Optimization method for regulating frequency of virtual power plant

Abstract

The invention relates to a virtual power plant frequency regulation optimization method, and relates to the field of power system frequency regulation, and the method comprises the steps: collecting a global power grid frequency and a local power grid frequency in real time, a first precise timestamp being attached to the global power grid frequency, and a second precise timestamp being attached to the local power grid frequency; calculating a time deviation based on the first precise timestamp and the second precise timestamp; when the time deviation is greater than the critical deviation value, executing a self-checking mode; when the time deviation is smaller than a critical deviation value, obtaining a global power grid frequency deviation and a local power grid frequency deviation through standard frequency difference based on the global power grid frequency and the local power grid frequency; dynamic weight distribution is carried out based on the global power grid frequency deviation and the local power grid frequency deviation, and the current frequency deviation is calculated; and performing frequency modulation optimization on the global power grid frequency and the local power grid frequency based on the current frequency deviation. According to the invention, efficient, accurate and automatic frequency regulation capability is realized, and the stability and reliability of frequency modulation are obviously improved.

Description

Technical Field

[0001] The present invention relates to the field of power system frequency regulation, and in particular to an optimization method for frequency regulation of a virtual power plant. Background Art

[0002] At present, with the large-scale grid connection of renewable energy, virtual power plants (VPPs), as the core technology for aggregating distributed energy, have become a key means to improve the flexibility of the power grid and the ability to absorb renewable energy. The deep integration of large-scale resource integration and high-precision frequency regulation is the core requirement to support the stable operation of power grids with a high proportion of new energy.

[0003] To effectively address the slow frequency regulation response and inflexibility of traditional power plants, technical experts in related fields have adopted a centralized optimization model to dispatch the power output of generators. Through this innovative approach, they have significantly shortened frequency regulation response time and significantly improved the frequency regulation accuracy of virtual power plants.

[0004] Regarding the above-mentioned related technologies, although this technology has greatly shortened the frequency regulation time and improved the frequency regulation accuracy, in the face of large-scale distributed energy access, the centralized optimization model has a single frequency regulation mode and is overly dependent on cloud data, resulting in bottlenecks in robustness. In particular, the frequency regulation data of a single frequency regulation mode will have large deviations and the centralized architecture is difficult to handle local problems. Summary of the Invention

[0005] In order to improve the existing technology for handling the access of large-scale distributed energy, which is difficult to achieve real-time and accurate frequency regulation, the present invention provides an optimization method for regulating the frequency of a virtual power plant.

[0006] The present invention provides a method for optimizing frequency regulation of a virtual power plant, which adopts the following technical solutions: Step 1: Collecting the global grid frequency and the local grid frequency in real time, wherein the global grid frequency is attached with a first precise time stamp, and the local grid frequency is attached with a second precise time stamp; Step 2: Calculate a time offset based on the first precise timestamp and the second precise timestamp; Step 3: When the time deviation is greater than a preset critical deviation value, executing a preset self-test mode; Step 4: when the time deviation is less than the critical deviation value, obtaining a global grid frequency deviation and a local grid frequency deviation by subtracting a preset standard frequency from the global grid frequency and the local grid frequency; Step 5: performing preset dynamic weight allocation based on the global grid frequency deviation and the local grid frequency deviation and calculating the current frequency deviation; Step 6: Optimize the global grid frequency and the local grid frequency based on the current frequency deviation.

[0007] By adopting the above technical solution, the time deviation is obtained based on the sending time of the first precise timestamp in the global power grid frequency and the second precise timestamp in the local power grid frequency. The two are confirmed to be data from the same time based on the time deviation being less than the critical deviation value, which effectively improves the reliability of the data, solves the problem of mismatch between the local power grid frequency deviation and the global power grid frequency deviation, and solves the traditional single frequency modulation method through frequency modulation optimization through dynamic weight allocation, effectively improving the accuracy of frequency modulation.

[0008] Optionally, the method for performing the dynamic weight allocation based on the global grid frequency and the local grid frequency includes: Step 50: Dynamically adjust the preset default weight coefficient according to the time deviation, and define the adjusted default weight coefficient as the modified weight coefficient; Step 51: Calculate the current frequency deviation based on the modified weight coefficient in combination with the global grid frequency deviation and the local grid frequency deviation, and output the calculated current frequency deviation.

[0009] By adopting the above technical solution, the current frequency deviation is output by allocating weight coefficients according to the time deviation, and more weight coefficients are allocated to data with smaller time deviations. This can flexibly cope with data processing under different time deviations, making frequency adjustment more adaptable.

[0010] Optionally, the method of executing the self-test mode when the time deviation is greater than the critical deviation value includes: Step 30: Decomposing the first timestamp of the global power grid frequency to obtain a global sending time and a global arrival time; Step 31: Determine a global delay time difference based on the global sending time and the global arrival time; Step 32: Executing a preset local mode when the global delay time difference exceeds a preset global communication threshold, wherein the local mode is a mode in which a weight coefficient corresponding to the local grid frequency deviation is larger than a weight coefficient corresponding to the global grid frequency deviation during calculation of the current frequency deviation; Step 33: When the global delay time difference is lower than the global communication threshold, exit the local mode and execute steps 5 to 6; Step 34: Decompose the second timestamp of the local power grid frequency to obtain a local sending time and a local arrival time; Step 35: Determine a local delay time difference based on the local sending time and the local arrival time; Step 36: Executing a preset global mode when the local delay time difference exceeds a preset local communication threshold, wherein the global mode is a mode in which the weight coefficient corresponding to the global grid frequency deviation is larger than the weight coefficient corresponding to the local grid frequency deviation during calculation of the current frequency deviation; Step 37: When the local delay time difference is lower than the local communication threshold, exit the global mode and execute steps 5 to 6.

[0011] By adopting the above technical solution, the corresponding global mode and local mode are executed according to the global delay time difference and the local delay time difference, and the proportion of the global grid frequency deviation and the local grid frequency deviation in calculating the current frequency deviation is flexibly adjusted, so as to better meet the actual frequency regulation needs and solve the problem that the global grid frequency and the local grid frequency delay are too large to effectively perform frequency regulation.

[0012] Optionally, the method of executing the local mode when the global delay time difference exceeds the global communication threshold includes: Step 320: Accumulate the global disconnection time when the global delay time difference exceeds the global communication threshold; Step 321: searching for a historical global power grid frequency deviation, and using the historical global power grid frequency deviation as the global power grid frequency deviation; Step 322: When the global disconnection time is less than a preset global connection time threshold, calculating and outputting the current frequency deviation based on a preset global tilt weight coefficient in combination with the global grid frequency deviation and the local grid frequency deviation, wherein the global tilt weight coefficient is a preset low weight coefficient assigned to the global grid frequency and a preset high weight coefficient assigned to the local grid frequency; Step 323: When the global disconnection time exceeds the global connection time threshold, frequency adjustment is performed based on the local power grid frequency.

[0013] By adopting the above technical solution, by accumulating the global disconnection time and comparing the global connection time threshold according to the global disconnection time, it is possible to avoid switching modes back and forth when the global delay time difference exceeds the global communication critical value in a short period of time, thereby reducing unnecessary operations. When the global disconnection time exceeds the global connection time threshold, frequency regulation is performed according to the global power grid frequency data to avoid erroneous global data from participating in frequency regulation.

[0014] Optionally, when the global disconnection time exceeds the global connection time threshold, the method for performing frequency regulation based on the local power grid frequency includes: Step 3230: Accumulating the local grid frequency deviations of the local nodes to obtain a local grid frequency deviation sum; Step 3231: Obtain a theoretical local power grid frequency deviation based on the local power grid frequency deviation and a preset deviation and average value algorithm; Step 3232: Subtracting the local grid frequency deviation from the theoretical local grid frequency deviation to obtain a local deviation; Step 3233: Adjusting the dynamic weight distribution based on the local deviation according to a preset deviation level weighting method; Step 3234: Calculate a weighted average of local deviations based on the local deviations and the deviation level weight coefficients; Step 3235: Output the weighted average value of the local deviation as the local power grid frequency deviation.

[0015] By adopting the above technical solution, a standard value is confirmed by calculating the theoretical local grid frequency deviation, and the allocation weight coefficient is dynamically adjusted according to the difference between the local grid frequency deviation and the standard value, so as to avoid the problem of ineffective frequency regulation when the difference between the local grid frequency deviations of individual local nodes is too large.

[0016] Optionally, the specific method of executing the global mode when the local delay time difference exceeds the local communication critical value includes: Step 360: Accumulate local disconnection time; Step 361: When the local disconnection time is less than a preset local connection time threshold, calculating and outputting the current frequency deviation based on a preset local tilt weight coefficient in combination with the global grid frequency deviation and the local grid frequency deviation, wherein the local tilt weight coefficient is a preset high weight coefficient assigned to the global grid frequency and a preset low weight coefficient assigned to the local grid frequency; Step 362: When the local disconnection time exceeds the local connection time threshold, perform frequency adjustment based on the global grid frequency.

[0017] By adopting the above technical solution, by accumulating the local disconnection time and comparing the local connection time threshold according to the local disconnection time, it is possible to avoid switching modes back and forth when the local delay time difference exceeds the local communication critical value in a short period of time, thereby reducing unnecessary operations. When the local disconnection time exceeds the local connection time threshold, frequency modulation is performed according to the global power grid frequency data to avoid erroneous local data from participating in frequency modulation.

[0018] Optionally, when the local disconnection time is less than the local connection time threshold, the method of calculating and outputting the current frequency deviation based on the local tilt weight coefficient in combination with the global grid frequency deviation and the local grid frequency deviation includes: Step 3610: defining the local delay time difference exceeding the local communication threshold as an abnormal delay time difference; Step 3611: Determine a local abnormal node based on the abnormal delay time difference; Step 3612: Obtain the historical local power grid frequency deviation corresponding to the local abnormal node; Step 3613: If the historical local power grid frequency deviation does not exist, determine other local nodes based on the local abnormal node; Step 3614: Calculate the local grid frequency deviations corresponding to the other local nodes according to a preset multi-node median algorithm to obtain a median local grid frequency deviation, and output the median local grid frequency deviation as the local grid frequency deviation; Step 3615: if the historical local grid frequency deviation exists, output the historical local grid frequency deviation as the local grid frequency deviation; Step 3616: Calculate the current frequency deviation based on the local tilt weight coefficient in combination with the global grid frequency deviation and the local grid frequency deviation and output it.

[0019] By adopting the above technical solution, when the local abnormal node data does not participate in frequency regulation, the frequency is regulated by obtaining the historical local frequency deviation of the local node, thereby solving the problem that the local abnormal node lacks historical data and cannot be frequency regulated. If there is no historical data, the median local power grid frequency deviation is calculated through the multi-node median algorithm for frequency regulation, further improving the accuracy and reliability of frequency regulation.

[0020] Optionally, the method of calculating the local grid frequency deviations corresponding to the other local nodes according to the multi-node median algorithm to obtain the median local grid frequency deviation, and outputting the median local grid frequency deviation as the local grid frequency deviation includes: Step 36140: Screening the other local nodes based on the local abnormal node to determine adjacent nodes and the number of nodes, wherein the adjacent nodes are the other local nodes whose corresponding local delay time differences do not exceed the local communication critical value; Step 36141: If the number of nodes is less than a preset reliable reference number, executing a preset historical data compensation method; Step 36142: If the number of nodes is greater than the reliable reference number, the adjacent nodes are sorted according to a preset ascending algorithm based on the local grid frequency deviations corresponding to the adjacent nodes to obtain an adjacent node order; Step 36143: Determine a median adjacent node based on the adjacent node sequence and a preset median algorithm, and define the local power grid frequency deviation corresponding to the median adjacent node as a median local power grid frequency deviation for output.

[0021] By adopting the above technical solution, the deviation of local power grid frequency data is reduced by setting a reliable reference number to screen other local nodes. At the same time, the median local power grid frequency deviation is determined by the incremental algorithm and the median algorithm, thereby improving the accuracy of the frequency regulation data and ensuring that relatively accurate frequency regulation can be achieved even in the absence of historical local power grid frequency.

[0022] Optionally, if the number of nodes is less than the reliable reference number, executing the method for historical data compensation includes: Step 361410: reducing the reliable reference quantity according to a preset equal weakening algorithm to obtain a weakened reference quantity; Step 361411: Compare the weakening reference number with the number of adjacent nodes; Step 361412: If the number of the adjacent nodes is less than the weakening reference number, summing the local grid frequency deviations of the adjacent nodes to obtain a sum of adjacent node frequency deviations; Step 361413: Calculate an average value based on the adjacent node frequency deviations and output it as the median local grid frequency deviation; Step 361414: If the number of adjacent nodes is greater than or equal to the weakening reference number, execute steps 36142 to 36143.

[0023] By adopting the above technical solution, remedial measures are provided when the number of nodes cannot reach the reliable reference number, solving the problem of insufficient data volume and ensuring that relatively accurate median local power grid frequency deviation can be effectively obtained under different numbers of nodes.

[0024] Optionally, a method for processing the local abnormal node is further included, the method comprising: Step 3617: If the local abnormal node exists, define the corresponding local abnormal node as a marked node; Step 3618: Restart the marked node and accumulate the number of restarts. Step 3619: When the restart count is less than a preset self-start count threshold, redetermine the local delay time difference of the marking node, and define the local delay time difference as a secondary delay time difference; Step 3620: If the secondary delay time difference does not exceed the local communication critical value, the marked node is regarded as a preset normal node and output; Step 3621: If the secondary delay time difference exceeds the local communication threshold, continue to perform the restart operation; Step 3622: If the restart times exceed the self-start times critical value, it is defined as a fault node and a preset warning notification SMS is sent to the mobile phone of the current on-duty personnel.

[0025] By adopting the above technical solution, local abnormal nodes are marked, restarted, and the number of restarts is accumulated. The node status is judged based on the number of restarts and the secondary delay time difference. If the number of restarts exceeds the critical value, it is regarded as a faulty node and an early warning notification is sent, which effectively solves the problem of local abnormal nodes and prevents local nodes from always being in the local abnormal node state and unable to participate in frequency modulation.

[0026] In summary, the present invention includes at least one of the following beneficial technical effects: By matching the local grid frequency with the global grid frequency at the time of the timestamp transmission, the local grid frequency and the global grid frequency generated at the same time are obtained. The local grid frequency deviation and the global grid frequency deviation are allocated according to the time deviation for dynamic frequency adjustment, realizing dynamic frequency adjustment based on data reliability. The reliability of the global and local grid frequencies is determined based on the global and local delay time differences, and corresponding local and global modes are set to respond accordingly. This ensures that frequency regulation can be achieved even in the absence of global or local grid frequency deviations. When the local abnormal node is solved by the multi-node centering algorithm and there is no historical local grid frequency, the approximate value of the local grid frequency deviation of the local abnormal node is obtained to replace the current local abnormal node for frequency regulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 This is a flow chart of a method for optimizing frequency regulation of a virtual power plant in an embodiment of the present application; Figure 2 is a flow chart of a method for dynamically allocating weights based on global grid frequency and local grid frequency in an embodiment of the present application; Figure 3 is a flowchart of a method for executing a preset local mode when a global delay time difference exceeds a preset global communication threshold in an embodiment of the present application; Figure 4 It is a flowchart of a method in an embodiment of the present application for calculating the local grid frequency deviations corresponding to other local nodes according to a preset multi-node median algorithm to obtain the median local grid frequency deviation, and outputting it as the local grid frequency deviation. DETAILED DESCRIPTION

[0028] The present invention is further described in detail below with reference to the accompanying drawings and embodiments.

[0029] The embodiment of the present invention discloses a method for evaluating frequency regulation capability. Figure 1 , an optimization method for adjusting frequency of a virtual power plant includes: Step 1: Collect global grid frequency and local grid frequency in real time.

[0030] The global grid frequency is attached with a first precise time stamp, and the local grid frequency is attached with a second precise time stamp.

[0031] Global grid frequency refers to the combined frequency value of the entire power grid system, collected at key nodes using pre-set data acquisition equipment. It reflects the overall operational status of the power grid. Local grid frequency refers to the frequency measured by data acquisition equipment in each local grid area, reflecting the current operating status of the power grid in each region. The first and second precise timestamps are timestamps inserted in the frequency data frame header to accurately record the time of generation of the global and local grid frequencies.

[0032] Step 2: Calculate the time offset based on the first precise timestamp and the second precise timestamp.

[0033] Time deviation is data used to determine whether the local grid frequency and the global grid frequency are generated at the same time. The specific calculation method is to subtract the transmission time of the local grid frequency from the global grid frequency to obtain the deviation value. By attaching a precise timestamp to the frequency data, the time deviation can be accurately obtained by subtracting the transmission time of the local grid frequency from the global grid frequency.

[0034] Step 3: When the time deviation is greater than a preset critical deviation value, a preset self-test mode is executed.

[0035] The critical deviation value is a pre-set deviation value. When it exceeds the critical deviation value, it indicates that the coordinated frequency regulation results of the local and global power grids will have a large error. The self-check mode is a mode for self-correcting the data and patterns related to the virtual power plant's frequency regulation and dynamic weight coefficients. The specific self-check mode is divided into global mode and local mode, which will be introduced in the subsequent process and will not be repeated here. If the time deviation is greater than the critical deviation value, it means that the local grid frequency and the global grid frequency may not be the same data at the same time, so the self-check mode must be executed. The specific implementation of the self-check mode is introduced in the subsequent steps and will not be repeated here.

[0036] Step 4: When the time deviation is less than the critical deviation value, a global grid frequency deviation and a local grid frequency deviation are obtained by subtracting the global grid frequency from the local grid frequency by a preset standard frequency.

[0037] The standard frequency is a standardized frequency value obtained through analysis of multiple experimental data. When this frequency is used as the standard frequency, the frequency regulation accuracy is the highest, and it represents the frequency that the power grid should maintain under ideal conditions.

[0038] Global grid frequency deviation and local grid frequency deviation refer to the difference between the global grid frequency and the local grid frequency and the standard frequency respectively. They can clearly reflect the degree to which the current global and local grid frequencies deviate from the ideal state, thereby providing accurate data basis for subsequent frequency adjustment.

[0039] When the time deviation is less than the critical deviation threshold, it means that the local grid frequency and the global grid frequency are data at the same time. At this time, the global grid frequency deviation and the local grid frequency deviation obtained by subtracting the standard frequency are relatively accurate and reliable, and can be used for subsequent precise frequency regulation.

[0040] Step 5: Perform preset dynamic weight allocation based on the global grid frequency deviation and the local grid frequency deviation and calculate the current frequency deviation.

[0041] Dynamic weight allocation is a method that flexibly adjusts the weight of the global and local grid frequency deviations in frequency regulation based on their actual conditions. For example, in local mode, dynamic weight allocation assigns a high weight to the local grid frequency deviation and a low weight to the global grid frequency deviation. The current frequency deviation refers to the frequency deviation based on which the grid frequency is regulated. When the time deviation is less than the critical deviation threshold, indicating that the local and global grid frequencies are at the same time, weights are assigned based on the global and local grid frequency deviations to calculate the current frequency deviation.

[0042] Step 6: Optimize the global grid frequency and the local grid frequency based on the current frequency deviation.

[0043] The current frequency deviation is analyzed, and a preset frequency modulation coefficient is selected based on the preset deviation level range to multiply the current frequency deviation. The adjusted power is then used to control the power increase or decrease required by the generators in the power grid. Frequency modulation is optimized based on the adjusted power to maintain the global and local grid frequencies within a stable range. The deviation level range is divided into layers based on the deviation value. For example, if the current frequency deviation is less than 0.1Hz in absolute value, a 5% frequency modulation coefficient is required. The frequency modulation coefficient is a multiplier with a default value of 50%. Adjusted power is delegated to local nodes for frequency modulation. Local nodes are specifically divided into slow resources (such as energy storage systems and gas turbines) and fast resources (such as coal-fired units).

[0044] Reference Figure 2, the method for dynamic weight allocation based on global grid frequency and local grid frequency includes: Step 50: Dynamically adjust the preset default weight coefficient according to the time deviation, and define the adjusted default weight coefficient as the modified weight coefficient.

[0045] The default weight coefficient is a pre-set weight value assigned to the local grid frequency deviation and the global grid frequency deviation, which is 0.5. Initially, this weight coefficient is used to calculate the global and local grid frequency deviations. The modified weight coefficient is a new weight coefficient obtained by dynamically adjusting the default weight coefficient based on time deviation, which better reflects the actual operation of the current grid.

[0046] As an important reference factor, time deviation can reflect the timeliness and reliability of data. When the time deviation changes, the default weight coefficient also needs to be adjusted accordingly. For example, when the time deviation is smaller, the weight of the local grid frequency deviation is increased and the weight of the global grid frequency deviation is reduced to ensure the accuracy of frequency regulation. Therefore, the weight needs to be corrected.

[0047] Step 51: Calculate the current frequency deviation based on the modified weight coefficient in combination with the global grid frequency deviation and the local grid frequency deviation, and output it.

[0048] The global grid frequency deviation and the local grid frequency deviation are multiplied by their corresponding correction weight coefficients, and the multiplied results are added and averaged to obtain the current frequency deviation.

[0049] The method of executing the self-test mode when the time deviation is greater than the critical deviation value includes: Step 30: Decompose the first timestamp of the global grid frequency to obtain a global sending time and a global arrival time.

[0050] The global transmission time is the time when the global grid frequency is transmitted. The global arrival time is the time when the global grid frequency is received. When the global grid frequency is received, the current time is automatically recorded in the timestamp and defined as the global arrival time.

[0051] The first timestamp data includes the global sending time and the global arrival time. The global sending time and the global arrival time are stored in the preset position of the first timestamp data packet. The format is similar to the IP datagram. The global sending time and the global arrival time can be obtained by receiving the data at this position.

[0052] Step 31: Determine a global delay time difference based on the global sending time and the global arrival time.

[0053] The global delay time difference refers to the transmission time of the global power grid frequency. The global delay time difference is determined by subtracting the global transmission time from the global arrival time.

[0054] Step 32: Execute a preset local mode when the global delay time difference exceeds a preset global communication threshold.

[0055] The local mode is a mode in which, during the calculation of the current frequency deviation, the weight coefficient corresponding to the local grid frequency deviation is larger than the weight coefficient corresponding to the global grid frequency deviation.

[0056] The global communication threshold is a pre-set standard value used to measure the normality of global grid frequency transmission delays. When the global delay time difference exceeds the global communication threshold, it means that a significant delay has occurred in the global grid frequency transmission process, reducing the reliability of the global grid frequency. Therefore, local mode is implemented, focusing more on calculating the current frequency deviation based on the local grid frequency deviation to ensure the relative accuracy of frequency regulation.

[0057] Step 33: When the global delay time difference is lower than the global communication threshold, exit the local mode and execute steps 5 to 6.

[0058] When the global delay time difference is lower than the global communication critical value, it means that the current global grid frequency deviation reliability meets the requirements and can participate in frequency regulation. The current frequency deviation is calculated based on the global grid frequency deviation and the local grid frequency deviation, and then frequency regulation is performed based on the current frequency deviation.

[0059] Step 34: Decompose the second timestamp of the local grid frequency to obtain the local sending time and the local arrival time.

[0060] The local sending time refers to the time when the local grid frequency is sent, and the local arrival time refers to the time when the local grid frequency is received. When the local grid frequency is received, the current time is automatically recorded in the timestamp and defined as the local arrival time. The second timestamp data contains the local sending time and local arrival time, which are obtained by disassembly.

[0061] Step 35: Determine the local delay time difference based on the local sending time and the local arrival time.

[0062] The local global delay time difference refers to the transmission time of the local power grid frequency. The local delay time difference is obtained by subtracting the local transmission time from the local arrival time.

[0063] Step 36: When the local delay time difference exceeds the preset local communication threshold, a preset global mode is executed. The global mode is a mode in which the weight coefficient corresponding to the global grid frequency deviation is larger than the weight coefficient corresponding to the local grid frequency deviation during the calculation of the current frequency deviation.

[0064] The local communication critical value refers to the maximum acceptable delay critical value obtained through analysis of multiple experiments. When the local delay time difference exceeds the local communication critical value, it indicates that the current local power grid frequency deviation reliability is very low, and participating in frequency regulation will affect accuracy, so the global mode must be executed.

[0065] Step 37: When the local delay time difference is lower than the local communication threshold, exit the global mode and execute steps 5 to 6.

[0066] When the local delay time difference is lower than the local communication critical value, it means that the current local power grid frequency deviation reliability meets the requirements and can participate in frequency modulation. The current frequency deviation is calculated based on the local power grid frequency deviation and the global power grid frequency deviation, and then frequency modulation is performed based on the current frequency deviation.

[0067] Reference Figure 3 The method for executing the local mode when the global delay time difference exceeds the global communication threshold includes: Step 320: Accumulate the global disconnection time when the global delay time difference exceeds the global communication threshold.

[0068] The global disconnection time refers to the duration during which the global grid frequency loses effective synchronization with the local grid due to transmission delay exceeding the global communication critical value.

[0069] When the global delay difference exceeds the global communication threshold, it indicates a significant delay in global grid frequency transmission, potentially leading to loss of synchronization with the local grid. At this point, global disconnection time begins to accumulate. By recording this time, we can better understand the persistence of global grid frequency anomalies and provide a basis for subsequent decision-making.

[0070] Step 321: Search for historical global power grid frequency deviations, and use the historical global power grid frequency deviations as the global power grid frequency deviation.

[0071] Historical global grid frequency deviation refers to the global grid frequency deviation recorded in preset historical records. Each received global grid frequency deviation and local grid frequency deviation are recorded in the historical records. When the preset historical record capacity is exceeded, the newly recorded data overwrites the oldest recorded time. When the global delay time difference exceeds the global communication threshold, it indicates that the current global grid frequency data is unreliable, so the historical global grid frequency deviation is retrieved to replace it for subsequent frequency adjustment operations.

[0072] Step 322: When the global disconnection time is less than the preset global connection time threshold, the current frequency deviation is calculated based on the preset global tilt weight coefficient and the global grid frequency deviation and the local grid frequency deviation, and is output.

[0073] The global tilt weight coefficient is a preset low weight coefficient allocated to the global grid frequency and a preset high weight coefficient allocated to the local grid frequency. Specifically, the weight coefficient of 0.3 allocated to the global grid frequency and the weight coefficient of 0.7 allocated to the local grid frequency are added and averaged to obtain the current frequency deviation. As the global disconnection time continues to increase and is less than the global connection time threshold, the weight of the local grid frequency allocation is further increased and the weight of the global grid frequency allocation is reduced.

[0074] The global connection time threshold refers to the pre-set time standard used to determine the time required for the global power grid frequency transmission delay to return to normal. When the global disconnection time is less than this threshold, it means that although the global power grid frequency has temporarily experienced a transmission delay, it is still possible to quickly restore the normal connection state. At this time, the current frequency deviation is calculated based on the global tilt weight coefficient. By giving the global power grid frequency a lower weight coefficient and the local power grid frequency a higher weight coefficient, it is possible to ensure the accuracy of frequency regulation while taking into account the impact of the global power grid frequency returning to normal. This ensures that during the global power grid frequency delay, the local power grid frequency can play a more important role in frequency regulation, making the entire power grid frequency regulation process more stable and reliable.

[0075] Step 323: When the global disconnection time exceeds the global connection time threshold, frequency adjustment is performed based on the local grid frequency.

[0076] When the global disconnection time exceeds the global connection time threshold, it indicates that the global grid frequency is not a short-term abnormality and the probability of returning to normal is very low. Therefore, the full weight of the local grid frequency is directly allocated, and the global grid frequency no longer participates in frequency regulation.

[0077] When the global disconnection time exceeds the global connection time threshold, the method for frequency regulation based on the local power grid frequency includes: Step 3230: Accumulate the local grid frequency deviations of the local nodes to obtain a local grid frequency deviation sum.

[0078] The local grid frequency deviation sum refers to a value obtained by accumulating and summing the local grid frequency deviations of all local nodes. Subsequent operations are performed after the local grid frequency deviation sum is obtained.

[0079] Step 3231: Obtain a theoretical local grid frequency deviation based on the local grid frequency deviation and a preset deviation and average value algorithm.

[0080] The deviation sum average algorithm calculates the average deviation by dividing the sum of the local grid frequency deviations by the number of local nodes. The result is the theoretical local grid frequency deviation. The theoretical local grid frequency deviation is an expected value for the local grid frequency deviation and is the standard value that the local grid frequency deviation should fall within during normal operation.

[0081] Step 3232: Obtain a local deviation based on the difference between the local grid frequency deviation and the theoretical local grid frequency deviation.

[0082] Local deviation refers to the difference between the actual deviation of the local grid frequency and the theoretical deviation. It reflects the degree of deviation of the local grid frequency from the ideal state under the current operating state.

[0083] Step 3233: Adjust the dynamic weight distribution based on the local deviation according to the preset deviation level weight method.

[0084] The deviation level weighting method categorizes the deviation level based on the absolute value of the local deviation, with each level corresponding to a different weight adjustment strategy. For example, when the absolute value of the local deviation is between 0.1 and 0.2 Hz, it is considered a mild deviation level. In this case, the weight of fast resources in the dynamic weight allocation should be appropriately increased. Fast resources, such as coal-fired units and flexible loads, can respond relatively quickly to frequency regulation needs and help quickly correct mild frequency deviations. When the absolute value of the local deviation is between 0.2 and 0.3 Hz, it is considered a moderate deviation level. In addition to further increasing the weight of fast resources, it is also necessary to appropriately increase the weight of slow resources, such as energy storage systems and gas turbines. These resources can leverage their stable regulation capabilities and work in synergy with fast resources for more effective frequency regulation. If the absolute value of the local deviation exceeds 0.3 Hz, it enters the severe deviation level. In addition to significantly increasing the weights of both fast and slow resources, the overall grid frequency regulation strategy may also need to be reassessed and optimized, such as by adjusting the frequency regulation coefficient, to address more severe frequency deviations and ensure that the grid frequency returns to a stable range as quickly as possible.

[0085] Step 3234: Calculate the weighted average of the local deviation based on the local deviation and the deviation level weight coefficient.

[0086] The weighted average of local deviations is calculated by multiplying the local deviation by the corresponding deviation level weight coefficient, adding all the multiplied results, and taking the average. This value is used to reflect the overall frequency deviation of the local grid, allowing for more accurate frequency regulation optimization based on the local grid frequency. In this way, the frequency regulation strategy can be appropriately adjusted based on the actual deviation of different local nodes, making the frequency regulation process more consistent with the actual operating conditions of the local grid, thereby improving the stability and effectiveness of the frequency regulation of the entire virtual power plant.

[0087] Step 3235: Output the weighted average value of the local deviation as the local power grid frequency deviation.

[0088] Frequency regulation is performed based on this weighted average of local deviations, replacing the local grid frequency deviation. According to the preset frequency regulation strategy, the local grid frequency is further fine-tuned to ensure that the local grid frequency is as close as possible to the standard frequency, maintaining the stable operation of the grid system. For example, if the weighted average of local deviations is positive, indicating that the local grid frequency is generally high, frequency regulation can be implemented by appropriately reducing the output power of local power generation equipment; if it is negative, the output power of local power generation equipment is increased accordingly.

[0089] The specific method of executing the global mode when the local delay time difference exceeds the local communication critical value includes: Step 360: Accumulate local disconnection time.

[0090] Local disconnection time refers to the duration of time the local grid frequency loses effective synchronization with the global grid due to transmission delays exceeding the local communication threshold. When the local delay time difference exceeds the local communication threshold, it indicates a significant delay in local grid frequency transmission, potentially causing loss of effective synchronization with the global grid. At this point, local disconnection time begins to accumulate. This operation, similar to global disconnection time, records the duration of the grid frequency anomaly, providing more accurate data support for subsequent decision-making.

[0091] Step 361: When the local disconnection time is less than the preset local connection time threshold, the current frequency deviation is calculated based on the preset local tilt weight coefficient and the global grid frequency deviation and the local grid frequency deviation, and is output.

[0092] The local tilt weight coefficient is a preset high weight coefficient allocated to the global grid frequency and a preset low weight coefficient allocated to the local grid frequency. Specifically, the weight coefficient of 0.8 allocated to the global grid frequency and the weight coefficient of 0.2 allocated to the local grid frequency are added and averaged to obtain the current frequency deviation. As the local disconnection time continues to increase and is less than the local connection time threshold, the weight of the global grid frequency allocation is further increased and the weight of the local grid frequency allocation is reduced.

[0093] When the local disconnection time is less than the local connection time threshold, it indicates that the local delay time difference may be less than the local communication critical value later, so the local tilt weight coefficient should be executed first based on the local grid frequency deviation and the global grid frequency deviation for joint frequency adjustment.

[0094] Step 362: When the local disconnection time exceeds the local connection time threshold, frequency adjustment is performed based on the global grid frequency.

[0095] When the local disconnection time exceeds the local connection time threshold, it means that the possibility of the local delay time difference being less than the local communication critical value becomes very small. There is no need to wait for the local delay time difference to recover to the local communication critical value range. The local grid frequency deviation is no longer reliable, so frequency regulation is directly performed through the global grid frequency.

[0096] Among them, when the local disconnection time is less than the local connection time threshold, the method of calculating and outputting the current frequency deviation based on the local tilt weight coefficient in combination with the global grid frequency deviation and the local grid frequency deviation includes:

[0097] Step 3610: Define the local delay time difference that exceeds the local communication critical value as an abnormal delay time difference.

[0098] Step 3611: Determine the local abnormal node based on the abnormal delay time difference.

[0099] A local abnormal node refers to an unreliable local power grid frequency deviation with excessively long communication transmission time. When an abnormal delay time difference occurs in a local node, it means that its local delay time difference exceeds the local communication critical value and the transmission time is too long. Therefore, it must be identified as a local abnormal node for subsequent operations.

[0100] Step 3612: Obtain the historical local power grid frequency deviation corresponding to the local abnormal node.

[0101] The historical local grid frequency deviation is the local grid frequency deviation recorded in the historical records. When a local abnormal node appears, it means that the local grid frequency deviation is no longer reliable. The historical local grid frequency deviation before the abnormality can be obtained for subsequent operations.

[0102] Step 3613: If there is no historical local power grid frequency deviation, determine other local nodes based on the local abnormal nodes.

[0103] Other local nodes refer to local nodes that are physically adjacent to the local abnormal node and do not include the local abnormal node. The specific determination method is to search through the preset local power grid topology structure, find the adjacent nodes directly connected to the local abnormal node, and determine them as other local nodes.

[0104] When there is no historical local grid frequency deviation, it means that the local grid frequency deviation of the current local abnormal node is unreliable and there is no effective historical local grid frequency deviation for alternative frequency adjustment, so subsequent operations must be performed by determining other local nodes.

[0105] Step 3614: Calculate the local grid frequency deviations corresponding to other local nodes according to the multi-node median algorithm to obtain a median local grid frequency deviation, and output it as the local grid frequency deviation.

[0106] The multi-node centering algorithm analyzes multiple nodes to obtain an approximate value. The specific steps will be described later and will not be repeated here. The multi-node centering algorithm uses an approximate value of the local grid frequency deviation to replace the local grid frequency deviation of the current abnormal node for frequency regulation.

[0107] Step 3615: If there is a historical local power grid frequency deviation, the historical local power grid frequency deviation is output as the local power grid frequency deviation.

[0108] When there is a historical local grid frequency deviation, it means that the local abnormal node can rely on the local grid frequency deviation before the abnormality to replace the current local grid frequency deviation for frequency regulation, reducing unnecessary tedious operations. Therefore, the historical local grid frequency deviation is output as the local grid frequency deviation.

[0109] Step 3616: Calculate the current frequency deviation based on the local tilt weight coefficient in combination with the global grid frequency deviation and the local grid frequency deviation and output it.

[0110] The specific calculation method is similar to step 361 and will not be described in detail here. The local grid frequency deviation obtained through the above process is multiplied by the local tilt weight coefficient, and the global grid frequency deviation is multiplied by the local tilt weight coefficient, and then the average is taken to obtain the current frequency deviation.

[0111] Reference Figure 4 The method of calculating the local grid frequency deviations corresponding to other local nodes according to the multi-node median algorithm to obtain a median local grid frequency deviation and outputting it as the local grid frequency deviation includes: Step 36140: Based on the local abnormal node, other local nodes are screened to determine adjacent nodes and the number of nodes.

[0112] The adjacent nodes are other local nodes whose corresponding local delay time differences do not exceed the local communication critical value. The adjacent nodes are determined by screening by judging whether the local delay time differences of other local nodes exceed the local communication critical value.

[0113] The number of nodes refers to the number of adjacent nodes. The default value is set to 0. Specifically, the number of nodes is accumulated each time an adjacent node is filtered out.

[0114] The local delay time difference does not exceed the local communication critical value, indicating that the local node is not a local abnormal local node, and its local power grid frequency deviation can be adopted.

[0115] Step 36141: If the number of nodes is less than the preset reliable reference number, execute the preset historical data compensation method.

[0116] The reliable reference number is a pre-set standard value used to determine whether the number of nodes is sufficiently reliable. If the number of nodes is less than this value, it indicates that the number of available adjacent nodes is small, and the local grid frequency deviation calculated based on these nodes may not be accurate or reliable. In this case, the preset historical data compensation method is implemented.

[0117] The historical data compensation method is a remedial method to prevent the multi-node centering algorithm from failing to execute normally when the number of nodes is less than the reliable reference number. The specific steps will be introduced in the subsequent process and will not be repeated here. If the number of nodes participating in the multi-node centering algorithm is less than the reliable reference number, then the algorithm will not be able to execute and even if it is executed, the reliability of the data obtained will be very low, so historical data compensation must be performed.

[0118] Step 36142: If the number of nodes is greater than the reliable reference number, the adjacent nodes are sorted according to a preset ascending algorithm based on the local grid frequency deviations corresponding to the adjacent nodes to obtain an adjacent node order.

[0119] An ascending algorithm sorts data from smallest to largest. The order of adjacent nodes is determined by sorting the local grid frequency deviations corresponding to adjacent nodes using this algorithm. If the number of nodes participating in the multi-node centering algorithm exceeds the reliable reference number, the data obtained by this algorithm will have a certain degree of reliability, and the multi-node centering algorithm can be executed.

[0120] Step 36143: Determine the median adjacent node based on the adjacent node sequence and the preset median algorithm, and define the local grid frequency deviation corresponding to the median adjacent node as the median local grid frequency deviation for output.

[0121] The median algorithm is an algorithm for obtaining the median in ordered data. For example, if the sequence is 1, 2, 4, 5, 5, 6, 10, then the median algorithm will calculate the median value of this sequence to be 5. When the median reaches a certain number of data points, the larger the number of data points, the smaller the deviation between the median data point and all the data points in the sequence. The median local grid frequency deviation refers to the local grid frequency deviation corresponding to the median adjacent node determined based on the adjacent node sequence and the median algorithm. The data obtained through this method will be very similar to the local grid frequency deviation of the local abnormal node, so the median local grid frequency deviation should be used for subsequent frequency modulation operations.

[0122] If the number of nodes is less than the reliable reference number, the method for performing historical data compensation includes: Step 361410: Reduce the reliable reference quantity according to the preset equal weakening algorithm to obtain the weakened reference quantity.

[0123] The equal-reduction algorithm reduces the number of reliable reference nodes by a fixed amount. For example, if the number of reliable reference nodes is originally 5, after one equal-reduction algorithm, the number of reliable reference nodes will be reduced to 4. This reduces the number of nodes required, making it possible to perform relevant calculations based on existing nodes even if the number of adjacent nodes is insufficient.

[0124] Step 361411: Compare the number of adjacent nodes based on the weakened reference number.

[0125] When the number of adjacent nodes cannot meet the reference number, the node selection algorithm will deadlock and will remain in the process of waiting for the number of adjacent nodes to meet the reference number. Therefore, to avoid this situation, the reference number can be reduced to obtain a weakened reference number to replace the reliable reference number, and the number of adjacent nodes can be compared based on the weakened reference number for subsequent operations.

[0126] Step 361412: If the number of adjacent nodes is less than the weakening reference number, summing the local grid frequency deviations of the adjacent nodes to obtain a sum of the adjacent node frequency deviations; When the number of adjacent nodes is less than the weakening reference number, it means that the current number of adjacent nodes still does not meet the conditions for executing the multi-node centering algorithm, so the local power grid frequency deviations of the adjacent nodes must be summed for subsequent operations.

[0127] Step 361413: Calculate the average value based on the frequency deviations of adjacent nodes and output it as the median local grid frequency deviation.

[0128] Because the current number of adjacent nodes is too small, the median is no longer reliable. Therefore, the average value of the adjacent node deviation and the sum divided by the number of adjacent nodes is calculated as the median local power grid frequency deviation for subsequent frequency regulation operations. When the amount of data is too small, the average value is closer to all the data in the group than the median.

[0129] Step 361414: If the number of adjacent nodes is greater than or equal to the weakening reference number, execute steps 36142 to 36143.

[0130] When the number of adjacent nodes is greater than or equal to the weakening reference number, it indicates that the requirements for executing the multi-node centering algorithm are met, so the specific steps of the multi-node centering algorithm are executed.

[0131] The method also includes a method for handling local abnormal nodes, which includes: Step 3617: When a local abnormal node exists, define the corresponding local abnormal node as a marked node.

[0132] When there is a local abnormal node, it indicates that the local node has an abnormality and needs to be processed, so it is defined as a marked node to facilitate subsequent processing operations.

[0133] Step 3618: Perform restart work on the marked node and accumulate the number of restarts.

[0134] Restarting the current marked node attempts to restore normal communication. Restart counts refer to the number of times a marked node has been restarted. After each restart, the number of restarts is accumulated and recorded to monitor the severity of the node's abnormality. Because most marked node failures are caused by internal factors such as data anomalies and excessive local latency, a simple self-restart can restore normal communication.

[0135] Step 3619: When the number of restarts is less than a preset self-starting number threshold, the local delay time difference of the marking node is re-determined, and the local delay time difference is defined as the secondary delay time difference.

[0136] The self-start count threshold is a range used to prevent node damage caused by frequent restarts. The secondary delay difference is the value obtained by re-determining the local delay difference of the marked node. If the number of restarts is less than the self-start count threshold, it indicates that there is still a possibility of restoring normal communication. The secondary delay difference can be used for further analysis and judgment. If the number of restarts is less than the self-start count threshold, it indicates that normal communication can still be restored.

[0137] Step 3620: If the secondary delay time difference does not exceed the local communication critical value, the marked node is regarded as a preset normal node and output.

[0138] When the secondary delay time difference does not exceed the local communication critical value, it indicates that the marked node has resumed normal communication, so it can be regarded as a normal node and subsequently operated according to the frequency modulation strategy of a normal local node without any special treatment.

[0139] Step 3621: If the secondary delay time difference exceeds the local communication critical value, continue to execute the restart operation.

[0140] If the secondary delay time difference exceeds the local communication critical value, it indicates that the marked node has not yet restored normal communication and needs to be restarted again. The hope is that restarting again will improve the local delay time difference and bring it back within the local communication critical value, thereby restoring the node's normal function in grid frequency regulation. After each restart, the relationship between the secondary delay time difference and the local communication critical value is continuously observed, and subsequent actions are determined based on the results.

[0141] Step 3622: If the number of restarts exceeds the critical value of the number of self-starts, it is defined as a fault node and a preset warning notification SMS is sent to the mobile phone of the current on-duty person.

[0142] If the number of restarts exceeds the threshold for self-starts, it indicates that after multiple restart attempts, the marked node has not been able to restore normal communication. The local delay difference continues to exceed the local communication threshold, and the node has been identified as a faulty node. At this point, an alert SMS message is sent to the on-duty personnel, allowing them to be notified and take further action.

[0143] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiment. All technical solutions based on the concept of the present invention are within the scope of protection of the present invention. It should be noted that for those skilled in the art, various improvements and modifications that do not depart from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A method for optimizing frequency regulation of a virtual power plant, characterized in that: include: Step 1: Collecting the global grid frequency and the local grid frequency in real time, wherein the global grid frequency is attached with a first precise time stamp, and the local grid frequency is attached with a second precise time stamp; Step 2: Calculate a time offset based on the first precise timestamp and the second precise timestamp; Step 3: When the time deviation is greater than a preset critical deviation value, executing a preset self-test mode; Step 4: when the time deviation is less than the critical deviation value, obtaining a global grid frequency deviation and a local grid frequency deviation by subtracting a preset standard frequency from the global grid frequency and the local grid frequency; Step 5: performing preset dynamic weight allocation based on the global grid frequency deviation and the local grid frequency deviation and calculating the current frequency deviation; Step 6: Optimize the global grid frequency and the local grid frequency based on the current frequency deviation.

2. The method for optimizing frequency regulation of a virtual power plant according to claim 1, characterized in that: The method for performing the dynamic weight allocation based on the global grid frequency and the local grid frequency includes: Step 50: Dynamically adjust the preset default weight coefficient according to the time deviation, and define the adjusted default weight coefficient as the modified weight coefficient; Step 51: Calculate the current frequency deviation based on the modified weight coefficient in combination with the global grid frequency deviation and the local grid frequency deviation, and output the calculated current frequency deviation.

3. The method for optimizing frequency regulation of a virtual power plant according to claim 1, characterized in that: The method for executing the self-test mode when the time deviation is greater than the critical deviation value includes: Step 30: Decomposing the first timestamp of the global power grid frequency to obtain a global sending time and a global arrival time; Step 31: Determine a global delay time difference based on the global sending time and the global arrival time; Step 32: Executing a preset local mode when the global delay time difference exceeds a preset global communication threshold, wherein the local mode is a mode in which a weight coefficient corresponding to the local grid frequency deviation is larger than a weight coefficient corresponding to the global grid frequency deviation during calculation of the current frequency deviation; Step 33: When the global delay time difference is lower than the global communication threshold, exit the local mode and execute steps 5 to 6; Step 34: Decompose the second timestamp of the local power grid frequency to obtain a local sending time and a local arrival time; Step 35: Determine a local delay time difference based on the local sending time and the local arrival time; Step 36: Executing a preset global mode when the local delay time difference exceeds a preset local communication threshold, wherein the global mode is a mode in which the weight coefficient corresponding to the global grid frequency deviation is larger than the weight coefficient corresponding to the local grid frequency deviation during calculation of the current frequency deviation; Step 37: When the local delay time difference is lower than the local communication threshold, exit the global mode and execute steps 5 to 6.

4. The method for optimizing frequency regulation of a virtual power plant according to claim 3, characterized in that: The method of executing the local mode when the global delay time difference exceeds the global communication threshold includes: Step 320: Accumulate the global disconnection time when the global delay time difference exceeds the global communication threshold; Step 321: searching for a historical global power grid frequency deviation, and using the historical global power grid frequency deviation as the global power grid frequency deviation; Step 322: When the global disconnection time is less than a preset global connection time threshold, calculating and outputting the current frequency deviation based on a preset global tilt weight coefficient in combination with the global grid frequency deviation and the local grid frequency deviation, wherein the global tilt weight coefficient is a preset low weight coefficient assigned to the global grid frequency and a preset high weight coefficient assigned to the local grid frequency; Step 323: When the global disconnection time exceeds the global connection time threshold, frequency adjustment is performed based on the local grid frequency.

5. The method for optimizing frequency regulation of a virtual power plant according to claim 4, characterized in that: When the global disconnection time exceeds the global connection time threshold, the method for frequency regulation based on the local power grid frequency includes: Step 3230: Accumulating the local grid frequency deviations of the local nodes to obtain a local grid frequency deviation sum; Step 3231: Obtain a theoretical local power grid frequency deviation based on the local power grid frequency deviation and a preset deviation and average value algorithm; Step 3232: Subtracting the local grid frequency deviation from the theoretical local grid frequency deviation to obtain a local deviation; Step 3233: Adjusting the dynamic weight distribution based on the local deviation according to a preset deviation level weighting method; Step 3234: Calculate a weighted average of local deviations based on the local deviations and the deviation level weight coefficients; Step 3235: Output the weighted average value of the local deviation as the local power grid frequency deviation.

6. The method for optimizing frequency regulation of a virtual power plant according to claim 3, characterized in that: The specific method of executing the global mode when the local delay time difference exceeds the local communication threshold includes: Step 360: Accumulate local disconnection time; Step 361: When the local disconnection time is less than a preset local connection time threshold, calculating and outputting the current frequency deviation based on a preset local tilt weight coefficient in combination with the global grid frequency deviation and the local grid frequency deviation, wherein the local tilt weight coefficient is a preset high weight coefficient assigned to the global grid frequency and a preset low weight coefficient assigned to the local grid frequency; Step 362: When the local disconnection time exceeds the local connection time threshold, perform frequency adjustment based on the global grid frequency.

7. The method for optimizing frequency regulation of a virtual power plant according to claim 6, characterized in that: When the local disconnection time is less than the local connection time threshold, the method for calculating and outputting the current frequency deviation based on the local tilt weight coefficient in combination with the global grid frequency deviation and the local grid frequency deviation includes: Step 3610: defining the local delay time difference exceeding the local communication threshold as an abnormal delay time difference; Step 3611: Determine a local abnormal node based on the abnormal delay time difference; Step 3612: Obtain the historical local power grid frequency deviation corresponding to the local abnormal node; Step 3613: If the historical local power grid frequency deviation does not exist, determine other local nodes based on the local abnormal node; Step 3614: Calculate the local grid frequency deviations corresponding to the other local nodes according to a preset multi-node median algorithm to obtain a median local grid frequency deviation, and output the median local grid frequency deviation as the local grid frequency deviation; Step 3615: if the historical local grid frequency deviation exists, output the historical local grid frequency deviation as the local grid frequency deviation; Step 3616: Calculate the current frequency deviation based on the local tilt weight coefficient in combination with the global grid frequency deviation and the local grid frequency deviation and output it.

8. The method for optimizing frequency regulation of a virtual power plant according to claim 7, characterized in that: The method of calculating the local grid frequency deviations corresponding to the other local nodes according to the multi-node median algorithm to obtain the median local grid frequency deviation, and outputting the median local grid frequency deviation as the local grid frequency deviation includes: Step 36140: Screening the other local nodes based on the local abnormal node to determine adjacent nodes and the number of nodes, wherein the adjacent nodes are the other local nodes whose corresponding local delay time differences do not exceed the local communication critical value; Step 36141: If the number of nodes is less than a preset reliable reference number, executing a preset historical data compensation method; Step 36142: If the number of nodes is greater than the reliable reference number, the adjacent nodes are sorted according to a preset ascending algorithm based on the local grid frequency deviations corresponding to the adjacent nodes to obtain an adjacent node order; Step 36143: Determine a median adjacent node based on the adjacent node sequence and a preset median algorithm, and define the local power grid frequency deviation corresponding to the median adjacent node as a median local power grid frequency deviation for output.

9. The method for optimizing frequency regulation of a virtual power plant according to claim 8, characterized in that: If the number of nodes is less than the reliable reference number, the method for performing historical data compensation includes: Step 361410: reducing the reliable reference quantity according to a preset equal weakening algorithm to obtain a weakened reference quantity; Step 361411: Compare the weakening reference number with the number of adjacent nodes; Step 361412: If the number of the adjacent nodes is less than the weakening reference number, summing the local grid frequency deviations of the adjacent nodes to obtain a sum of adjacent node frequency deviations; Step 361413: Calculate an average value based on the adjacent node frequency deviations and output it as the median local grid frequency deviation; Step 361414: If the number of adjacent nodes is greater than or equal to the weakening reference number, execute steps 36142 to 36143.

10. The method for optimizing frequency regulation of a virtual power plant according to claim 7, characterized in that: Also included is a method for processing the local abnormal node, the method comprising: Step 3617: If the local abnormal node exists, define the corresponding local abnormal node as a marked node; Step 3618: Restart the marked node and accumulate the number of restarts. Step 3619: When the restart count is less than a preset self-start count threshold, redetermine the local delay time difference of the marking node, and define the local delay time difference as a secondary delay time difference; Step 3620: If the secondary delay time difference does not exceed the local communication critical value, the marked node is regarded as a preset normal node and output; Step 3621: If the secondary delay time difference exceeds the local communication threshold, continue to perform the restart operation; Step 3622: If the restart times exceed the self-start times critical value, it is defined as a fault node and a preset warning notification SMS is sent to the mobile phone of the current on-duty personnel.

Images