index support for the subsequent black start scheme and black start partition.

Description

Technical Field

[0001] This invention relates to the field of power system restoration and proposes a method for assessing the spatiotemporal support capability of a new energy and energy storage combined power generation system during black start-up. Background Technology

[0002] To promote energy structure transformation and ensure its sustainable development, the power industry urgently needs to accelerate the construction of a new power system based on new energy sources. With the large-scale integration of new energy sources into the power grid, the grid structure and dynamic behavior of the power system are becoming increasingly complex, increasing the risk of large-scale power outages caused by faults. In recent years, the concept of a "resilient grid" has been proposed. By building a resilient grid, it is possible to proactively anticipate and prepare for various disturbances, actively defend against them, quickly restore important power loads, and enhance the grid's ability to cope with disturbances and disasters. Although the power system is constantly improving, the risk of large-scale blackouts cannot be completely eliminated. Therefore, developing black-start schemes suitable for new power systems and fully exploring the black-start capabilities of new energy units is of significant practical importance.

[0003] Existing research indicates that energy storage can mitigate fluctuations in wind and solar power output and accelerate grid power recovery. With the continuous decline in the unit cost of energy storage equipment and the ongoing development of control technology, photovoltaic-storage power stations have the potential to serve as black-start power sources. Previous literature has proposed using energy storage-type wind farms as black-start power sources to assist in the startup of thermal power units, and has conducted simulation analyses of various transient processes, exploring the feasibility of the scheme in depth. Other research has designed a self-starting controller and self-starting control strategy based on feedback linear control for the electromechanical transient characteristics of the self-starting process of doubly-fed asynchronous wind turbines. Meanwhile, deep learning, due to its powerful fitting ability, also has unique advantages and promising application prospects in the field of power system recovery. For transmission lines charging during the black-start phase, existing research has utilized artificial neural networks (ANNs) to quickly predict the maximum peak voltage of the black-start closing overvoltage, specifically employing an error backpropagation algorithm for training.

[0004] Based on the above, significant research has been conducted on the self-starting characteristics of new energy sources, but these findings have not yet been incorporated into the black start emergency power guarantee capability assessment system. Furthermore, the existing black start scheme assessment framework does not account for the spatiotemporal fluctuations in the output of new energy units. Therefore, for the black start phase, which urgently requires start-up power support, periodic evaluation of the black start support potential of different new energy power plants or rolling prediction of their spatiotemporal black start support capabilities would greatly improve the optimized use of start-up power during the initial black start phase. Summary of the Invention

[0005] In view of this, this invention proposes a method for evaluating the spatiotemporal support capability of a combined renewable energy and energy storage power generation system during black start. This method addresses the problem of scarce starting power in the initial stage of black start, considers the randomness of renewable energy output, and takes into account the safety margin level and continuous support time of the black start power source during the black start process, thereby evaluating its support capability. Results show that this method can optimize the evaluation of different black start schemes in a regional power grid, improve the utilization effect of starting power in the initial black start stage, and provide indicator support for subsequent black start zoning.

[0006] This invention provides a method for assessing the spatiotemporal support capability of a new energy and energy storage combined power generation system during black start, characterized by the following steps:

[0007] S1: Identify the key technical factors that restrict the safety of black start-up in the initial stage of new energy participation;

[0008] S2: Based on the key technical factors that restrict the safety of black start in the initial stage of new energy participation as determined in step 1, define the spatiotemporal support capability of black start of new energy power generation system.

[0009] S3: Build a simulation model of new energy sources and their combined energy storage power generation systems participating in the initial black start;

[0010] S4: Based on step S3, build a simulation model and use dynamic link library simulation technology to construct a neural network training sample set;

[0011] S5: Establish a black-start spatiotemporal support capability assessment model based on LSTM neural network, and evaluate the black-start spatiotemporal support capability of the new energy and energy storage combined power generation system defined in step S2 based on the black-start spatiotemporal support capability assessment model and the neural network training sample set constructed in step S4.

[0012] Furthermore, in step S1, the key technical factors that restrict the safety of new energy participation in the initial black start are determined as follows: the drop in system frequency and voltage when the plant auxiliary machine starts is a key technical factor that restricts new energy as a black start power source to participate in the initial black start, and the frequency and voltage impact on the system when the maximum capacity auxiliary machine starts is controlled within a certain limit range.

[0013] Furthermore, in step S2, based on the analysis of key technical factors restricting the safety of black start in the initial stage of new energy participation in step S1, the spatiotemporal support capability of black start is defined from two aspects: safety margin level and continuous support time.

[0014] First, regarding the safety margin level, based on the mathematical expression of margin as "distance / limit", the initial black-start safety margin index is defined from both frequency and voltage perspectives. and :

[0015] (1)

[0016] (2)

[0017] in, A safety margin indicator characterizing the system's frequency level; A safety margin indicator characterizing the system voltage level; The maximum acceptable frequency offset is to maintain system stability; The maximum frequency offset of the actual frequency of the connected load system; To maintain system stability, the maximum acceptable voltage drop; This refers to the actual voltage drop of the system.

[0018] The safety margin level for new energy power generation systems participating in the initial black start is defined as follows:

[0019] (3)

[0020] In the formula, The safety margin level for new energy power generation systems during the initial black start-up phase; This is a weighting coefficient, representing the relative importance of voltage indicators compared to frequency indicators; a value of 1 indicates that both are equally important.

[0021] Secondly, for the continuous support time, the safety margin level of the system under known external conditions is obtained by using time-domain simulation. Since the light intensity changes continuously with time, the photovoltaic output also changes at different times. Each sample is designated to represent a time section on the time axis, and this is continuously extended to form the black start support capability in the time dimension.

[0022] Furthermore, in step S4, an influencing factor analysis is performed on the black start spatiotemporal support capability defined in step S2. The energy storage configuration, energy storage SOC value, start-up distance, light intensity, start-up load size, and photovoltaic unit installed capacity are used as input features of the neural network. Based on the initial black start simulation model built in step S3, dynamic link library technology is used to obtain sample data through time-domain simulation, calculate the safety margin level, and use it together with the influencing factors as sample data. After labeling and preprocessing the sample data, a neural network training sample set is formed, which is divided into training data and test data according to the proportion.

[0023] Compared with the prior art, the present invention has the following beneficial effects:

[0024] 1. Considering the large-scale grid connection of new energy sources under the new power structure, this invention takes into account the spatiotemporal fluctuations of the output of new energy units, establishes an evaluation model for new energy black start schemes, and improves the existing evaluation and optimization system for black start schemes.

[0025] 2. This invention fully considers various technical problems of initial black start and, through preliminary simulation verification, has determined that the fluctuation of system frequency and voltage during the start-up of auxiliary equipment is the key factor restricting the participation of new energy sources as black start power sources in the initial black start.

[0026] 3. Considering that the output of new energy units varies with time, and that LSTM neural networks have unique advantages in processing time-related data, this invention comprehensively considers the offset level of frequency and voltage and the over-limit time. It defines the black start spatiotemporal support capability from two aspects: safety margin level and continuous support time. The black start spatiotemporal support capability evaluation model constructed using LSTM neural networks makes the black start scheme evaluation results more applicable to new power systems. Attached Figure Description

[0027] Figure 1 This is a schematic diagram showing the system frequency and voltage safety level.

[0028] Figure 2 A schematic diagram illustrating the spacetime support capabilities during black-out startup.

[0029] Figure 3 A schematic diagram of the initial black start structure for photovoltaic and energy storage participation;

[0030] Figure 4 This is a schematic diagram of the V / f control strategy;

[0031] Figure 5 This is a schematic diagram of an LSTM cell structure;

[0032] Figure 6 For the black start restoration topology of the regional power grid;

[0033] Figure 7 A comparative chart evaluating different recovery plans;

[0034] Figure 8 This is a flowchart illustrating a method for assessing the spatiotemporal support capability of a new energy and energy storage combined power generation system during black start, according to an embodiment of the present invention. Detailed Implementation

[0035] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0036] like Figure 8As shown in the figure, this invention provides a method for assessing the spatiotemporal support capability of a new energy and energy storage combined power generation system during black start, comprising the following steps:

[0037] S1: Identify the key technical factors that restrict the safety of black start in the initial stage of new energy participation.

[0038] In traditional power system black start, the main technical issues involved in the initial recovery phase include: generator self-excitation, operational overvoltage during line charging, and system voltage and frequency verification when starting auxiliary equipment of thermal power units.

[0039] Firstly, during the initial black start phase, the process of searching for a recovery path from the starting power source to the tripped unit with critical time constraints often involves long-distance transmission lines. If the unit capacity is small, self-excitation issues are highly likely to occur. When self-excitation occurs, a mismatch will appear between the generator terminal voltage and the excitation current. In some cases, the generator voltage becomes difficult to control, causing the voltage at individual points in the system to exceed the allowable value, endangering the insulation safety of the lines and transformers. However, when a new energy source and its combined energy storage power generation system participates in the initial black start phase as the black start power source, there are no traditional generator units, so the self-excitation issue can be disregarded.

[0040] Secondly, during the initial stage of black start, it is necessary to charge the unloaded line. The charging path is usually a series path consisting of a generator, transformer, and line. Generally, the total line length is relatively long, the voltage level is high, and overvoltage situations are quite severe. Overvoltage is caused by the oscillating process of conversion and transition between the magnetic energy stored in the inductor and the electrostatic energy in the capacitor during operations such as closing the unloaded line. However, for 220kV lines, the permissible operating overvoltage level is generally 3p.u. For lines with lower voltage levels, due to the smaller reactive power of charging, it is not necessary to check for operating overvoltage.

[0041] Finally, in the initial stage of system recovery, the black-start power supply is relatively small, and the system is relatively vulnerable. Close monitoring and coordination of the output and the rate of load input on the source side are necessary to prevent system frequency and voltage fluctuations from exceeding the system's tolerance limits. The critical loads input during black starts are mostly large auxiliary machines from large power plants, and many are asynchronous motors. Asynchronous motors remain stalled until they reach their rated torque, and the large starting current can severely impact the already fragile plant power system recovery process, causing voltage and frequency drops on the plant power side, ultimately leading to black start failure. Therefore, a black-start power supply with sufficient capacity is needed to meet the active and reactive power demands during auxiliary machine startup.

[0042] In summary, when using new energy sources and their combined energy storage power generation systems as black start power sources, the following situations need to be considered:

[0043] 1) The drop in system frequency and voltage during the startup of auxiliary equipment is a key technical factor that restricts new energy sources from participating in the initial black start as black start power sources.

[0044] 2) Since the system grid structure is relatively simple in the early stage of black start and has a weak ability to withstand load impact, the auxiliary equipment of thermal power plants usually needs to be started in sequence. Therefore, it is only necessary to consider that the frequency voltage impact on the system caused by the start of the auxiliary equipment with the maximum capacity is within a certain limit.

[0045] S2: Based on the key technical factors that restrict the safety of black start in the early stage of new energy participation, define the spatiotemporal support capability of black start of new energy power generation system.

[0046] Based on the analysis of key technical factors restricting the safety of black start in the initial stage of new energy participation in step S1, the spatiotemporal support capability of black start is defined from two aspects: safety margin level and continuous support time. Specifically, it includes the following steps:

[0047] S201: Regarding the safety margin level, based on the mathematical expression of margin as "distance / limit", the initial black-start safety margin index is defined from both frequency and voltage perspectives. and :

[0048] (1)

[0049] (2)

[0050] in, A safety margin indicator characterizing the system's frequency level; A safety margin indicator characterizing the system voltage level; The maximum acceptable frequency offset is to maintain system stability; The maximum frequency offset of the actual frequency of the connected load system; To maintain system stability, the maximum acceptable voltage drop; This refers to the actual voltage drop of the system.

[0051] The safety margin level for new energy power generation systems participating in the initial black start is defined as follows:

[0052] (3)

[0053] In the formula, The safety margin level for new energy power generation systems during the initial black start-up phase; This is a weighting coefficient, representing the relative importance of voltage indicators compared to frequency indicators; a value of 1 indicates that both are equally important.

[0054] According to standard requirements for frequency quality, the system frequency is classified as safe, critically safe, and unsafe when the frequency deviation is below 0.2Hz, above 0.5Hz, and above 0.8Hz, respectively. This invention classifies safety levels by assigning a frequency voltage safety level a range of 0-100. For simplified analysis, [the following is used...] Figure 1 The piecewise linear function shown represents the frequency security level. Then, in the figure... , , The values ​​are 0.2, 0.5, and 0.8 respectively, in Hz; similarly, the voltage safety level can also be expressed as... Figure 1 ,at this time , , The values ​​are taken as 0.05, 0.1 and 0.15 respectively according to the standard requirements for voltage quality, with the unit being pu. That is, when the voltage deviation is below 0.05pu, above 0.1pu and above 0.15pu respectively, the voltage of the system is called safe, critically safe and unsafe.

[0055] S202: For the sustained support time, the safety margin level of the system under external conditions is determined by time-domain simulation. However, the light intensity changes continuously over time, and thus the photovoltaic output also changes at different times. Therefore, each sample is designated to represent a time segment on the time axis, and this process is continuously extended to form the black-start support capability in the time dimension. The black-start spatiotemporal support capability is as follows: Figure 2 As shown.

[0056] S3: Build a simulation model of new energy sources and their energy storage combined power generation systems participating in the initial black start.

[0057] S301: Taking a photovoltaic-storage combined power generation system as an example, a simulation model of photovoltaic and energy storage participation in the initial black start is built. The structure of the simulation model is as follows: Figure 3 As shown. After receiving the black start command, the photovoltaic-storage power station starts the energy storage-assisted photovoltaic units. The photovoltaic-storage combined power generation system serves as the black start power source to charge the transmission lines and gradually starts the auxiliary equipment of the thermal power plant, working together with the thermal power units to restore the power supply to the entire local power grid.

[0058] S302: Formulate control strategies for the optical storage system to participate in the initial black start under islanded operation conditions.

[0059] In islanded operation, the voltage and frequency of the regional power grid need to be kept stable during black start. The AC side of the photovoltaic-storage power station adopts a V / f control strategy. Figure 4 This is a block diagram of the V / f control strategy for the photovoltaic-storage power station.

[0060] The V / f control strategy is a dual-loop control system based on voltage and current. The outer loop controls voltage, while the inner loop controls current. The two loops work together, with the inner loop nested within the outer loop, to ultimately stabilize the system's voltage and frequency parameters. The three-phase voltage i is measured from the microgrid system using a measuring device. cn and u n The direct-axis quadrature-axis command voltage u is obtained after Parker transformation. d and u q The two command voltages and the preset reference voltage u d_ref and u q_ref After subtraction, the direct-axis and quadrature-axis current components i are obtained through a PI controller. cd_ref and i cq_ref This process is called voltage outer loop control. Two current components i cd and i cq The current component i obtained from the system and subjected to the Parker transformation cd_ref and i cq_ref After subtraction, the direct-axis and quadrature-axis command voltage components u are obtained through another PI controller. d_ref and u q_ref This is called current inner loop control. Then the two voltage components u d_ref and u q_ref After the inverse transformation of the Parker transform, the modulated voltage is obtained. After SPWM modulation, the modulated pulse is output to the gate of the inverter to realize the turn-on and turn-off control of the inverter switching transistors.

[0061] Ultimately, the inverter controls the output voltage and frequency based on the collected data. When the load or power output of the microgrid changes, the system will experience a brief fluctuation process before finally achieving voltage and frequency stability.

[0062] S4: Based on the simulation model built in step S3, construct a neural network training sample set using dynamic link library simulation technology.

[0063] Specifically, an influencing factor analysis is performed on the black-start spatiotemporal support capability defined in step S2. Energy storage configuration, energy storage SOC value, start-up distance, irradiance, start-up load, and the installed capacity of the photovoltaic unit are used as input features of the neural network. Based on the initial black-start simulation model built in step S3, dynamic link library technology is used to obtain sample data through time-domain simulation, calculate the safety margin level, and use it together with the influencing factors as sample data. After labeling and preprocessing the sample data, a neural network training sample set is formed. Step S4 specifically includes the following steps:

[0064] S401: Analysis of Influencing Factors

[0065] First, the main obstacle for photovoltaic (PV) power plants as black-start power sources stems from the uncertainty of solar irradiance. Simultaneously, under maximum power point tracking (MPPT) operation, the PV modules' ability to participate in system frequency regulation varies with irradiance. Furthermore, limited by converter capacity, the reactive power output of PV modules also exhibits uncertainty with irradiance variations. Second, under varying irradiance conditions, energy storage devices can smooth PV output fluctuations and significantly improve the system's frequency response characteristics. Voltage stability can be maintained through the coordinated control of reactive power compensation devices such as static var compensators (SVCs) and static var generators (SVGs). In islanded grid operation, similar to the black-start process, research on energy storage configuration and frequency control has demonstrated that energy storage plays the most crucial role in autonomous operation.

[0066] Existing research indicates that during black start, there is a mapping relationship between the line length and the overvoltage of the black start no-load charging line. Furthermore, the line length and voltage level also affect the power delivery of the line; excessive starting distance leads to excessive voltage and power loss, reducing the success rate of starting the load.

[0067] Therefore, energy storage configuration, energy storage SOC value, start-up distance, and solar irradiance are used as input features for the neural network. In addition, the start-up load and the installed capacity of the photovoltaic units are also used as input information for model training. Details are shown in the table below:

[0068] Table 1 Summary of Input Features

[0069]

[0070] S402: Using dynamic link library technology to perform rapid time-domain simulation and obtain sample data.

[0071] In the architecture of the Windows operating system, Dynamic Link Libraries (DLLs) play a crucial role. They are the primary means of sharing resources among multiple processes and an important tool for the operating system to provide system services. DLLs are language-independent, provide generic interfaces, and are only loaded into memory when software calls them. Furthermore, they can be released after use, thus reducing the program's memory footprint.

[0072] Since acquiring sample data using a single Simulink simulation is too slow, the simulation model is encapsulated as a DLL, and the simulation is performed by calling the model from the program, thus speeding up the acquisition of sample data.

[0073] S403: After determining the input and output features, constructing a well-trained supervised training set typically includes acquiring sample data, sample labeling, data preprocessing, and dividing the training and test sets.

[0074] (1) Obtaining sample data. Based on the Matlab / Simulink simulation platform, a model was established in which the photovoltaic storage is used as a black start power source to participate in the initial black start. By changing the influencing factors such as start distance, black start load size, and light intensity, sample data was generated.

[0075] (2) Sample labeling. A black-start safety margin index was established based on both system frequency and voltage to determine whether the system is safe during a black start, and then all samples were labeled. The black-start safety margin index is shown in formula (3).

[0076] (3) Data preprocessing. To ensure that all features are compared at a uniform order of magnitude and to accelerate the convergence speed of the neural network model, the 0-1 normalization method is used to normalize the sample features. The input vector of the model is shown in the following formula.

[0077] (4)

[0078] (5)

[0079] In the formula: For the i-th sample feature, and These are the minimum and maximum values ​​of the feature, respectively.

[0080] (4) Divide the data into training data and test data. Randomly select 90% of the samples as training samples and 10% of the samples as test samples.

[0081] S5: Establish a black-start spatiotemporal support capability assessment model based on LSTM neural network, and assess the black-start spatiotemporal support capability of the new energy and energy storage combined power generation system defined in step S2 based on the black-start spatiotemporal support capability assessment model and the neural network training sample set constructed in step S4.

[0082] This invention introduces the LSTM neural network algorithm. The LSTM neural network was first proposed by Hochreiter et al. and improved by Graves. It is an improvement on RNN and solves the gradient vanishing problem that easily occurs in RNN.

[0083] LSTM cell structure as follows Figure 5 As shown. An LSTM unit contains one or more cell nuclei, which describe the current state of the LSTM unit. Figure 5There are three control gates: Input Gate, Output Gate, and Forget Gate. The outputs of the three gates are connected to a multiplication unit, thereby controlling the network's input, output, and the state of the Cell unit, respectively. After processing the first input O1, as long as the Input Gate is kept in the closed state (equivalent to the multiplication coefficient being 0) and the Forget Gate is kept in the open state (equivalent to the multiplication coefficient being 1), the network's output will continue to be affected by O1.

[0084] Then, a black-start spatiotemporal support capability assessment model is constructed based on LSTM. Based on the black-start spatiotemporal support capability assessment model and the neural network training sample set constructed in step S4, the black-start spatiotemporal support capability of the new energy and energy storage combined power generation system defined in step S2 is assessed.

[0085] (1) Model training: According to the weight parameters between each layer of the network, the weight coefficient of the t-th layer of the network is determined by the data of the previous t-1 layer. The previous input and output are used as the input and output of the LSTM network model for training, and the model parameters are adjusted. The cost function is the squared reconstruction error.

[0086] (2) Model Evaluation: To test the evaluation effect of the established model, two conventional evaluation indicators, Mean Absolute Percentage Error (MAPE) and Root Mean Squared Error (RMSE), were used to analyze the error of the prediction results. The calculation methods of the two evaluation indicators are shown in the following formulas.

[0087] (6)

[0088] (7)

[0089] In the formula: and These represent the MAPE and RMSE error values, respectively; N is the total length of the test data. and These are the predicted and actual values ​​of the black-start safety margin level for the i-th data sample, respectively.

[0090] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

[0091] The following explains the effectiveness of the black-start spatiotemporal support capability assessment method for new energy and energy storage combined power generation systems proposed in this invention. Figure 6 Taking a certain regional power grid as an example, we will conduct an analysis.

[0092] Due to the developed photovoltaic energy storage industry in a certain region, it is proposed to use its photovoltaic-energy storage system as the black start power source in this project. The photovoltaic power station has a total installed capacity of 30MW, the energy storage power station has a rated power of 12MW and a rated capacity of 24MW, the upper and lower limits of the energy storage SOC are 90% and 10% respectively, the energy storage charge and discharge efficiency is 90%, and the maximum power discharge and charging available time are 42.3min and 77.7min respectively.

[0093] The blackout start time was selected from 8:30 to 10:30 am, and the light intensity sampling time was 5 minutes.

[0094] There are three power plants available for startup—G1, G2, and G3. Specific line data and parameters for the black-start power plants are shown in Tables 2 and 3.

[0095] Black boot recovery scheme one: G1 is started by the optical storage black boot system via L1 line and L2 line.

[0096] Black start recovery scheme 2: The photovoltaic-storage black start system starts G2 via line L1, transformer T1, and line L3.

[0097] Black start recovery scheme three: The photovoltaic-storage black start system starts G3 via L1 line, transformer T1, L4 line, and transformer T2.

[0098] Table 2 Parameters of Photovoltaic-Storage Black Start Power Plants in Regional Power Grid

[0099]

[0100] Table 3 Parameters of Regional Power Grid Photovoltaic-Storage Black Start Line

[0101]

[0102] Figure 7This evaluation assesses the spatiotemporal support capabilities of three different black-start schemes for the regional power grid. Due to sufficient energy storage capacity, all schemes can smooth out fluctuations in photovoltaic power plant output within a 2-hour recovery period, ensuring stable power output. As sunlight intensity gradually increases over time, the safety margins of all three recovery schemes steadily improve. Compared to recovery scheme three, recovery schemes one and two have higher safety margins due to their shorter start-up distance, lower line voltage levels, and lower auxiliary power. Once sunlight intensity reaches a certain level, both schemes can safely proceed with a black start. While recovery scheme three can also be implemented, its safety margin is lower, posing a certain risk. Compared to recovery scheme two, recovery scheme one has an even shorter start-up distance and lower auxiliary power, thus maintaining a higher safety margin even in the initial stages of recovery when sunlight intensity is low.

[0103] In summary, this method can evaluate and optimize different black start schemes for new energy participation in regional power grids while taking into account the randomness of new energy output, thereby improving the utilization effect of initial black start on starting power and providing indicator support for subsequent black start zoning.

[0104] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for assessing the spatiotemporal support capability of a new energy and energy storage combined power generation system during black start, characterized in that, Includes the following steps: S1: Identify the key technical factors that restrict the safety of black-start in the initial stage of new energy participation; S2: Based on the key technical factors that restrict the safety of black start in the initial stage of new energy participation as determined in step 1, define the spatiotemporal support capability of black start of new energy power generation system. S3: Build a simulation model of new energy sources and their combined energy storage power generation systems participating in the initial black start; S4: Based on step S3, build a simulation model and use dynamic link library simulation technology to construct a neural network training sample set; S5: Establish a black-start spatiotemporal support capability assessment model based on LSTM neural network, and assess the black-start spatiotemporal support capability of the new energy and energy storage combined power generation system defined in step S2 based on the black-start spatiotemporal support capability assessment model and the neural network training sample set constructed in step S4. In step S2, based on the analysis of key technical factors restricting the safety of black start in the initial stage of new energy participation in step S1, the spatiotemporal support capability of black start is defined from the aspects of safety margin level and continuous support time. First, regarding the safety margin level, based on the mathematical representation of margin as "distance / limit", the initial black-start safety margin index is defined from both frequency and voltage perspectives. and : （1）； （2）； in, A safety margin indicator characterizing the system's frequency level; A safety margin indicator characterizing the system voltage level; The maximum acceptable frequency offset is to maintain system stability; The maximum frequency offset of the actual frequency of the connected load system; To maintain system stability, the maximum acceptable voltage drop; This refers to the actual voltage drop of the system. The safety margin level for new energy power generation systems participating in the initial black start is defined as follows: （3）； In the formula, The safety margin level for new energy power generation systems during the initial black start-up phase; This is a weighting coefficient, representing the relative importance of voltage indicators compared to frequency indicators; a value of 1 indicates that both are equally important. Secondly, for the continuous support time, the safety margin level of the system under known external conditions is obtained by using time-domain simulation. Since the light intensity changes continuously with time, the photovoltaic output also changes at different times. Each sample is designated to represent a time section on the time axis, and this is continuously extended to form the black start support capability in the time dimension.

2. The method for assessing the spatiotemporal support capability of a new energy and energy storage combined power generation system for black start-up as described in claim 1, characterized in that, In step S1, the key technical factors that restrict the safety of new energy participation in the initial black start are determined as follows: the drop in system frequency and voltage when the plant auxiliary machine starts is the key technical factor that restricts new energy as a black start power source to participate in the initial black start. The frequency and voltage impact on the system when the maximum capacity auxiliary machine starts is controlled within a certain limit.

3. The method for assessing the spatiotemporal support capability of a new energy and energy storage combined power generation system for black start-up as described in claim 1, characterized in that... In step S4, the influencing factors of the black start spatiotemporal support capability defined in step S2 are analyzed. The energy storage configuration, energy storage SOC value, start-up distance, light intensity, start-up load size and photovoltaic unit installed capacity are used as input features of the neural network. Based on the initial black start simulation model built in step S3, dynamic link library technology is used to obtain sample data through time domain simulation, calculate the safety margin level, and use it together with the influencing factors as sample data. After labeling and preprocessing the sample data, a neural network training sample set is formed, which is divided into training data and test data according to the proportion.

Images