Load aggregation control method for distribution network in high proportion distributed power supply scenario

Abstract

This invention relates to the field of predictive control technology for distribution networks, specifically to a load aggregation control method for distribution networks in scenarios with a high proportion of distributed generation (DG). This method determines predictive stability by analyzing the differences between predicted and actual data from each historical prediction of DG; it obtains the true reliability of DG by analyzing prediction deviations between DG and similar DGs, and then samples true deviations; based on the distribution of these true deviation samples, it determines a severity index of the deviation and, combined with the reliability of the true deviation samples, obtains an instability correction coefficient. After correcting the predictive instability, it adjusts the error weight matrix in the MPC algorithm for control. This invention analyzes the impact of true deviations in cases of high prediction deviations, corrects predictive instability results, and adjusts the weights in the MPC algorithm to more accurately reflect the fluctuation impact of each DG, thereby improving the accuracy of the closed-loop control results of the MPC.

Description

Technical Field

[0001] This invention relates to the field of predictive control technology for distribution networks, and specifically to a control method for distribution network load aggregation in scenarios with a high proportion of distributed power sources. Background Technology

[0002] With the large-scale integration of distributed power sources such as photovoltaics and wind power, the distribution network has transformed from a traditional single-source radial structure into a complex network of interconnected multiple power sources. Among them, distributed power sources, represented by photovoltaics and wind power, are growing in the distribution network. While this high proportion of distributed energy injection brings clean electricity to the grid, its inherent intermittency, volatility, and instability also require close attention to the safety and stability of the distribution network.

[0003] The essence of distribution network load aggregation control is to package and integrate a large number of small distributed energy resources that are geographically dispersed and owned by different users, forming a resource pool of a certain scale that can respond to grid dispatch commands and appear as a single controllable unit to the outside world. The commonly used MCP algorithm is a model-based advanced optimization control strategy that performs dynamic collaborative optimization in distribution network load aggregation control, coordinating controllable units within the aggregation to cope with distributed power source fluctuations.

[0004] However, because load aggregation contains a large number of controllable units, its prediction accuracy for distributed power output and load demand is highly dependent. In practical applications, the fluctuations of each energy storage unit have a certain degree of randomness, leading to frequent fluctuations in grid power. This, in turn, results in different availability of prediction results from different power sources, compromising the accuracy of subsequent MPC closed-loop control results. Summary of the Invention

[0005] To address the technical problems in the prior art, the present invention aims to provide a control method for distribution network load aggregation in scenarios with a high proportion of distributed power sources. The specific technical solution adopted is as follows:

[0006] The first aspect of this application provides a method for controlling distribution network load aggregation in a high-proportion distributed generation scenario, the method comprising:

[0007] At each distributed power source, obtain the predicted output power sequence and the actual power sequence after each historical control sampling.

[0008] For each distributed power source, the deviation index of each control sampling is obtained by measuring the trend deviation between the actual power sequence and the predicted output power sequence after each historical control sampling. Based on the magnitude of the deviation index, deviation samples are selected from all control samples. The distribution characteristics of the deviation samples and the corresponding deviation index in the time series are analyzed to obtain the prediction instability.

[0009] For each distributed power source, the true reliability of each deviation sample is determined based on the difference between the deviation index of each deviation sample and that of the same type of distributed power source. Based on the true reliability, the true deviation samples are selected from the deviation samples. The degree of continuous distribution and clustering of the true deviation samples in time series is analyzed to determine the deviation severity index of the distributed power source. Combining the deviation severity index and the distribution of the true reliability corresponding to the true deviation samples, the true impact of the deviation is analyzed to obtain the instability correction coefficient.

[0010] The instability of a distributed power source is determined by combining the predicted instability and the instability correction coefficient of each distributed power source; the error weight adjustment in the MPC algorithm is then controlled based on the instability.

[0011] Furthermore, the method for obtaining the deviation index includes:

[0012] For any single control sampling of a distributed power source, the correlation between the predicted output power sequence and the actual power sequence after the control sampling is negatively correlated and normalized to obtain the trend deviation of the control sampling.

[0013] After calculating the numerical difference between the predicted power output sequence and the actual power sequence at each time step, the mean of all differences is used as the numerical deviation of the control sample.

[0014] The product of the trend deviation and the numerical deviation of the control sample is normalized to obtain the deviation index of the control sample.

[0015] Furthermore, the method for obtaining the deviation sampling includes:

[0016] In the timing sequence of each distributed power source, control samples with deviation indicators greater than a preset deviation threshold are used as deviation samples.

[0017] Furthermore, the method for obtaining the predicted instability includes:

[0018] For any distributed power source, the mean of the deviation index of all deviation samples in the time series is calculated and used as the predicted deviation index of the distributed power source.

[0019] The predicted instability of the distributed power source is obtained by normalizing the product of the total number of time-series deviation samples and the predicted deviation index.

[0020] Furthermore, the method for obtaining the authenticity and credibility includes:

[0021] For any one deviation sampling of a distributed power source, calculate the difference between the deviation index of the distributed power source and each other of the same type of distributed power source at the time of the deviation sampling, and normalize the mean of the difference between the deviation index of the distributed power source and all other distributed power sources of the same type to obtain the true credibility of the distributed power source at the time of the deviation sampling.

[0022] When the load device is in an abnormal state during the deviation sampling, the true reliability of the deviation sampling is recorded as the preset highest reliability value.

[0023] Furthermore, the method for obtaining the true deviation sampling includes:

[0024] For each distributed power source, deviation samples with a true credibility greater than a preset credibility threshold are used as true deviation samples in the time sequence.

[0025] Furthermore, the method for obtaining the severity index of the deviation includes:

[0026] For any distributed power source, sample the real deviations of consecutively adjacent distributions in the time series of the distributed power source as a continuous deviation distribution sequence; calculate the time interval between every two consecutive deviation distribution sequences, and perform a negative correlation mapping on the sum of all time intervals as the continuous aggregation degree of the distributed power source.

[0027] The total number of samples of all real deviations in the continuous deviation distribution sequence of the distributed power source is taken as the distribution proportion of the distributed power source.

[0028] The product of the distribution ratio and the degree of continuous aggregation of the distributed power source is normalized to obtain the severity index of the deviation of the distributed power source.

[0029] Furthermore, the method for obtaining the instability correction coefficient includes:

[0030] For any distributed power source, the sum of the true confidence values ​​of all deviation samples on the time series of the distributed power source is used as the true impact index of the distributed power source.

[0031] The total number of all real deviation samples on the time series of the distributed power source is taken as the number of real feedbacks of the distributed power source; the product of the deviation severity index, the number of real feedbacks, and the real impact index of the distributed power source is normalized and taken as the instability correction coefficient of the distributed power source.

[0032] Furthermore, the method for obtaining the instability includes:

[0033] The product of the instability correction factor and the predicted instability of each distributed power source is normalized to obtain the instability of each distributed power source.

[0034] Furthermore, the control of error weight adjustment in the MPC algorithm based on instability includes:

[0035] Obtain the error weight matrix in the current MPC algorithm; multiply the instability of each distributed power source by the error weight matrix to obtain the optimized error weight matrix of each distributed power source, and output the optimal control strategy of each distributed power source through the MCP algorithm.

[0036] Secondly, this application provides a distribution network load aggregation control system for a high-proportion distributed power generation scenario, the system comprising:

[0037] The data acquisition module is used to acquire the predicted output power sequence and the actual power sequence after each historical control sampling at each distributed power source;

[0038] The instability analysis module is used to obtain the deviation index for each control sampling for each distributed power source by measuring the trend deviation between the actual power sequence and the predicted output power sequence after each historical control sampling; based on the magnitude of the deviation index, it filters the deviation samples from all control samples; and analyzes the distribution characteristics of the deviation samples and the corresponding deviation index in the time series to obtain the predicted instability.

[0039] The correction analysis module is used to determine the true reliability of each deviation sample for each distributed power source based on the difference between the deviation index of each deviation sample and that of the same type of distributed power source; to filter out the true deviation samples from the deviation samples based on the true reliability; to analyze the degree of continuous distribution and clustering of the true deviation samples in time series to determine the deviation severity index of the distributed power source; and to analyze the true impact of the deviation and obtain the instability correction coefficient by combining the deviation severity index and the distribution of the true reliability corresponding to the true deviation samples.

[0040] The adjustment control module is used to determine the instability of the distributed power source by combining the predicted instability and instability correction coefficient of each distributed power source; and to control the adjustment of error weights in the MPC algorithm based on the instability.

[0041] Thirdly, this application provides a computer device including a memory and a processor. The memory is used to store computer program code, and the processor is used to call and run the computer program code from the memory to perform the method as described in the first aspect of this application or any embodiment of the first aspect.

[0042] Fourthly, this application provides a computer program product comprising computer program code, which, when executed, performs the method as described in the first aspect of this application or any embodiment thereof.

[0043] Fifthly, this application provides a computer-readable storage medium that stores computer program code, which, when executed, performs the method as described in the first aspect of this application or any embodiment thereof.

[0044] The present invention has the following beneficial effects:

[0045] This invention reflects the instability of prediction results by analyzing the differences between predicted and actual conditions for each distributed power source in each historical prediction. Then, by analyzing the prediction deviations between each distributed power source and similar distributed power sources, considering the impact of temporary scheduling, it measures the credibility of the actual impact of deviations on each distributed power source, obtains the credibility, and determines the sampling time of the actual deviations. Since the short-term effects of predictions may show a clustered distribution of severe deviations, the distribution of the actual deviation sampling is used to determine the severity, and the credibility of the deviations is combined to reflect the degree of the actual impact in the instability analysis, resulting in an instability correction coefficient. This coefficient is used to correct the prediction instability, making the analysis of the predicted instability results more accurate. The final instability is then used to adjust the error weight matrix in the MPC algorithm to obtain the optimal control strategy for each distributed power source. This invention addresses the issue of high prediction deviations by analyzing the impact of actual deviations on the prediction instability results based on the synchronization deviations of similar distributed power sources, and then optimizing the weights in the MPC algorithm to more accurately reflect the fluctuation impact of each distributed power source and improve the accuracy of the closed-loop control results of the MPC. Attached Figure Description

[0046] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0047] Figure 1 This is a flowchart of a distribution network load aggregation control method in a high-proportion distributed power generation scenario, provided by an embodiment of the present invention.

[0048] Figure 2 This is a structural diagram of a distribution network load aggregation control system in a high-proportion distributed power generation scenario, provided by an embodiment of the present invention.

[0049] Figure 3 This is a schematic diagram of a computer device structure provided in one embodiment of the present invention. Detailed Implementation

[0050] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a distribution network load aggregation control method in a high-proportion distributed power generation scenario proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0051] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0052] The following description, in conjunction with the accompanying drawings, details a specific scheme for a distribution network load aggregation control method in a high-proportion distributed power generation scenario provided by the present invention.

[0053] Please see Figure 1 The diagram illustrates a flowchart of a distribution network load aggregation control method in a high-proportion distributed power generation scenario, according to an embodiment of the present invention. The method includes:

[0054] S1: Obtain the predicted output power sequence and actual power sequence after each historical control sampling at each distributed power source.

[0055] In this embodiment of the invention, power sensors are installed at distributed power sources, such as wind turbines and photovoltaic panels, to monitor their operating status, collect real-time power output, and operational data such as temperature and voltage. Simultaneously, the corresponding load side is monitored, and faulty or out-of-service loads are marked with abnormal status using smart meters or flexible load controllers to facilitate more comprehensive regulation.

[0056] In one specific embodiment of the present invention, in order to realize the MPC algorithm for subsequent distribution network strategy regulation, at the control sampling time of regulation, each distributed power source is input into the trained LST+attention mechanism model based on historically collected operating data such as temperature, power, and voltage to predict the output power sequence for a future period of time. The time interval between each control sampling prediction is the duration of the predicted output power sequence. The control sampling involved in the analysis can be limited to sampling within a historical day. It should be noted that the limitation of the historical time series can be adjusted by the implementer, similar to the time period of a phased regulation task requiring regulation, and is not restricted here.

[0057] In this embodiment of the invention, the prediction time is set to 30 minutes. The prediction method is a publicly known technique known to those skilled in the art. The specific prediction settings can be adjusted by the implementer according to the specific implementation scenario, and no restrictions are imposed here.

[0058] S2: For each distributed power source, obtain the deviation index of each control sampling by the trend deviation between the actual power sequence and the predicted output power sequence after each historical control sampling; filter the deviation samples from all control samples based on the magnitude of the deviation index; analyze the distribution characteristics of the deviation samples and the corresponding deviation index in the time series to obtain the prediction instability.

[0059] For each distributed power source's historical prediction data, prediction stability can be assessed by the degree of deviation between actual data collection and prediction. The greater the deviation between the predicted output power after control sampling and the actual power, the lower the effectiveness of the prediction result. In this embodiment of the invention, the method for obtaining the deviation index includes:

[0060] For any single control sampling of a distributed power source, the correlation between the predicted output power sequence and the actual power sequence after the control sampling is negatively correlated and normalized to obtain the trend deviation of the control sampling. In this embodiment of the invention, the Pearson correlation coefficient can be used to obtain the correlation between the sequences. The larger the correlation, the more similar the prediction and the actual result are, and the lower the deviation. The method of obtaining the correlation between sequences using the Pearson correlation coefficient is a well-known technique to those skilled in the art. The calculation of the correlation between sequences can also use dynamic time warping or Spearman rank correlation coefficient, which will not be elaborated here.

[0061] It should be noted that negative correlation mapping and normalization are numerical techniques used by those skilled in the art. For example, negative correlation mapping can be in the form of a negative exponent with the natural constant as the base or inverse proportional form, and the choice of normalization can be linear normalization or standard normalization, etc. The specific method is not limited here.

[0062] Furthermore, after calculating the numerical difference between the predicted power output sequence and the actual power sequence at each time step, the average of all differences is used as the numerical deviation of the control sampling, reflecting the degree of difference between the overall values. The greater the numerical deviation, the less similar the predicted result is to the actual result.

[0063] Finally, the product of the trend deviation and the numerical deviation of the control sample is normalized to obtain the deviation index of the control sample. Through the deviation analysis between the trend and the numerical values, the larger the deviation index, the smaller the similarity between the prediction and the actual results.

[0064] Then, samples with large deviations are selected to represent samples whose prediction results are relatively uncontrollable. In this embodiment of the invention, on the timing of each distributed power source, control samples with deviation indices greater than a preset deviation threshold are used as deviation samples. The preset deviation threshold is set to 0.6, and the specific value can be adjusted by the implementer and is not limited here.

[0065] For distributed generation, a large number of significant deviations from the predicted outcome indicate a higher probability of prediction failure, requiring greater attention to error. Therefore, the instability of the prediction is measured by the number of deviation samples and the severity of the deviation index. In this embodiment of the invention, the method for obtaining prediction instability includes:

[0066] For any distributed generation, the mean of the deviation index of all deviation samples in the time series is calculated as the prediction deviation index of the distributed generation, reflecting the degree of instability with high deviation. The product of the total number of deviation samples in the time series of the distributed generation and the prediction deviation index is normalized to obtain the prediction instability of the distributed generation. The more deviation samples there are and the greater the deviation, the higher the instability and the larger the error of the distributed generation in the prediction analysis.

[0067] S3: For each distributed power source, determine the true reliability of each deviation sample based on the difference between the deviation index and the deviation index of each deviation sample of the same type of distributed power source; select the true deviation samples from the deviation samples based on the true reliability; analyze the degree of continuous distribution and clustering of the true deviation samples in time series to determine the deviation severity index of the distributed power source; combine the deviation severity index and the distribution of the true reliability corresponding to the true deviation samples to analyze the true impact of the deviation and obtain the instability correction coefficient.

[0068] When distributed power sources exhibit predictive instability, it may also be due to temporary power dispatch. When temporary power dispatch is required, grid loads of the same type usually need to be dispatched in a unified manner. Therefore, distributed power sources of the same type should have similar deviations when the same deviation is sampled.

[0069] Therefore, further similarity analysis is performed on similar deviation indicators. The lower the similarity, the higher the probability of a real prediction error. In this embodiment of the invention, the method for obtaining the true reliability includes:

[0070] For any single deviation sample of a distributed power source, the difference in deviation indices between this distributed power source and every other distributed power source of the same type at that deviation sample is calculated. Then, the mean of these differences is normalized to obtain the true reliability of the distributed power source at that deviation sample. When the overall difference in deviation indices between the distributed power source and other distributed power sources of the same type is high, it indicates a higher probability that the predicted deviation of this distributed power source is actually unstable, thus increasing its true reliability.

[0071] Specifically, when the load device is in an abnormal state during the deviation sampling, it indicates that there is a difference on the load side itself, and the reliability of its prediction deviation is relatively high. Therefore, no synchronous analysis is performed. Thus, the true reliability of the deviation sampling is recorded as the preset highest reliability value. In this embodiment of the invention, the preset highest reliability value is set to 1. The implementer can adjust it as they see fit, and there are no restrictions here.

[0072] The authenticity and reliability can be used to further filter out samples that are more unstable and significant. In this embodiment of the invention, in the time series of each distributed power source, the deviation sample with authenticity and reliability greater than the preset confidence threshold is taken as the true deviation sample. The preset confidence threshold is set to 0.5, and the specific value can be adjusted by the implementer.

[0073] Based on the sampling of actual deviations in distributed power sources, it can reflect the significant situation of high prediction errors in historical forecasts. Since the continuous prediction and adjustment process will mitigate and correct the errors, when there are high clustered distributions of actual deviations in a short period of time, it reflects that the instability of actual prediction is more serious and there is a possibility of short-term prediction failure. In the future, more attention should be paid to the impact of its deviations.

[0074] Preferably, in this embodiment of the invention, the method for obtaining the index of severe deviation by analyzing the aggregation distribution includes:

[0075] First, for any distributed power source, the true deviations of consecutively adjacent distributions in the time series are sampled as a continuous deviation distribution sequence, representing sampling periods with consistently high and significant deviations. Then, the time interval between each pair of adjacent continuous deviation distribution sequences is calculated, and the sum of all time intervals is negatively correlated to represent the continuous clustering degree of the distributed power source. The degree of clustering of short-term high deviations is determined by the intervals between consecutive high deviation sequences; the shorter the time interval, the more significant the high clustering.

[0076] Furthermore, in the time series of this distributed power source, the total number of all real deviation samples in the continuous deviation distribution sequence is taken as the distribution proportion of this distributed power source. The higher the proportion of continuous existence, the more serious the prediction instability stage exists, and the higher the demand for instability attention.

[0077] Finally, the product of the distribution ratio and the continuous aggregation degree of the distributed power source is normalized to obtain the deviation severity index of the distributed power source. By combining the proportion of continuous real deviation sampling and the time interval analysis, the aggregation severity of the real deviation is quantified. The larger the deviation severity index, the higher the instability concern requirement.

[0078] In particular, if there are no multiple continuously distributed deviation sequences, that is, there are no time intervals that can be analyzed, then continuous clustering analysis is not performed, and the distribution proportion reflects the severity of the deviation.

[0079] The predicted instability is adjusted based on the analysis of the actual deviation sampling. The more severe the deviation reflected by the actual deviation sampling and the higher its proportion, the higher the reliability of the instability analysis. In this embodiment of the invention, the method for obtaining the instability correction coefficient includes:

[0080] For any distributed power source, the sum of the true confidence scores of all deviation samples on the time series of that distributed power source is used as the true impact index of that distributed power source. This comprehensive confidence score reflects the participation of true deviations in the overall impact of the overall deviations. Furthermore, the total number of true deviation samples on the time series of that distributed power source is used as the true feedback quantity of that distributed power source. A higher number of true deviation samples indicates a significant overall prediction error and a greater need for attention.

[0081] Finally, the product of the severity index of the distributed power source deviation, the number of actual feedbacks, and the actual impact index is normalized and used as the instability correction coefficient of the distributed power source. The larger the instability correction coefficient, the higher the error of the predicted instability situation, and the greater the degree of correction that needs to be paid attention to, and the more reliable the instability analysis.

[0082] S4: Combine the predicted instability and instability correction coefficient of each distributed power source to determine the instability of the distributed power source; control the adjustment of error weights in the MPC algorithm based on the instability.

[0083] By adjusting the predicted instability using an instability correction coefficient, the true degree of instability concern in the predicted results can be reflected more accurately and reliably. In this embodiment of the invention, the product of the instability correction coefficient and the predicted instability for each distributed power source is normalized to obtain the instability of each distributed power source. The larger the instability correction coefficient, the higher the demand for instability control, and therefore the greater the instability. The error weights of the distributed power sources in the MPC algorithm are adjusted according to the instability situation to improve the accuracy of the control results.

[0084] In this embodiment of the invention, the error weight matrix in the current MPC algorithm is obtained. The error weight matrix corresponds to the error weight at each moment in the prediction time domain. Its function is to adjust the system's emphasis on tracking error. The larger the weight, the more the system focuses on reducing the deviation between the output and the reference value at that moment. Therefore, the instability of each distributed power source is multiplied by the error weight matrix to obtain the optimized error weight matrix for each distributed power source. Then, the optimal control strategy for each distributed power source is output through the MCP algorithm. The control strategy may include: active power output command, energy storage charging and discharging power, flexible load adjustment, etc., and then control command allocation is performed. The optimal control strategy for each distributed power source is decomposed into commands and then issued and executed. After the control period after sampling ends, MPC optimization is repeated to achieve cyclic control of the distribution network load aggregation in a high proportion of distributed power source scenarios.

[0085] In summary, this invention reflects the instability of prediction results by analyzing the differences between the predicted and actual situations for each distributed power source in each historical prediction. Then, by analyzing the prediction deviations between each distributed power source and similar distributed power sources, considering the impact of temporary scheduling, it measures the credibility of the actual impact deviation of each distributed power source, obtains the true credibility, and determines the sampling time of the true deviation. Since the short-term effects of predictions may show a clustered distribution of severe deviations, the distribution of the sampled true deviations is used to determine the severity of the situation. The credibility of the deviations is then considered to reflect the degree of the true impact in the instability analysis, resulting in an instability correction coefficient. This coefficient is used to correct the prediction instability, making the analysis of the predicted instability results more accurate. The final instability is then used to adjust the error weight matrix in the MPC algorithm to obtain the optimal control strategy for each distributed power source. This invention, for cases with high prediction deviations, analyzes the impact of the true deviations on the prediction instability results based on the synchronization deviations of similar distributed power sources, and optimizes the weights in the MPC algorithm to more accurately reflect the fluctuation impact of each distributed power source and improve the accuracy of the closed-loop control results of the MPC.

[0086] This application also provides a distribution network load aggregation control system for a high proportion of distributed power generation scenarios. Please refer to [link / reference]. Figure 2 The diagram illustrates a structural diagram of a distribution network load aggregation control system in a high-proportion distributed power generation scenario according to an embodiment of the present invention. The system includes: a data acquisition module 201, an instability analysis module 202, a correction analysis module 203, and an adjustment control module 204.

[0087] The data acquisition module 201 is used to acquire the predicted output power sequence and the actual power sequence after each historical control sampling at each distributed power source;

[0088] The instability analysis module 202 is used to obtain the deviation index of each control sampling for each distributed power source by the trend deviation between the actual power sequence and the predicted output power sequence after each historical control sampling; to filter the deviation samples from all control samples based on the magnitude of the deviation index; and to obtain the predicted instability by analyzing the distribution characteristics of the deviation samples and the corresponding deviation index in the time series.

[0089] The correction analysis module 203 is used to determine the true reliability of each deviation sample for each distributed power source based on the difference between the deviation index of each deviation sample and that of the same type of distributed power source; to filter out the true deviation samples from the deviation samples based on the true reliability; to analyze the degree of continuous distribution and clustering of the true deviation samples in time series to determine the deviation severity index of the distributed power source; and to analyze the true impact of the deviation and obtain the instability correction coefficient by combining the deviation severity index and the distribution of the true reliability corresponding to the true deviation samples.

[0090] The adjustment control module 204 is used to determine the instability of the distributed power source by combining the predicted instability and instability correction coefficient of each distributed power source; and to control the adjustment of the error weight in the MPC algorithm based on the instability.

[0091] It should be noted that the system provided in the above embodiments is only an example of the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer equipment can be divided into different functional modules to complete all or part of the functions described above. In addition, the distribution network load aggregation control system and the distribution network load aggregation control method embodiment in a high proportion of distributed power generation scenario provided in the above embodiments belong to the same concept, and their specific implementation process can be found in the method embodiment, which will not be repeated here.

[0092] This application also provides a computer device; please refer to [link / reference]. Figure 3 The illustration shows a schematic diagram of a computer device structure according to an embodiment of the present invention. The computer device includes a memory 301, a processor 302, and a computer program 303 stored in the memory 301 and running on the processor 302. When the processor 302 executes the computer program 303, the computer device can execute any of the aforementioned high-proportion distributed power source scenario distribution network load aggregation control methods.

[0093] This application also provides a computer program product that, when run on a computer device, enables the computer device to execute any of the aforementioned high-proportion distributed power source scenario distribution network load aggregation control methods.

[0094] This application also provides a computer-readable storage medium storing computer program code. When the computer program code is run on a computer device, the computer device can execute any of the aforementioned high-proportion distributed power source scenario distribution network load aggregation control methods.

[0095] In the embodiments provided in this application, it should be understood that the computer device, computer program product and computer-readable storage medium provided are all used to perform the corresponding methods provided above, and therefore the beneficial effects they can achieve can be referred to the beneficial effects of the methods provided above, which will not be repeated here.

[0096] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0097] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A method for controlling load aggregation in a distribution network under a high proportion of distributed power generation scenarios, characterized in that, The method includes: At each distributed power source, obtain the predicted output power sequence and the actual power sequence after each historical control sampling. For each distributed power source, the deviation index of each control sampling is obtained by measuring the trend deviation between the actual power sequence and the predicted output power sequence after each historical control sampling. Based on the magnitude of the deviation index, deviation samples are selected from all control samples. The distribution characteristics of the deviation samples and the corresponding deviation index in the time series are analyzed to obtain the prediction instability. For each distributed power source, the true reliability of each deviation sample is determined based on the difference between the deviation index of each deviation sample and that of the same type of distributed power source. Based on the true reliability, the true deviation samples are selected from the deviation samples. The degree of continuous distribution and clustering of the true deviation samples in time series is analyzed to determine the deviation severity index of the distributed power source. Combining the deviation severity index and the distribution of the true reliability corresponding to the true deviation samples, the true impact of the deviation is analyzed to obtain the instability correction coefficient. The instability of a distributed power source is determined by combining the predicted instability and the instability correction coefficient for each distributed power source; the error weight adjustment in the MPC algorithm is then controlled based on the instability. The method for obtaining the predicted instability includes: For any distributed power source, the mean of the deviation index of all deviation samples in the time series is calculated and used as the predicted deviation index of the distributed power source. The product of the total number of time-series deviation samples of the distributed power source and the prediction deviation index is normalized to obtain the prediction instability of the distributed power source. The method for obtaining the instability correction coefficient includes: For any distributed power source, the sum of the true confidence values ​​of all deviation samples on the time series of the distributed power source is used as the true impact index of the distributed power source. The total number of all real deviation samples on the time series of the distributed power source is taken as the number of real feedbacks of the distributed power source; the product of the deviation severity index, the number of real feedbacks, and the real impact index of the distributed power source is normalized and taken as the instability correction coefficient of the distributed power source.

2. The method for controlling distribution network load aggregation in a high-proportion distributed power generation scenario according to claim 1, characterized in that, The method for obtaining the deviation index includes: For any single control sampling of a distributed power source, the correlation between the predicted output power sequence and the actual power sequence after the control sampling is negatively correlated and normalized to obtain the trend deviation of the control sampling. After calculating the numerical difference between the predicted power output sequence and the actual power sequence at each time step, the average of all differences is used as the numerical deviation of the control sample. The product of the trend deviation and the numerical deviation of the control sample is normalized to obtain the deviation index of the control sample.

3. The method for controlling distribution network load aggregation in a high-proportion distributed power generation scenario according to claim 1, characterized in that, The method for obtaining the deviation sampling includes: In the timing sequence of each distributed power source, control samples with deviation indicators greater than a preset deviation threshold are used as deviation samples.

4. The method for controlling distribution network load aggregation in a high-proportion distributed power generation scenario according to claim 1, characterized in that, The methods for obtaining the authenticity and credibility include: For any one deviation sampling of a distributed power source, calculate the difference between the deviation index of the distributed power source and each other of the same type of distributed power source at the time of the deviation sampling, and normalize the mean of the difference between the deviation index of the distributed power source and all other distributed power sources of the same type to obtain the true credibility of the distributed power source at the time of the deviation sampling. When the load device is in an abnormal state during the deviation sampling, the true reliability of the deviation sampling is recorded as the preset highest reliability value.

5. The method for controlling distribution network load aggregation in a high-proportion distributed power generation scenario according to claim 1, characterized in that, The method for obtaining the true deviation sampling includes: For each distributed power source, deviation samples with a true credibility greater than a preset credibility threshold are used as true deviation samples in the time sequence.

6. The method for controlling distribution network load aggregation in a high-proportion distributed power generation scenario according to claim 1, characterized in that, The method for obtaining the severity index of the deviation includes: For any distributed power source, sample the real deviations of consecutively adjacent distributions in the time series of the distributed power source as a continuous deviation distribution sequence; calculate the time interval between every two consecutive deviation distribution sequences, and perform a negative correlation mapping on the sum of all time intervals as the continuous aggregation degree of the distributed power source. In the time series of the distributed power source, the total number of samples of all real deviations in the continuous distribution sequence of deviations is taken as the distribution proportion of the distributed power source. The product of the distribution ratio and the degree of continuous aggregation of the distributed power source is normalized to obtain the severity index of the deviation of the distributed power source.

7. The method for controlling distribution network load aggregation in a high-proportion distributed power generation scenario according to claim 1, characterized in that, The method for obtaining the instability includes: The product of the instability correction factor and the predicted instability of each distributed power source is normalized to obtain the instability of each distributed power source.

8. The method for controlling distribution network load aggregation in a high-proportion distributed power generation scenario according to claim 1, characterized in that, The control of error weight adjustment in the MPC algorithm based on instability includes: Obtain the error weight matrix in the current MPC algorithm; multiply the instability of each distributed power source by the error weight matrix to obtain the optimized error weight matrix of each distributed power source, and output the optimal control strategy of each distributed power source through the MCP algorithm.

Images